Identification of Bacteria Associated with Post-Operative Wounds of Patients with the Use of Matrix-Assisted Laser Desorption/Ionization Time-of-Flight Mass Spectrometry Approach

The bacterial infection of post-operative wounds is a common health problem. Therefore, it is important to investigate fast and accurate methods of identifying bacteria in clinical samples. The aim of the study was to analyse the use of the MALDI-TOF MS technique to identify microorganism wounds that are difficult to heal. The most common bacteria are Escherichia coli, Staphylococcus spp., and Enterococcus spp. We also demonstrate the effect of culture conditions, such as the used growth medium (solid: Brain Heart Infusion Agar, Mueller Hilton Agar, Glucose Bromocresol Purple Agar, and Vancomycin Resistance Enterococci Agar Base and liquid: Tryptic Soy Broth and BACTEC Lytic/10 Anaerobic/F), the incubation time (4, 6, and 24h), and the method of the preparation of bacterial protein extracts (the standard method based on the Bruker guideline, the Sepsityper method) to identify factors and the quality of the obtained mass spectra. By comparing the protein profiles of bacteria from patients not treated with antibiotics to those treated with antibiotics based on the presence/absence of specific signals and using the UniProt platform, it was possible to predict the probable mechanism of the action of the antibiotic used and the mechanism of drug resistance.


Introduction
Quick and reliable methods are needed to identify bacteria from post-operative wounds. The rapid identification of the pathogen causing the infection will enable to implement of an appropriate therapy [1]. Currently, the matrix-assisted laser desorption ionization time-of-flight mass spectrometry technique (MALDI-TOF MS) is used with great success [2]. The MALDI-TOF MS technique allows for quick and accurate identification of clinically relevant microorganisms [3,4].
Post-operative wound infection, also known as surgical site infections (SSIs), is a common healthcare problem. It is estimated that 2-7% of all operations develop SSIs, and 11% of all deaths in intensive care patients are related to post-operative wound infections [5].
Research confirms the incidence of SSIs (2.1%) and shows that infections occur even when surgeons take aseptic precautions during surgery and patients are treated before and after surgery [6]. Infections are associated with the bacterial contamination of the wound during surgery or later due to the deterioration of skin barriers during wound care [7]. actions on the 50S subunit of the bacterial ribosome, possibly by influencing the initiation of the peptide chain. The mechanisms of resistance are similar to those of quinolone resistance and are characteristic of, among others, Enterococcus spp., Streptococcus spp., and Staphylococcus spp. [27].
The aim of the research was a new approach to development a spectrometric method to identify bacterial strains isolated from post-operative wounds of patients undergoing antibiotic therapy using the MALDI-TOF MS technique and the influence of culture conditions on the level of the identification and quality of mass spectra. Moreover, generated proteins profiles of strains derived from control patients (no antibiotic treatment) and those subjected to antibiotic therapy were compared based on the presence/absence of specific signals of bacterial molecular profile.
Then, the signals were analysed following their name and function by using the UniProt database as a pilot study. Utilization of the UniProt database was performed as an attempts to indicate the protein complexes responsible for molecular changes in bacterial profiles-untargeted analysis. Due to the use of MALDI-TOF MS technique and the UniProt database, we compared the obtained signals and, on this basis, it was possible to investigate whether there are factors contributing to the development of antibiotic resistance in isolates from patients undergoing antibiotic therapy.

Bacterial Identification
The sample preparation procedure and the MALDI-TOF MS analysis allowed us to obtain MS spectra and identify bacterial strains on post-operative wounds of patients. Bacterial strains were analysed in single-spectrum and MSP (Main spectra profiles) modes using the MALDI Biotyper 3.0 platform. Identification results and score values for both RAW and MSP spectra for each medium are summarized in Table 1.
The use of four culture media allowed for greater microbiological diversity. Due to this, it was possible to analyse the number of not only the species but also the strains colonizing difficult post-operative wounds.
