Proteomic Analysis of Thermus thermophilus Cells after Treatment with Antimicrobial Peptide

Abstract

In recent years, the study of antimicrobial peptides (AMPs) has garnered considerable attention due to their potential in combating antibiotic-resistant pathogens. Mass spectrometry-based proteomics provides valuable information on microbial stress responses induced by AMPs. This work aims to unravel the proteomic alterations induced by the amyloidogenic antimicrobial peptide R23I, encompassing both inhibitory and non-inhibitory concentrations. This study investigates the effects of the R23I peptide on the protein abundance of Thermus thermophilus (T. thermophilus) at different concentrations (20, 50, and 100 μg/mL). We found 82 differentially expressed proteins, including 15 upregulated and 67 downregulated proteins. We also compared the protein identification results between the PEAKS and IdentiPy programs. Our proteomic analysis revealed distinct patterns of protein expression, suggesting compensatory mechanisms in response to the R23I peptide. Notably, the alterations predominantly affected membrane and cytoplasmic proteins that play a central role in critical cellular processes such as transcription, translation, and energy conversion. This study sheds light on the complex interactions between antimicrobial peptides and bacterial responses, offering insights into microbial adaptability and potential implications for antimicrobial strategies and the understanding of microbial physiology.

1. Introduction

In recent years, the attention of many researchers has been directed to the study of the structural features and mechanisms of action of antibacterial peptides [1,2,3,4]. This trend is due to the relevance of finding of new candidate molecules against the increasing number of pathogenic bacterial strains resistant to the action of traditional antibiotics [5,6,7]. Antibacterial peptides belong to a diverse group of antimicrobial peptides (AMPs), which can be divided into natural and synthetic [8]. Natural AMPs are produced in the host’s body constantly or in response to infection with a pathogen, while synthetic AMPs can be developed with the required properties [9,10]. The synthetic peptides and chemically modified natural AMPs can have new mechanisms of antimicrobial action, and their modification can enhance their antimicrobial effect as well as delay the resistance of pathogens to them [11,12]. The antimicrobial properties of such synthetic AMPs can be determined by their effects on the cellular processes of model bacteria [13,14]. Proteomic studies of bacterial cells provide valuable information about the proteins expressed in response to the extracellular and intracellular activity of AMPs [15,16].

Label-free quantitative proteomics, based on the tandem mass spectrometry of digested proteins, is a powerful tool for the analysis and quantification of complex protein mixtures [17]. It is used for the analysis of both prokaryotic proteomes [18] and proteomes from eukaryotic tissues [19,20] or bodily fluids [21]. Modern quantitative approaches provide data for hundreds of proteins [22] and are relatively cheap [19].

The R23I peptide is the first artificial hybrid AMP developed on the combination of a cell-penetrating peptide sequence and the amyloidogenic region of the S1 ribosomal protein from Thermus thermophilus [23]. AMP exhibits both amyloidogenic and antimicrobial properties; in particular, R23I is able to co-aggregate with the S1 ribosomal protein from T. thermophilus and cause inhibition of bacterial growth in cell culture. The S1 ribosomal protein is the largest protein of the 30S ribosomal subunit and is associated with many important cellular processes, including the biosynthesis of other proteins [24,25]. In addition, the S1 protein is prone to aggregation, which is associated with the presence of amyloidogenic regions [26,27]. Such amyloidogenic regions can be in use as a basis for the synthesis of amyloid-antibacterial peptides [28]. That is why this protein was chosen as a target for the action of the synthetic amyloid-antimicrobial peptide R23I [29]. Antimicrobial peptides are a diverse group of molecules. Their mechanism of action is aimed at destroying the cell membrane or against intracellular targets. Ribosomes are one of the key targets, and interaction with them can lead to antimicrobial effects [30,31]. A number of works have been published describing the effect of antimicrobial peptides on the functional centers of the ribosome [32,33,34]. There are several known strategies for interrupting protein biosynthesis by antimicrobial peptides: disrupting the translation process by erroneous insertions of amino acids, preventing (blocking) the release of the newly synthesized polypeptide chain, and disrupting protein folding [35,36,37]. Investigating the mechanisms of action of AMPs is important for the discovery of new peptides, the development of new, more effective antibiotics, as well as for studying the regulation of the translation process. The mechanisms of the AMPs’ action can also be studied using T. thermophilus [38].

