More than a Toxin: Protein Inventory of Clostridium tetani Toxoid Vaccines

Clostridium tetani is the etiological agent of tetanus, a life-threatening bacterial infection. The most efficient protection strategy against tetanus is a vaccination with the C. tetani neurotoxin, which is inactivated by formaldehyde-crosslinking. Since we assumed that besides the tetanus toxin, other proteins of C. tetani may also be present in toxoid preparations, we analyzed commercially available vaccines from different countries in respect to their protein content using mass spectrometry. In total 991 proteins could be identified in all five analyzed vaccines, 206 proteins were common in all analyzed vaccines and 54 proteins from the 206 proteins were potential antigens. The additionally present proteins may contribute at least partially to protection against C. tetani infection by supporting the function of the vaccine against the devastating effects of the tetanus toxin indirectly. Two different label-free protein quantification methods were applied for an estimation of protein contents. Similar results were obtained with a Total Protein Approach (TPA)-based method and Protein Discoverer 2.2 software package based on the minora algorithm. Depending on the tetanus toxoid vaccine and the quantification method used, tetanus neurotoxin contributes between 14 and 76 % to the total C. tetani protein content and varying numbers of other C. tetani proteins were detected.


Introduction
Toxigenic strains of Clostridium tetani can cause tetanus, a life-threatening bacterial infection. Spores of this obligate anaerobic, saprophytic bacterium, which preferentially lives in warm and moist habitats, are present in environments all over the world and may enter the human body via contamination of wounds. These may be the result of injuries in case of adults or children or occur in case of newborns during birth when the umbilical cord is cut (neonatal tetanus). Under anaerobic conditions in dirt-contaminated wounds, poorly blood supplied or necrotic tissue, the spores may germinate and the vegetative cells can proliferate and produce exotoxin. Tetanospasmin, the extremely potent neurotoxin of C. tetani, blocks inhibitory neurotransmitters of the central nervous system, preventing relaxation of muscles and leading to muscle rigidity and spastic paralysis that are typical for tetanus [1]. Based on the frequent abundance of C. clostridium spores in the environment and a high fatality rate, which is almost 100% in case of lacking immune protection and intensive medical treatment, tetanus is a serious threat for human health. However, it can easily be prevented by vaccination with tetanus toxoid vaccine, which was already introduced in 1924, making it one of the oldest used vaccines (for review, see reference [2]).
Tetanus vaccine production is based on a C. tetani strain, designated as the "Harvard strain", which was initially collected in the United States of America in the 1920s and subsequently distributed

Preparation of Vaccines Used in This Study
Commercially available diphtheria toxoid and tetanus toxoid (DT) vaccines used for proteome analyses are shown in Table 1. Since the protein content of the vaccine analyzed in this study was considered low, 10 vaccine doses (0.5 mL each) were pooled and precipitated by addition of 10% (w/v) trichloroacetic acid (TCA) and incubation at 4 • C for 16 h to get a concentrated sample for mass spectrometry analysis [11]. After incubating over-night at 4 • C the samples were centrifuged at 8000 x g for 30 min at 4 • C. The precipitated proteins were dried on ice and resuspended in rehydration buffer (2% sodium deoxycholate, 10 mM dithiothreitol (DTT), 50 mM Tris, pH 8.0). To reverse the formaldehyde cross-linking of the inactivated toxins, the samples were incubated for 20 min at 95 • C [12]. The protein amount one vaccine dose was determined using a spectrophotometer (NanoDrop LITE, Thermo Fisher Scientific, Bremen, Germany) at 280 nm.

Russia
Microgen Tetanus toxoid ≥ 5 BU/mL Diphtheria toxoid ≥ 5 Lf/mL *1: Flocculation units (Lf) are defined as the equivalent of toxin to one unit of standard antitoxin or the amount of toxin that flocculates one unit of an international reference antitoxin [13,14]. *2: International units (I.U.) are defined as a certain weight of the international standard for antitoxin [15] and are equal to the corresponding amount of toxin. *3: Binding units (BU) correspond to the amount of tetanus toxoid to bind directly or compete with tetanus antibody [16] and, therefore, are equal to Lf or I.U.