Identifying the most significant number of bacterial species was possible through the use of non-selective media: BHI and MH, which allow the isolation of a wide variety of microorganisms. These media have been successfully used to isolate bacteria from clinical specimens [28,29]. Compared to other culture media, BHI was the only one that enabled the isolation of Lactobacillus pentosus and Providenciastuartii. Fewer bacterial species were identified on the selective BCP medium to isolate Enterobacteriaceae bacteria [30]. The BCP medium was the only one that allowed Enterobacterkobei to be detected. The smallest number of bacterial species involved in post-operative wound infections was isolated on the VRE medium, a selective medium for the culture of vancomycin-resistant enterococci, such as E. faecalisand E. faecium [31].
In summary, we obtained 75 different bacterial isolates, among which 71 were identified. Bacteria from 50 samples were identified with species confidence and 21 with genus confidence. The use of four different microbiological media allowed us to identify 34 strains belonging to 15 species and nine types of bacteria ( Table 2). The highest percentage concerned bacteria of the genus Enterococcus (26.5%) and Staphylococcus (23.5%). E. coli accounted for 11% of all the identified bacteria. Each of the other types of bacteria accounted for <10% of all identified microorganisms.  coli and S. aureus in post-operative wound infections; they also indicate Staphylococcus, Enterococcus, and Streptococcus bacteria. The research shows that P. aeruginosa also plays a role in surgical site infections [32,33]. The MALDI-TOF MS analysis did not allow the identification of three bacterial isolates (score < 1.7). This indicates the need to extend the currently available databases, also indicated by the literature data [34]. The insufficient number of mass spectra in the base of testers may also cause incorrect identification, which was proven by Body et al. [35].
As for bacterial species commonly found in clinical samples, the available databases are robust and sufficient to conclude that they are enough for the routine identification in clinical microbiology [36]. On the other hand, known human pathogens represent only 1% of the population of all microbial species in the world; therefore, the databases will not be fully complete [37]. Extending the databases will enable more accurate identification of microorganisms. For example, a group previously thought to be a single Candida haemulonii is now identified as a complex of species, three of which are regularly detected in clinical trials [38]. A similar situation applies to bacteria of the Klebsiella genus [39]. The newly discovered species belonging to these types of microorganisms appear to be of clinical importance. Updating the databases can solve the problem of misidentification or nonidentification [40].
From the swabs taken from a patient treated with three antibiotics (B21) and two antibiotics, ciprofloxacin combined with metronidazole (B12), it was impossible to isolate the bacteria from any of the solid media used, which may indicate that the antibiotic therapy worked effectively. The presence of Bacillus spp. in the patient sample B21 may have been caused by the contamination of the sample.
The type of swabs used was also taken into account. It can be concluded that the analysis performed for liquid and solid swabs gave a similar number of identified bacteria at the species level.

Mass Spectra Analysis
The analysis of mass spectra obtained for S. aureus bacteria isolated from each of the four culture media (Figure 1), for which score >2 showed that a different quality of the spectrum was obtained depending on the medium used. The highest identification score value was obtained for the bacteria isolated from the VRE medium (2.47) and BHI (2.35). The medium from which the bacteria giving the lowest identification scores were isolated was the selective BCP medium (2.03).
A salt blend added to the VRE medium and pigments in the BCP medium can reduce the yield or inhibit the analyte is ionized by blocking/suppressing the ionization process [41,42]. On the other hand, Złoch et al. investigated the effect of various solid culture media on the identification of S. aureus [43]. Their results showed that each medium allowed for the correct identification of microorganisms, and the differences between the score values were minimal.