T. thermophilus is a model organism and is widely used in the field of structural biology [39,40]. On the one hand, three-dimensional models of T. thermophilus proteins play a significant role in solving problems in this area to explain decoding mechanisms [41]. The emergence of atomic structures 30S and 50S from T. thermophilus and other organisms made it possible to summarize decades of research on the translation apparatus [39,42,43]. On the other hand, thermophilic and mesophilic species have similar mechanisms of adaptations to the action of antibiotic compounds [44,45].

The coincubation of a bacterial cell with an antimicrobial peptide causes a cellular response to the action of the drug, which is demonstrated in suppressing the growth and changing the vital activity of the cell [46]. Cells often react with changes in the quantitative and qualitative composition of synthesized proteins in response to stress caused by the action of an antimicrobial peptide [47,48]. Differential protein expression analysis is a method that allows the identification of the differences between the proteins synthesized by the cell in the control and in the experiment with the tested effector on protein expression [49,50]. HPLC-MS is a technique that is widely used for the proteomics of complex mixtures of proteins and peptides. Of course, studies of the transcriptome can also provide interesting quantitative information about changes in cellular processes under the influence of the effector [51]. However, as noted in the works [52,53], it is often difficult to establish a correlation between the level of expression of mRNA and protein in the cell. In addition, due to the fact that the considered active antimicrobial substance R23I is a peptide acting according to the proposed mechanism based on directed peptide–protein aggregation, we focused directly on the analysis of the proteins and not the RNA. The peptide and proteins, the synthesis of which is indirectly influenced by this peptide, can be analogs of the effectors and components of the bacterial response to the antimicrobial peptide [54,55,56,57].

In our earlier work, we described the change in the growth of the bacterial culture of T. thermophilus, its proteome, and cell morphology under the action of the R23I peptide [23]. In this article, we focus not only on the inhibitory concentrations of the peptide but also on the concentrations that did not impact on the grow bacteria. R23I is an amyloidogenic peptide synthesized based on the S1 ribosomal protein sequence. We assume that the peptide with an amyloidogenic ability interacts with a cellular target by the mechanism of directed coaggregation (Figure 1). This hypothetical mechanism was described earlier [58]. At the moment, the reason for the death of bacterial cells treated with the R23I peptide remains unclear. The study of the cell proteome using mass spectrometry can expand our understanding of the bacterial stress responses caused by AMP activity. Thus, the aim of this work was to identify proteins whose expression is altered by the action of the amyloid-antimicrobial peptide of inhibitory and non-inhibitory concentrations.

2. Results and Discussion

2.1. Changes in Protein Profiles after R23I Peptide Treatment

The R23I peptide inhibits the growth of T. thermophilus cell culture at concentrations above 40 μg/mL, as shown in Figure 2, with a concentration of 20 μg/mL being the highest concentration that does not inhibit the bacterial cells’ growth.

T. thermophilus bacteria were treated with the amyloidogenic antibacterial peptide R23I during one day at one subinhibitory and two inhibitory dosages (20, 50, and 100 μg/mL, respectively). The resulting cells were disrupted, and proteins were isolated from them and then analyzed by chromatography–mass spectrometry. We identified a total of 269,814 peptides across the three dosages and control. Of the identified peptides, 7279 were unique, resulting in 1070 proteins identified. That is, approximately half of the proteins known from the UniProt database were identified. Moreover, this result is slightly better than that obtained using MALDI-TOF-MS [59]. Furthermore, 291 proteins were present in all samples and were used for principal component analysis (PCA) (Figure 3A), which showed that the treatment groups were well separated from the controls. Regarding the dosages, the proteins with 20 μg/mL dosage formed a distinct cluster in the PCA plot, while the proteins with 50 and 100 μg/mL dosage seemed to form one group, suggesting that the protein abundance differed between the subinhibitory and inhibitory conditions.