Tryptic Digest and C18 Clean up
About 10 µg soluble proteins prepared from the vaccine samples (see above) were transferred to 10 kDa vivacon 500 membrane filters and the flow-through was discarded by centrifugation for Proteomes 2019, 7, 15 3 of 11 30 min at 12,000 × g. The tryptic digest of the prepared vaccines samples occurred within the modified Filter Aided Sample Preparation (FASP) protocol [11,17]. The proteins were reduced by addition of 200 µl of reduction buffer (25 mM DTT, 8 M urea, 50 mM triethylammonium bicarbonate buffer (TEAB)) for 30 min at 37 • C. Alkylation of sulfhydryl groups were carried out with 200 µl alkylation buffer (25 mM chloroacetamide (CAA), 8 M urea, 50 mM TEAB) for additional 30 min on a shaker at 600 rpm in the dark. The proteins were subsequently washed with 300 µl of 8 M urea in 50 mM TEAB followed by another washing step with 200 µL 6 M urea in 50 mM TEAB. Afterwards 0.5 µg mass spectrometry grade LysC endopeptidase was added onto the filter unit and incubated on a shaker at 37 • C and 600 rpm for 3 h, followed by a second digest with 1 µg trypsin and 250 µL dilution buffer (50 mM TEAB) to reach a final concentration of 1 M urea. The sample was incubated over-night at 37 • C at 600 rpm on a shaker. Peptides were then collected by centrifugation at 12,000 × g for 20 min. For acidification of the peptide solution 20 µL of 10% trifluoroacetic acid (TFA) was added to reach a final concentration of 0.5% TFA. To desalt the peptides a clean-up with C18 stage tips were performed. Prior to LC-MS/MS analysis, peptides were vacuum dried and resuspended in 0.1% trifluoroacetic acid (TFA) [8].

Mass Spectrometry
For mass spectrometric analyses peptides (250 or 2500 ng, respectively) were loaded onto a nanoflow Ultimate 3000 HPLC (Dionex, Sunnyvale, CA, USA). Separation was carried out using an EASY-Spray column (Thermo Fisher Scientific; C18 with 2 µm particle size, 50 cm × 75 µm) with a flow rate of 200 nl min −1 and increasing acetonitrile concentrations over 120 min. Method duration including equilibration and column wash was in total 160 min. All samples were analyzed on an Orbitrap Fusion mass spectrometer (Thermo Fisher Scientific, Waltham, MA, USA). The mass spectrometer was operating with 2000 V spray voltage, 275 • C transfer tube temperature, 300-2000 (m/z) scan range for the MS 1 detection in the Orbitrap, a maximum injection time of 50 ms, an automatic gain control (AGC) target of 400,000 and an Orbitrap resolution of 120.000. The most intense ions were selected for collision-induced dissociation with collision energy of 35%. For ion trap detection a maximum injection time of 250 ms and an AGC target of 100 were set [8,17]. Resulting raw data files were analyzed using the Clostridium tetani E88 database (Proteome Id: UP000001412) and the C. diphtheriae ATCC 700971/NCTC 13129/biovar gravis database (Proteome Id: UP000002198) in UniProt (www.uniprot/proteomes) and the Proteome Discoverer 1.4 program package (Thermo Fisher Scientific, Bremen, Germany). As described by Schäfer and co-workers [18] theoretical masses for peptides were generated by trypsin digestion with a maximum of two missed cleavages. Product ions were compared to the measured spectra using the following parameters: carbamidomethyl modification on cysteine was set as fixed and oxidation of methionine as dynamic modification. Mass tolerance was set to 10 ppm for survey scans and 0.6 Da for fragment mass measurements. For protein identification the thresholds were set on 1% false discovery rate (FDR).

Label-Free Quantitative Protein Analysis
For protein quantification, three samples from the German and Indian vaccine as well as three single vaccine doses from the Russian vaccine were prepared and analyzed by mass spectrometry. Based on the total protein approach (TPA) method [19,20] we used the peak area for protein quantification [21,22]. The total protein content was defined as the sum of peptide intensities integrated over the elution profile of each peptide and the amount of each identified protein was calculated as the ratio of their intensity to the sum of all intensities in the sample [23]. Only peaks ranged from 2 × 10 7 up to 10 11 where used for quantification [24]. In addition to the TPA-based quantification, the label-free quantification method based on the Minora algorithm [25] and included in the Protein Discoverer 2.2 program package (Thermo Fisher Scientific, Bremen, Germany) was applied.