Phyloproteomic relationships between isolates are presented in MSP dendrograms ( Figure 2). All the phyloproteomic trees proved the relationship between the bacterial strains isolated from the biological material and the reference strains present in the database. Proteomic trees were made in order to confirm the correct identification of the isolated bacterial strain. Their analysis showed that five groups of bacteria were identified on the BHI medium, four groups on the BCP, five groups on the VRE, and eight groups on the MH. On each of the dendrograms of bacteria evolved from the BHI, BCP, and MH media, bacteria of the genus Escherichia, Klebsiella, and Enterobacter were grouped into one cluster may prove their close relationship.   The listed species of bacteria belong to the large Enterobacteriaceae family, and the results from the phyloproteomic tree correlate with their genetic background. This is confirmed, among other things, by the studies by Paradis et al. [44]. No close relationship was found for the bacteria of the genus Staphylococcus triggered on the MH medium as they were divided into three groups based on species (S. haemolyticus, S. aureus, and S. epidermidis). The analysis of phyloproteomic trees made for bacteria evolved from the remaining growth media (BHI, BCP, and VRE) allows us to conclude that there is a close relationship between these strains.
The analysis of the phylogenetic trees by Złoch et al. confirmed the relative origin of the Staphylococcus species [2]. The analysis of the VRE dendrogram shows that Enterococcus bacteria were divided into two groups (E. Faecalis and E. faecium), and the genetic relationships presented by analogues phylogenetic trees were confirmed by the data obtained from the analysis of phyloproteomic trees [45]. Table 3 includes the specific peaks identified by the MALDI-TOF MS and their characteristics from the UniProt database. Moreover, the table does not include signals from uncharacterized proteins or proteins that are not associated with the action of antibiotics or antibiotic resistance. The values of m/z from the control sample and the test sample were compared. The samples tested were protein profiles obtained for S. aureus (patient treated with clindamycin), E. cloacae (patient treated with clindamycin) and E. faecalis (piperacillin combined with tazobactam). 10,883.9 -+ Ribosome hibernation promoting factor HPF primary metabolic process, regulation of translation [64] The control and test sample columns are highlighted as green and orange, respectively, to differentiate between them.

Identification of Proteins with Specific Peak to the Predicted Antibiotic Mechanism of Action and Bacterial Drug-Resistance
Eighteen specific signals related to virulence, drug resistance, or the administered antibiotic's mechanisms were identified for S. aureus /ATCC 29213 THL/. Among the signals appearing in the test sample, phenol-soluble modulin PSM-α-3 (m/z = 2656.8) can be distinguished. The studies of Queck et al. and Chatterjee et al. demonstrated that this protein is a rare example of combined antibiotic resistance and virulence factors in MRSA [46,47]. The increased PSM expression was shown to cause the increased virulence of the MRSA (methicillin-resistant S. aureus) and MDR (multidrug-resistant) S. aureus strains and contribute to the infection's severity [65,66].
In addition, it was shown that PSM proteins are involved in the structuring of the bacterial biofilm of S. aureus, which entails the spread of infection and increases the resistance to antibiotics [67]. Generally, PSM toxin is a major virulence factor of S. aureus and primarily causes cytolysis in red and white blood cells. Among α-type phenol soluble modulins, the PSM-α3 plays the most prominent role in S. aureus virulence and, moreover, is the most cytotoxic member of the family [67].
The presence of the TraR/DksA family transcriptional regulation (m/z = 3006.2) on bacterial conjugation plasmids suggests their role in pathogenicity, virulence, and antibiotic resistance [48]. In the case of bacteria isolated from control group patients, the presence of antibiotic transport proteins (characteristic for the signal m/z = 6355.1), uncharacterized in the database protein synthesized in response to antibiotic use (m/z = 6589.8) and exported protein (m/z = 6743.1) was also demonstrated. These proteins can contribute to the development of antibiotic resistance [53].
Characteristic signals for immunoglobulin G building protein A (m/z = 9898.5) and protein A (m/z = 9909.4) being a virulence factor, especially in MRSA strains [55], have been reported. In addition, a signal characteristic of GCN5-related N-acetyltransferase (GNAT) was also identified, the presence of which was first confirmed in kanamycin and gentamycin resistant bacteria [54]. This result may indicate resistance to more than one antibacterial agent in the analysed S. aureus strain.