Changes in the proteome were identified by calculating the fold change (FC) in protein abundance relative to the control condition. The significance of each observed fold change was determined by calculating p values. An FC  ≥  1.2 and an adjusted p value ≤ 0.05 were defined as the biological and statistical cutoff values, respectively. As the result, 82 proteins had significantly different abundances: 15 were upregulated and 67 were downregulated (Figure 3B). The protein with the highest FC was Q72JI4, an aminopeptidase involved in catalytic activity (see Supplementary File S1). The protein with the lowest FC had Q72J15, a leucine-, isoleucine-, valine-, threonine-, and alanine-binding protein involved in amino acid transport. Strikingly, a quarter of the revealed proteins (19 out of 82) were 30S and 50S ribosomal proteins, and all of them were downregulated according to the volcano plot ({S2 S4 S9 S11 S12 S15 S16 S18 S19} and {L2 L5 L13 L14 L15 L22 L23 L24 L27 L29} (Figure 3B and Figure 4).

At the same time, the proteins S2, S4, and S15 belong to the protein regulators of the small ribosomal subunit [24]. The S2 operon (rpsB–tsf operon) of the bacteria encodes the elongation factor Ts and S2 protein, which was also downregulated in our case. S4, together with protein S7, initiates the assembly of the 30S ribosomal subunit [60]. Thus, our data align with the extensive research on the effects of ribosomal protein S1 inactivation and its consequences [24]. Unlike the work of Sørensen et al. [24], in which this effect was studied by deleting the rpsA gene encoding S1 in E. coli, our study inactivated the S1 protein through its targeted coaggregation with the R23I peptide.

An interesting pattern was observed in the comparison between groups (Figure 3C). Some proteins were upregulated in the dose 20 group, while they were downregulated in the dose 50 μg/mL and 100 μg/mL groups, in particular, heat shock protein, oxidoreductase, and chaperone protein, suggesting that, at a subinhibitory dose, the bacteria could mobilize resources to overcome the damage and that, at inhibitory doses, these resources were depleted.

As demonstrated in Figure 3A, the samples revealed distinct clustering of the control and 20 μg/mL groups, while the 50 and 100 μg/mL groups clustered together. This suggests that the proteomic response differs between the subinhibitory and inhibitory concentrations of the peptide. As a rule, the subinhibitory concentrations of stress factors induce changes in specific regulatory proteins without clearly affecting the overall protein levels [61]. Additionally, the volcano plot in Figure 3B further emphasizes proteins with statistically significant differential abundance between the control and treatment groups. This analysis allows us to pinpoint which proteins are most affected by the peptide treatment, either by being upregulated or downregulated. Figure 3B,C present 82 differentially regulated and statistically significant proteins. Thus, out of the 1070 identified proteins, these 82 were selected for further analysis. Previous studies have also reported a relatively small number of proteins, among the total identified, that were differentially regulated under the influence of antimicrobial peptides [62,63]. This can be explained by the fact that antimicrobial peptides do not affect all cellular processes but act through specific mechanisms, typically targeting particular cellular components and functions [64,65].

2.2. Comparison of Proteomic Analysis Results Obtained Using the Proteomics Tools PEAKS and IdentiPy

The results of the presented proteomic analysis can be compared with the previously published results on the number of proteins [23]. Both programs found approximately 250–300 proteins common to the four groups (PEAKS—291 proteins, IdentiPy—257), with an intersection of 155 proteins, which is more than half of each set. However, if we examine the proteins that were identified as significantly different by groups as a result of statistical processing (Figure 5A), we see that the number of proteins common to the two programs was only 33. This is less than half of the protein sets identified as significantly different (PEAKS—82 proteins, IdentiPy—91). This suggests that the results of chromatogram processing by the search program can significantly affect the results of further analysis.

It should be noted that, in the previous work, mass spectrometry data analysis was conducted using earlier versions of the software, namely Peaks Studio 7.5 and IdentiPy [66], whereas, in the present study, PEAKS Studio 11 was used for these purposes. Additionally, integration with deep learning-based technology allows PEAKS Studio 11 to utilize the predicted retention time as another validation level to resolve ambiguity between peptides with similar ion matching scores [67]. Differences between software programs lead to variations in the identified proteins. Specifically, PEAKS Studio 11, utilized in this study, identified a greater number of peptides and consequently detected more proteins than the 2018 version of IdentiPy used in the previous study. Thus, the results obtained in this study facilitate the comparison of different software tools for mass spectrometry data analysis.