Proteome Prediction of C. tetani
Cellular localization and the characteristics of C. tetani proteins identified in different samples of vaccines were extracted from the UniProt database. Proteins with unknown localization were analyzed by PSORTb v3.0.2 [26].

Prediction of Putative Antigens
A putative antigen function of proteins identified in the vaccine samples was determined using the VaxiJen database (http://www.ddg-pharmfac.net/vaxijen/VaxiJen/VaxiJen.html) with a threshold of 0.4 [27].

Data Availability Statement
Mass spectrometry results have been deposited to the ProteomeXchange Consortium (http: //proteomecentral.proteomexchange.org) via the PRIDE partner repository [28]. Data are available via ProteomeXchange with identifier PXD009289.

Identification of Proteins
For protein analysis, the processed vaccine samples were subjected to mass spectrometry. Identified peptides were analyzed with the Proteome Discoverer 1.4. For further investigations on protein localization and function the UniProt database was used. In total 991 C. tetani proteins were identified in all five analyzed vaccines (Table S1). Interestingly, the number of proteins identified in the different samples varied significantly. The highest numbers were observed for the vaccines from India and Russia with 662 and 627 proteins, respectively. The vaccine from Germany comprised 566 different proteins, 427 proteins were found in the vaccine sample from Brazil and 369 proteins in the Bulgarian vaccine. 206 unique proteins were present in all five analyzed vaccines ( Figure 1). in the Protein Discoverer 2.2 program package (Thermo Fisher Scientific, Bremen, Germany) was applied.

Proteome Prediction of C. tetani
Cellular localization and the characteristics of C. tetani proteins identified in different samples of vaccines were extracted from the UniProt database. Proteins with unknown localization were analyzed by PSORTb v3.0.2 [26].

Prediction of Putative Antigens
A putative antigen function of proteins identified in the vaccine samples was determined using the VaxiJen database (http://www.ddg-pharmfac.net/vaxijen/VaxiJen/VaxiJen.html) with a threshold of 0.4 [27].

Data availability Statement
Mass spectrometry results have been deposited to the ProteomeXchange Consortium (http://proteomecentral.proteomexchange.org) via the PRIDE partner repository [28]. Data are available via ProteomeXchange with identifier PXD009289.

Identification of Proteins
For protein analysis, the processed vaccine samples were subjected to mass spectrometry. Identified peptides were analyzed with the Proteome Discoverer 1.4. For further investigations on protein localization and function the UniProt database was used. In total 991 C. tetani proteins were identified in all five analyzed vaccines (Table S1). Interestingly, the number of proteins identified in the different samples varied significantly. The highest numbers were observed for the vaccines from India and Russia with 662 and 627 proteins, respectively. The vaccine from Germany comprised 566 different proteins, 427 proteins were found in the vaccine sample from Brazil and 369 proteins in the Bulgarian vaccine. 206 unique proteins were present in all five analyzed vaccines ( Figure 1).

Figure 2.
Localization of proteins identified in the analyzed vaccines. The proportion of proteins in respect to their predicted localization is shown for the vaccines from different countries and for the proteins common in all vaccines (grey: without predicted localization, green: cytoplasmic proteins, blue: membrane proteins and black: secreted proteins). For proteins with more than one annotated localization, each localization was added to the corresponding group.

Bioinformatic Analysis of the 54 Common Unique Proteins in All Analyzed Vaccines
To characterize the immunological potential of C. tetani proteins identified as parts of the tetanus toxoid vaccine, the proteins were analyzed in respect to the precence of signal peptides (SP) for protein translocation, transmembrane (TM) domains and their potential antigenic properties ( Figure  3). From a total of 206 common proteins in all analyzed vaccines 23 proteins comprised a signal peptide (SP: 21, SP/TM: 2) and 10 proteins contained a transmembrane domain (TM: 8, TM/SP: 2), while 175 proteins comprised none of these motifs. Fifty-four proteins were predicted antigens based on a VaxiJen database screening (Table 2).