Compared with the control group, the test sample is mainly the signals characterized as 50S ribosomal protein L36 (m/z = 4290.7) and L33 (m/z = 5437.4 and 5872.7). The inhibition of the production of ribosomal proteins, thus, prevents the translation and biosynthesis of bacterial proteins [49]. Signals from various ABC transport proteins were also identified (m/z = 4365.1, 4379.9, 4822.6, 7022.9, and 8908.0). These proteins may affect the virulency of pathogenic strains and may be the target of antibiotics [50].
The lack of proteins from the ABC family will likely impair the import of many substances important for the life of bacterial cells and, thus, negatively affect their functioning [51]. Single-stranded DNA binding proteins (m/z = 4386.1) (ssDNA) (SSB), the signal of which does not appear only in the control sample, are necessary for the DNA replication and the protection of the delayed strand against the nuclease attack [52].
The patient who was identified with the presence of E. cloacae/MB_5277_05 THL/was treated with clindamycin. After using the antibiotic, a signal appeared corresponding to the Agmatinase protein (m/z = 3710.7; 5598.8), catalysing the reaction of agmatine hydrolysis to putrescine and urea. McCurtain et al. proved that the accumulation of agmatinase could result in the increased antibiotic tolerance by bacteria [58].
We found that the MS spectrum of bacteria after the administration of the antibiotic showed a signal from murein transglycosylase (m/z = 6417.9), responsible for breaking the glycosidic bond and for the catalysis of intramolecular and glycosylation [60]. These reactions can disintegrate the bacterial cell wall. On the other hand, the appearance of the BcsF cellulose biosynthesis protein was discovered (m/z = 7620.8), which is involved in the formation of a bacterial biofilm and, thus, increased the resistance to antibiotics [61].
Additionally, it can be assumed that the use of clindamycin may affect damage to the genetic material, because the mass spectrum showed a signal from the DNA polymerase V (m/z = 9410.181), which is part of the SOS response to the DNA damage [62]. The antibiotic administration likely caused dysregulation of the DNA transcription, which pretended the ParB partition protein signal (m/z = 10,315.8) and translation-the appearance of a factor favouring the hibernation of ribosomes HPF (m/z = 10,883.9) [63,64,68]. After the antibiotic treatment, the signal disappearance at m/z = 5571.8 for dehydrogenase (NADP (+)) was visible.
This protein is responsible, inter alia, for the binding of Mg 2+ and Mn 2+ ions; it has the activity of isocitrate dehydrogenase (NADP (+)) and takes part in the citric acid cycle [59]. On this basis, we concluded that this protein plays several important physiological roles, inter alia, in the reductive fatty acid biosynthesis, proper oxidation-reduction balance as well as in oxidative damage. Moreover, the lack of it in a bacterial cell may result in a disturbance of the cell metabolism and may have a negative impact on the integrity of cell membranes [59].
Patients with identified E. faecalis/ATCC 7080 THL/were treated with piperacillin combined with tazobactam. Both of the used drugs belong to the group of β-lactam antibiotics that inhibit the biosynthesis of the bacterial cell wall. A significant change in the mass spectrum of bacteria from the control sample, compared to the test group, is the appearance of the attenuator leader peptide (m/z = 2443.3) in the latter, which is a response to the action of the applied antibiotics [56]. The disappearance of the signal characteristic of the 50S ribosomal protein L30 (m/z = 6341.0) was demonstrated. This can result in disturbances in translation and protein synthesis [57].