2.3. The Analysis and Description of Changes in the Relative Amount of Proteins in Terms of Gene Ontology

In addition to the targeted analysis of differentially expressed proteins, gene ontology (GO) term overrepresentation and pathway enrichment analyses were performed to identify changes in biological functions (Figure 6). The GO location analysis pointed out that most proteins were decreased in the ribosome, followed by the cytoplasm and membrane, while an increased level was equally seen in the cytoplasm and membrane, although in lesser number (Figure 6A). The GO function demonstrated that proteins with decreased levels were mainly associated with the structural constituent of the ribosome, rRNA and nucleotide binding, and oxidoreductase activity; proteins with increased levels were mainly associated with metal ion binding and transferase activity (Figure 6B). Of interest, few proteins involved in lipid binding (degV protein, Q72K34) and the sulfur (sulfur carrier protein TtuD, Q72JV2) pathway were upregulated. The GO processes showed that proteins with decreased levels were mainly associated with translation, tricarboxylic acid cycle, proteolysis, and transport (Figure 6C). Despite the fact that increased proteins were also involved in proteolysis and transport, some unique terms were observed, namely, cell wall organization and lipid A and peptidoglycan biosynthetic processes. It appears that, in a state of severely compromised protein synthesis, the bacteria made an effort to allocate their remaining resources to maintaining the cell wall. We could speculate that R23I has some impact not only on ribosomes but on the membrane as well.

Some AMPs (for example, human antibacterial peptide LL-37) have been shown to induce oxidative stress in Escherichia coli, disrupting the flow of electrons through the electron transport chain [54]. For this relative quantitative indicator, a decrease in the content of the S-layer membrane protein was noted. Of note, Kirk et al. recently identified Clostridium difficile strains that were not susceptible to bacteriocins (a class of AMPs) due to the lack of a surface S-layer [55]. The decrease in the concentration of the S-layer protein of T. thermophilus is directly related to the adaptation mechanism to the action of the R23I peptide. Moreover, with an increase in the concentration of the R23I peptide, a decrease in the relative concentration of the S-layer protein was observed. Notably, the S-layer protein has important roles in survival, adhesion, and pathogenesis [68]. The protein is able to perform protein secretion and antibiotic export [69]. However, other authors show that S-layer proteins combined with antibiotics can inhibit bacterial growth [70]. It remains unclear whether the decrease in the content of S-layer proteins is associated with the antimicrobial effect of R23I. Similar dynamics were observed for other membrane proteins (Figure 7).

Protection of the bacterial population from the limiting factor through the mechanism of persistent formation. It is known that persisters have a low level of expression of tricarboxylic acid cycle genes [71]. In our work, we also described a decrease in the expression of genes of the tricarboxylic acid cycle, including malate dehydrogenase and oxoglutarate dehydrogenase (succinyl-transferring). This decrease may indicate an imbalance in energy processes, which could lead to cell death.

It was shown that the regulation of gene expression can be carried out both by the holo-enzyme and by subunits [72]. Mutations of the bacterial RNA polymerase can lead to changes in the physiology and virulence of bacterial cells [73]. We also observed changes in the profiles of ribosomal protein abundance. Probably, the changes are the result of the R23I peptide interaction through the mechanism of directed coaggregation.

Changes in the abundance of enzymes in the Krebs cycle (fumarate hydratase class II (Fumarase C) and succinate—CoA ligase [ADP-forming] subunit beta; oxoglutarate dehydrogenase (succinyl-transferring) and isocitrate dehydrogenase (NADP (+)); malate synthase; succinate dehydrogenase iron–sulfur subunit, citrate synthase, aconitate hydratase (aconitase) and succinate dehydrogenase flavoprotein subunit) can be the result of “tuning” the catabolic pathway in response to increased energy demands [71].