Figure 2.
Localization of proteins identified in the analyzed vaccines. The proportion of proteins in respect to their predicted localization is shown for the vaccines from different countries and for the proteins common in all vaccines (grey: without predicted localization, green: cytoplasmic proteins, blue: membrane proteins and black: secreted proteins). For proteins with more than one annotated localization, each localization was added to the corresponding group.

Bioinformatic Analysis of the 54 Common Unique Proteins in All Analyzed Vaccines
To characterize the immunological potential of C. tetani proteins identified as parts of the tetanus toxoid vaccine, the proteins were analyzed in respect to the precence of signal peptides (SP) for protein translocation, transmembrane (TM) domains and their potential antigenic properties (Figure 3). From a total of 206 common proteins in all analyzed vaccines 23 proteins comprised a signal peptide (SP: 21, SP/TM: 2) and 10 proteins contained a transmembrane domain (TM: 8, TM/SP: 2), while 175 proteins comprised none of these motifs. Fifty-four proteins were predicted antigens based on a VaxiJen database screening (Table 2).

Quantification of Proteins
The three vaccines with the highest number of different proteins were chosen for a more detailed characterization of the relative amounts of distinct proteins. In a first approach, DT vaccines were analyzed in respect to their protein content originating from C. tetani and C. diphtheriae. In all cases, the major part of the proteins of the DT vaccines analyzed were proteins from tetanus toxoid preparations, while C. diphtheriae proteins were present in much lower abundance, which corresponds to the different international units combined in the DT vaccines (see Table 1).
As next step, the relative amount of tetanus neurotoxin (TeNT) in the tetanus toxoid preparations was determined. TeNT represented 48.2 ± 12.2% in the German vaccine, 43.7 ± 5.8% in the vaccine from India and only 17.2 ± 3.0% in the vaccine from Russia when analyzed with the TPA-based method (Figure 5a). When the Proteome Discoverer software package 2.2 was used, TeNT reached 72.9 ± 8.4% in the German vaccine, 72.9 ± 2.8% in the Indian vaccine and 29.4 ± 1.1% in the vaccine from Russia when analyzed with the TPA-based method (Figure 5b). vaccine samples, 11.9 ± 5.8% in the Indian samples and 14.9 ± 0.1% in the Russian vaccine. When the Proteome Discoverer 2.2 program package based on the minora algorithm [25] was applied, a similar distribution was observed. In the German vaccine 95.5 ± 0.5% C. tetani and 4.5 ± 0.5% C. diphtheriae proteins were found, in the Indian vaccine 93.1 ± 1.3% C. tetani and 6.9 ± 1.3% C. diphtheriae proteins were observed and in the Russian vaccine 90.5 ± 0.1% C. tetani and 9.5 ± 0.1% C. diphtheriae proteins were detected (Figure 4b). Taken together, with this method higher relative amounts of TeNT were detected; however, in all cases the Russian vaccine contained the lowest amount of TeNT. Since the toxoid was adjusted in form of flocculation units [13,14], this meant that the Russian toxoid was the most immunologically active in this comparison.
When the protein content from C. tetani in the Russian vaccine was analyzed in more detail using the TPA-based method, the TeNT represented only 17.2 ± 3.0% of all C. tetani proteins, while the putative S-layer protein/N-acetylmuramoyl-L-alanine amidase CTC_00462 reached 16.7 ± 0.8% of the protein content, the putative S-layer protein CTC_00465 9.6 ± 1.1% and the remaining proteins summed up to 56.5 ± 3.0% (Figure 6a). Also in this case, the general tendency of protein distribution was similar when the Proteome Discoverer software package was used. In this case the TeNT reached 29.4 ± 1.1% of all C. tetani proteins, while the putative S (surface)-layer protein/N-acetylmuramoyl-L-alanine amidase CTC_00462 reached 17.3 ± 0.3%, the putative S-layer protein CTC_00465 13.5 ± 0.4% and the remaining proteins 39.8 ± 1.1% (Figure 6b).  As next step, the relative amount of tetanus neurotoxin (TeNT) in the tetanus toxoid preparations was determined. TeNT represented 48.2 ± 12.2% in the German vaccine, 43.7 ± 5.8% in the vaccine from India and only 17.2 ± 3.0% in the vaccine from Russia when analyzed with the TPAbased method (Figure 5a). When the Proteome Discoverer software package 2.2 was used, TeNT reached 72.9 ± 8.4% in the German vaccine, 72.9 ± 2.8% in the Indian vaccine and 29.4 ± 1.1% in the vaccine from Russia when analyzed with the TPA-based method (Figure 5b).
Taken together, with this method higher relative amounts of TeNT were detected; however, in all cases the Russian vaccine contained the lowest amount of TeNT. Since the toxoid was adjusted in form of flocculation units [13,14], this meant that the Russian toxoid was the most immunologically active in this comparison. When the protein content from C. tetani in the Russian vaccine was analyzed in more detail using the TPA-based method, the TeNT represented only 17.2 ± 3.0% of all C. tetani proteins, while the putative S-layer protein/N-acetylmuramoyl-L-alanine amidase CTC_00462 reached 16.7 ± 0.8% of the protein content, the putative S-layer protein CTC_00465 9.6 ± 1.1% and the remaining proteins summed up to 56.5 ± 3.0% (Figure 6a). Also in this case, the general tendency of protein distribution was similar when the Proteome Discoverer software package was used. In this case the TeNT reached 29.4 ± 1.1% of all C. tetani proteins, while the putative S (surface)-layer protein/N-acetylmuramoyl-Lalanine amidase CTC_00462 reached 17.3 ± 0.3%, the putative S-layer protein CTC_00465 13.5 ± 0.4% and the remaining proteins 39.8 ± 1.1% (Figure 6b). protein content, the putative S-layer protein CTC_00465 9.6 ± 1.1% and the remaining proteins summed up to 56.5 ± 3.0% (Figure 6a). Also in this case, the general tendency of protein distribution was similar when the Proteome Discoverer software package was used. In this case the TeNT reached 29.4 ± 1.1% of all C. tetani proteins, while the putative S (surface)-layer protein/N-acetylmuramoyl-Lalanine amidase CTC_00462 reached 17.3 ± 0.3%, the putative S-layer protein CTC_00465 13.5 ± 0.4% and the remaining proteins 39.8 ± 1.1% (Figure 6b).