The conducted pilot research indicates the possibility of using the UniProt database to the preliminary match of the proteins contained therein with the spectra generated by the MS, including predicting the mechanisms of the action of antibiotics and the mechanisms of the antibiotic resistance of bacteria. Notably, the products of antibiotic metabolism can also be biologically active substances [69]. To confirm the obtained results, it is necessary to perform additional analysis by the utilisation of, e.g., shotgun techniques, such as liquid chromatography with tandem mass spectrometry (LC-MS/MS) or combined sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE) and MALDI-TOF/TOF MS methods [70,71]. Table 4 compares the obtained identification indicators of the tested bacterial species in terms of the liquid medium used, the method of sample preparation, and the incubation time. Empty spaces in the table mean that it was not possible to obtain mass spectra from these samples. Both methods used allowed for the correct species identification at a similar level. The analysis of the obtained results showed that it was possible to correctly identify E. faecalis after 4 h of incubation using both media and both methods of sample preparation. The use of the same conditions and a 6-h incubation allowed for the identification at the species level of S. aureus. Identification of E. coli was possible only after the application of liquid TSB medium.

Selection of Sample Preparation Conditions to Determine Microorganisms Using the MALDI-TOF MS Technique
The standard method turned out to be a better method of preparing bacterial extracts for MALDI-TOF MS analysis compared with the Sepsityper method. This conclusion confirms Drevinek et al.'s research, which proved that the technique of using ethanol for pre-treatment and formic acid in combination with acetonitrile for the preparation of bacterial extracts showed the highest extraction efficiency and spectral quality. Researchers found this method universal for samples suspected of having Gram (+) or Gram (−) bacteria [72]. The research by Burckhardt et al. proved that 20 h of incubation was sufficient to detect multidrug-resistant bacteria and MRSA strains [73].
Pomastowski et al. showed that shortening the incubation time to 12 h enabled the generation of good-quality spectra and the correct identification of microorganisms. Their research showed that different incubation time intervals affected the quality of the recorded spectra; yet, they did not detect any significant qualitative changes in the peak patternsthey only observed an increase in the intensity (100% with 24-h incubation) [17]. The research results by van den Bijllaardt et al. suggested that 10 h of incubation of bacterial cells during tests related to reading sensitivity to antibiotics was sufficient [74]. Altun et al., on the other hand, obtained results that showed that a 5.5-h incubation of bacteria from blood cultures enabled the satisfactory identification of bacteria using MALDI-TOF MS and the Vitek 2 database [75].
Oviaño et al., in their research, proved that using the MALDI-TOF MS technique and the sonication-based extraction method enabled the precise identification of bacteria in the liquid bacterial culture in 15 min [76]. Haiko et al. used the method of the urine short incubation MALDI-TOF for the rapid identification of urinary pathogens [77]. They correctly identified 86% of G (−) bacteria; however, this method did not identify G (+) bacteria. Due to this approach, it was possible to shorten the analysis time to 4-6 h compared to the 24 h required by conventional methods.
Shortening the analysis time will significantly speed up microbiological diagnostics, which will additionally enable a faster application of the appropriate therapy for saving the patient's health and life. Additionally, Gajdács et al. proved that the MALDI-TOF MS technique compared to conventional biochemical methods showed higher efficiency in identifying anaerobic bacteria, such as Cutibacterium spp., Bacteroides/Parabacteroides spp., and Clostridium spp. in blood samples [78].
The obtained mass spectra were also analysed for every tested bacteria. Below is a comparison of spectra under specific conditions for exemplary bacterial species. The spectra obtained for bacteria isolated from two different media slightly differ from each other. Some of the signals for the TSB medium vary in intensity (Figure 3(A.1,A.2)). Additionally, due to using the BACTEC Lytic/10 Anaerobic/F Vial Culture Media, a more significant number of signals were obtained ( Figure 3B). Using two methods of preparing bacterial extracts, spectra of different quality were obtained. The signals received for samples prepared with the Sepsityper method had a higher intensity than those equipped with the standard method. On the other hand, the second method allowed obtaining more separated signals ( Figure 4A) and more of them ( Figure 4B). The differences may result from the different buffer composition of the analysed sample.