In our prior research, we demonstrated that peptide R23I has the potential to induce substantial changes in the proteome, influencing a multitude of proteins crucial to essential cellular functions [23]. Notably, proteins such as cytochrome c oxidase subunit 2, isocitrate dehydrogenase (NADP), and NADH-quinone oxidoreductase subunit 2 were affected, suggesting a decrease in ATP and NADP synthesis. This observation provides support for the theory proposing a decline in metabolic activity in response to peptide-induced stress.

Moving forward, we plan to investigate the time-dependent effects of the R23I peptide and its analogs [28,74] on the proteomes of pathogenic bacteria such as Staphylococcus aureus and Pseudomonas aeruginosa. This further research aims to enhance our understanding of how antimicrobial peptides interact with pathogenic bacterial cells and how they influence proteome changes. Studying changes in the bacterial proteome after different incubation periods with the peptide will help determine whether the peptide’s effects are immediate and direct or delayed and indirect.

3. Materials and Methods

3.1. Experimental Design and Sample Preparation for MS/MS Analysis

In previous studies, we found that the R23I peptide exhibits antibacterial activity in the concentration range of 50–500 μg/mL [23]. In this work, we used concentrations of 20, 50, and 100 μg/mL. Experimental samples were taken in three technical repetitions, and the control samples (T. thermophilus cells without R23I peptide) underwent two technical repetitions.

The cells of the control sample and with the addition of different concentrations of the peptide were incubated for 24 h at the temperature of 70 °C in 1.5 mL tubes in an ES-20/60 shaker-incubator (Biosan, Rīga, Latvia). The optical density of the cell culture was measured at 600 nm using a spectrophotometer SF-102 (Akvilon, Moscow, Russia). The cell suspension was destroyed by freezing and thawing for 5 cycles. Cell lysates were concentrated using an Eppendorf 5301 vacuum concentrator (Eppendorf, Hamburg, Germany). Then, the electrophoretic separation of proteins in PAGE was carried out according to the Laemmli method [75]. A sample buffer was added to homogenized samples and loaded onto a 12% polyacrylamide gel in the presence of SDS. The buffer consisted of 1.25 mL 0.5 M Tris-HCl, 2.5 mL 50% glycerol, 1 mL 10% SDS, 0.2 mL 0.5% bromophenol blue, and 3.55 mL deionized H₂O. The separation was carried out in two modes. First at 80–90 V before entering the separating gel, then at 180 V. The standard procedure for staining and washing the gels from Coomassie was used.

In-gel digestion was carried out according to the standard procedure [76]. High-molecular-weight, medium-molecular-weight, and low-molecular-weight fractions were cut out of the gel with a scalpel and placed into 1.5 mL tubes. To remove SDS, 1 mL of a mixture of 40% methanol and 5% acetic acid was added in tubes and then incubated at room temperature for 15 min. The gel was washed from the dye with 1 mL of a solution of 50% acetonitrile in 50 mM NH₄HCO₃ (pH 8.0) for 20 min at 56 °C. To reduce disulfide bonds, the samples were incubated for 30 min at 36 °C with 50 μL of a 5 mM solution of β-mercaptoethanol. For subsequent alkylation, the samples were incubated with 50 μL of a 15 mM iodoacetamide solution for 30 min in the dark at room temperature. After each step, the gel fragments were washed with deionized water. Dehydration of the samples was carried out by adding 1 mL of acetonitrile for 10 min, followed by vacuum concentration.

Enzymatic hydrolysis was carried out by incubating the sample with a trypsin solution (10 μg/mL trypsin in 200 μL 50 mM NH₄HCO₃) for 10 min at 4 °C. Then, the solution was removed, and 400 μL of 50 mM NH₄HCO₃ was added. The samples were incubated in a thermal shaker for 20 h at 37 °C, and then the solution was transferred into clean tubes and dried on a vacuum concentrator. Subsequently, 100 μL of a mixture of 80% acetonitrile and 0.1% TFA was added to the gel for the additional extraction of peptides and incubated for 30 min at 25 °C. The protein hydrolysate was dried for analysis. Before mass spectrometric analysis, the peptide fractions were pooled and then purified using a ZipTip Empore octadecyl C18.