Discussion
Proteomic data are scarce for C. tetani and only a few proteins have been identified in proteome studies up to now [29]. In frame of this study, we were able to identify 991 different proteins of C. tetani as components of tetanus toxoid vaccines. In addition, we found a highly variable number of proteins depending on the studied vaccines, which may indicate that different companies use different protocols for tetanus toxoid production.
Independent of the number of detected proteins and the protein quantification method used, the main component of all vaccines was tetanus neurotoxin. The high number of predicted cytoplasmic proteins found in this study, 64% of the total identified proteins, may be explained by cell lysis of bacteria, which leads to a release and accumulation of especially highly abundant and stable proteins such as components of glycolysis, pentose phosphate pathway, tricarboxylic acids cycle and protein turnover into the supernatant. Alternatively, nonclassical transporters or nonspecific leakage or release by exosomes was discussed for these group of proteins [5]. Shearing forces may be another source of membrane-associated proteins, resulting for example from stirring during cultivation.
In summary, besides the tetanus neurotoxin, which contributes between 14 and 76% to the total C. tetani protein content depending on the tetanus vaccine sample and the protein quantification method used, a high number of other C. tetani proteins were detected, which may contribute at least partially to a protection against C. tetani infection supporting the function of the vaccine against the devasting effects of the tetanus toxin indirectly.
Interestingly, Clostridium botulinum toxoid vaccination in Danish cows did not only reduce botulinum neurotoxin in cattle feces, but also the number of C. botulinum spores. This observation may support the idea that toxoid preparation may also protect against infection with pathogenic bacteria [30].

Conclusions
DT vaccines consist of about 20% C. diphtheriae and about 80% C. tetani proteins. Tetanus neurotoxin is a major protein in toxoid preparations; however, similar amounts of putative S-layer proteins are found as well. These and other proteins found may contribute to protection against C. tetani infection.
Author Contributions: Protein preparation, data analysis and visualization of results was carried out by J.M.; M.E.K. carried out mass spectrometry analyses and was responsible for data storage; conceptualization, supervision of experiments; writing of the draft as well as project administration was carried out by A.B.
Funding: This research received no external funding.