Patient and Sample Characterization
Biological material was the superficial swab from post-operative wounds from fifteen patients with SSIs following abdominal surgery of the Provincial Polyclinic Hospital in Toruń. Thirteen patients were treated with antibiotics administered through an intravenous infusion. Six patients received one antibiotic: cefotaxime (B1), clindamycin (two patient-B2, B17), meropenem (B4), ciprofloxacin (B11), and metronidazole (16), while the remaining patients received a combination of the antibiotic treatment with piperacillin and tazobactam (three patients-B7, B15, and B18), ciprofloxacin and metronidazole (two patients-B8 and B12), penicillin and clindamycin (B9) and clindamycin, metronidazole, and amoxicillin (B21). The other two patients were not treated with an antibiotic (control samples).
Before smearing, the post-operative wounds were washed with a sterile physiological saline solution. Two swabs were then collected from each patient. The first was performed with a dry swab stick, which was then placed in Amies liquid transport medium (DeltaSwab Amies, Barcelona, Spain) (swab sample "b"). The second swab was taken with a swab stick from a tube with Amies solid transport medium (BorPol, Gliwice, Poland) (swab sample "a").
The research was approved by the Bioethics Committee of Nicolaus Copernicus University in Toruń (decision no. 585/2017). All the study participants were informed orally and in writing about the purpose of the study, completed the questionnaire and a declaration of their voluntary consent to participate in the study.

Isolation of Post-Operative Wound Bacteria
Swab stick "a" was placed in 1 mL of the transport medium, and 1 mL of the transport medium was added to the swab sample "b". 100 µL of the obtained suspension was transferred onto the four growth media: two non-selective (BHI, MH) and two selective (VRE, BCP) and incubated for 24 h at 37 • C. After the incubation, to receive pure bacterial colonies, streaking was made and followed by incubation again under the same conditions as previously [79]. Protein extracts were made from pure bacterial colonies according to standard acetonitrile-formic acid extraction protocol [2].

MALDI-TOF MS Analysis
For the MALDI-TOF MS analysis, bacterial protein extracts from the obtained pure bacterial cultures were prepared based on the Bruker guideline with slight modifications (standard method). Approximately10 mg of bacterial colonies were added to 150 µL of ultrapure water and mixed thoroughly. Subsequently, 300 µL of absolute EtOH was added to the resulting suspension, and then thoroughly vortexed and centrifuged (15871 RCF, 5 min).
After that, the supernatant was discarded, and the pellet was dried in a vacuum concentrator at 37 • C for 7-10 min. We added 1-12 µL FA to the dried pellet in proportion to the amount of biological material and carefully mixed by pipetting. Next, the same volume of ACN was added, mixed and centrifuged (15871 RCF, 5 min). Afterwards, 1 µL of each sample was transferred onto MALDI MTP 384 polished steel target sample spot (BrukerDaltonik GmbH, Bremen, Germany) in duplicate.
After air-drying, the sample spots were overlaid with HCCA matrix dissolved in standard solvent (50% water, 47.5% ACN, and 2.5% TFA) at the final concentration of 10 mg/mL. The MALDI target plate with samples was analysed using the ultrafleXtreme MALDI-TOF/TOF mass spectrometer (BrukerDaltonik GmbH, Bremen, Germany) equipped with smart beam-II laser-positive mode according to the procedure described by Pomastowski et al. [80].
The calibration analysis was originally performed with a Bruker Bacterial Test Standard (BTS). The spectra were collected manually using the flexControl software with parameters m/z range: 2000-20,000, acceleration voltage at 25 kV, global attenuator offset at 20%, attenuator offset at 34%, range at 34%, laser power at 40%, and 500 shots in-one-single spectra to frequency 2500. The obtained spectra were smoothed and brought to the baseline.