3.2. Liquid Chromatography–Mass Spectrometry (LC-ESI-Orbitrap)

Chromatography–mass spectrometry (analysis) was carried out on an Orbitrap Elite ETD mass spectrometer (Thermo Scientific, Waltham, MA, USA) coupled with an EASY nLC 1000 nano-flow chromatograph (Thermo Scientific, Waltham, MA, USA).

The mixture of peptides was separated on C18 reversed-phase columns packed in laboratory conditions (Phenomenex, Torrance, CA, USA) with the following parameters: pore size 100 Å, particle size 1.7 μm, column length 150–200 mm, and column inner diameter 100 μm. For the concentration and additional purification of the samples, we used a Precolumn C18 PepMap 100 (Thermo Scientific, Waltham, MA, USA) with the following parameters: pore size 100 Å, particle size 5 µm, column length 5 mm, and column inner diameter 300 µm.

The elution of peptides was carried out in a concentration gradient of acetonitrile from 5% to 50% at a flow rate of 300 nL/min for 240 min. The loaded volume (of the sample) was 1–10 μL. The ionization of the sample molecules was carried out by the nano-electrospray method. The voltage applied to the emitter was varied from 1800 to 2000 V. The temperature at the inlet capillary was 200 °C. The fragmentation of the selected ions was activated by collisions with molecules of an inert gas (helium) in a high-energy chamber (HCD method). The selection of ions for fragmentation was carried out automatically.

3.3. Data Analysis

The raw mass spectrometry data were mapped to the proteome database from Thermus thermophilus (Uniport, strain ATCC BAA-163/DSM 7039/HB27, Tax ID: 262724, May 2023) using PEAKS Studio 11 (Bioinformatics Solutions Inc., Waterloo, ON, Canada), allowing for 20 ppm parent ion and 0.02 m/z fragment ion mass error, 2 missed cleavages, carbamidomethylation, and oxidation (M) as fixed modifications. To limit false-positive peptide identification, a 1% false discovery rate (FDR) was applied to peptide spectrum matches (PSMs), and subsequently, the identified peptides were filtered with ≥2 unique peptides in all biological replicates. Label-free quantitative analysis was carried out on the identified peptides by using such a module in the PEAKS Studio 11 (Bioinformatics Solutions Inc., Waterloo, ON, Canada). The peak areas were normalized to the total ion count (TIC) of the respective analysis run before performing pairwise comparison between the groups. The protein abundance ratio between the control and treatment conditions was filtered with fold change ratios of ≥1.2, and analysis of variance (ANOVA) was performed to test the statistical significance of the observed abundance changes, with p values below 0.05 considered to be statistically significant. The mean protein coverage for each condition was used for the heatmap of ribosomal proteins. The Z-scores were calculated for the rows, and the hierarchically clustered graph was built in the Python 3.10.14 Seaborn module. The heatmaps in Figure 3C and Figure 6 and the Venn diagrams in Figure 7 were visualized using the Pandas, Matplotlib, and Seaborn libraries of Python, based on data from PEAKS Studio 11. The log2 ratio values were calculated from the ratios of the abundance of each group relative to the average protein abundance across all groups. The gene ontology (GO) term overrepresentation and pathway enrichment analyses of differentially expressed proteins was based on annotation taken from the GO webpage (https://www.ebi.ac.uk/QuickGO/ (accessed on 13 September 2024)).

4. Conclusions

The protein abundance profile was shown to vary with the concentration of the R23I peptide. In total, 82 differentially expressed proteins were identified compared with controls and other samples (as determined by the volcano plot), of which 15 were upregulated and 67 were downregulated. The dynamics of protein expression is probably associated with compensatory changes due to the action of the peptide. Changes relative to the quantity of proteins under the action of R23I at a concentration of 20, 50, and 100 μg/mL are related to the membrane and cytoplasmic proteins that are associated with processes such as transcription, translation, and energy conversion. It has been demonstrated that the cell response of T. thermophilus to the action of the R23I peptide can be correlated with that of other pathogenic and non-pathogenic model organisms.