Validated mass spectra were processed with the use of the software provided by the manufacturer-flexControl and flexAnalysis and, subsequently, used for the bacterial identification via the MALDI Biotyper Compass platform (BrukerDaltonik GmbH, Bremen, Germany) based on both raw spectra (RAW) and Main Spectra (MSP) according to the manufacturer's protocol. Based on the MSPs, phyloproteomic dendrograms were generated to verify the correct identification of the isolated microorganisms by their relationship with the reference strains and the relationship between the isolated strains. The UniProt platform was used to predict the mechanism of the action of the antibiotic administered to the patient on the bacterial cell based on the obtained m/z values.

Selection of Sample Preparation Conditions to Determine Microorganisms with MALDI-TOF MS
Two Gram-positive strains (S. aureusBHI.B3a, E. faecalis VRE.B3a) and one Gramnegative (E. coli BHI.B4a) bacterial strain were selected for further analysis. From one-day cultures, one inoculation loop was transferred to a glass tube with 2 mL peptone water to obtain a bacterial suspension with a density equal to 0.5 McFarland. This process was performed in duplicate. Then, 100 µL of suspension was transferred to 9.9 mL liquid growth media TSB, vortexed and diluted 1:1 (v/v) in TSB. From the second tubes, 100 µL inoculum was transferred to 900 µL peptone water and mixed thoroughly.
Next, 200 µL of suspension was added to the vial with BACTEC Lytic/10 Anaerobic/F culture media using a syringe with a needle and mixed. Both of the bacterial cultures were incubated for 4, 6, and 24 h at 37 • C. After each incubation time, the bacterial extracts were prepared according to the previously described standard method and using the MBT Sepsityper IVD Kit according to the protocol provided by the manufacturer. We added 200 µL of Lysis Buffer to a 1 mL of sample and vortexed. The sample was centrifuged (18407 RCF, 2 min), the supernatant was discarded, and 1 mL of Washing Buffer was added, mixed thoroughly and centrifuged again (18407 RCF, 1 min).
The supernatant was discarded, and the pellet was dried at 37 • C using a vacuum concentrator (LABCONCO CentriVap DNA ® Concentrator, Labconco Corporation, Kansas City, MO, USA). After that, FA and ACN were added proportionally to the amount of the dried pellet. The MALDI-TOF MS analysis was performed as previously described. For each bacterial species, 12 samples were prepared, including two liquid media, three different incubation times, and two sample preparation methods (six samples for TSB medium for bacteria incubated consecutively 4, 6, and 24 h, and extracts prepared by the standard and the Sepsityper method). In total, 36 samples were analysed.

Conclusions
Using the MALDI-TOF MS approach, we confirmed the high involvement of Staphylococcus, Enterococcus, and Escherichia bacteria in the process of inducing post-operative wound infections. The conducted research proved that the selection of the culture medium and the preparation of bacterial extracts significantly affected the identification factor and the quality of the mass spectra obtained. The best quality of the spectra was obtained using non-selective media and the standard method of sample preparation, which is also confirmed by the literature data.
The use of non-selective media, such as the BHI or MH, also leads to obtaining more types of bacteria. The incubation time strongly influenced the bacterial identification; therefore, we concluded that the optimal time can significantly accelerate the microbiological diagnosis of post-operative wound infections and implement the individual therapy faster. The use of the UniProt database made it possible to indicate proteins influencing the emergence of drug resistance and predict the antibiotics' actions.
The obtained results of the studies showed that MALDI-TOF MS with the use of additional databases, such as UniProt, can be a relevant tool for the detection of antibiotic resistance. Microbiological proteomics is a powerful tool for basic microbiological research not only in determining microbial mechanisms and physiology but also as a clinical diagnosis and antimicrobial therapy guideline. The obtained results may enable us to verify the medical diagnosis; they bring hope for the development of methods enabling a faster diagnosis due to the detection of disease changes at the cellular level prior to the occurrence of clinical changes, the monitoring of the course of the disease treatment, and the development of drugs for individual patients.