Changes in the Protein Profile in Staphylococcal Strains from Patients Infected with the SARS-CoV-2 Virus

Coronavirus disease 2019 (COVID-19) has affected a huge number of patients; however, the mortality rate often results from bacterial co-infection. The phenomenon of co-infection itself is widely known, but only some publications describe this mechanism in relation to SARS-CoV-2 (Mirzaei et al. 2020). As the gut bacterial diversity COVID-19 patients is reduced, bacterial co-infections might be a critical element that promotes severities of COVID-19 and mortality rates (Hoque et al. 2021). It has also been shown that the commensal microbiota can regulate and be regulated by invading viruses, causing stimulatory or suppressive effects (Bernardo et al. 2002; Franco et al. 2011; Fetsch et al. 2014). In the viral infections of the respiratory tract, bacterial co-infections are very common and vary from 3.0% to 68.0% (Hoque et al. 2021). It is speculated that the mechanism of co-infection includes virus-induced airway damage, cell loss, goblet cell hyperplasia, altered mucus secretion, reduced ciliary beat frequency, function and clearance, reduced oxygen exchange, and damage to the immune system (Hoque et al. 2021).

Due to the high incidence of infections, high infection rate with a severe course, and the severity of post-infection complications, particular interest should be paid to the co-infection with S. aureus as it is the most commonly identified pathogen in patients with SARS-CoV-2 (Clancy and Nguyen 2019). These bacteria may significantly impact the outcome of SARS-CoV-2 infections. Additionally, infections of the lower respiratory tract caused by multidrug-resistant strains of S. aureus may lead to substantial morbidity and mortality (Clancy and Nguyen 2019). Moreover, these bacteria may colonize healthy people’s respiratory tract and cause infections of varying severity. These infections range from local to non-invasive, from relatively mild skin infections to severe, life-threatening sepsis, infective endocarditis, and necrotizing pneumonia (Lowy 1998, van Belkum 2009).

The risk of staphylococcal infections, which causes considerable mortality, increases in patients with viral infections (Morens et al. 2008). S. aureus is particularly often isolated from patients with SARS-CoV-2 infections, especially in fatal cases (Lai et al. 2020). The reasons for this phenomenon have yet to be fully understood, but the mechanism may be multifactorial. It can be assumed that the virulent strains of S. aureus are selected during viral infections or that there are adaptive changes in potentially pathogenic staphylococci. As bacterial cell populations are frequently heterogenous, we may observe variations in pathogenicity.

Genetic typing methods have been used to recognize more virulent strains and follow the spread of such strains. In the epidemiology of staphylococcal infection, spa typing is widely accepted. Spa typing amplifies fragment X of the spa gene responsible for the production of species-specific S. aureus protein A, which is further sequenced (Shopsin et al. 1999; Hallin et al. 2009; Hasman et al. 2010). Fragment X of the spa gene is responsible for the attachment of protein A to the cell wall and contains a variable number of short repeat sequences (SSRs), each 21–24 nucleotides long (Shopsin et al. 1999). Their formation is associated with spontaneous point mutations, duplications, or deletions of repetitions in staphylococcal evolution (Frénay et al. 1996; Shopsin et al. 1999). Replicates obtained from sequencing are sequentially numbered and processed using the Ridom Spa Type program (http://spaserver.ridom.de). The result is a numeric code that identifies the spa types along with sequence and region changes (Harmsen et al. 2003).

The molecular techniques in microbiology in the last decade have been supplemented with a MALDI-TOF technique. This method allows for identifying microorganisms at the species level by analysis of total protein mass spectra and comparison to reference spectra (Sauget et al. 2017). Considering the above, protein profiling is a promising method in bacterial typing. In contrast to molecular techniques, MALDI-TOF MS requires minimal sample preparation, and is quick and inexpensive. This technique has successfully typed Gram-negative and Gram-positive bacteria (Sauget et al. 2017). It has been observed that modification of protein expression is reflected in the peak intensity. On the other hand, it is impossible to distinguish if the loss of the signal results from a lack of gene encoding the particular protein or a lack of gene expression. In contrast, a slight change in protein’s m/z value (peak shift) can be used as a reliable biomarker since it usually results from slight protein changes (Böhme et al. 2012; Josten et al. 2013; Østergaard and Møller 2018). Our previous research showed the increased expression of proteins crucial for the pathogenesis of bacteria during concomitant viral infection (Jarzembowski et al. 2013). The aim of our current research was to determine whether the SARS-CoV-2 virus infection affects the protein profile of S. aureus, and to find a biomarker for the risk assessment of bacterial co-infection development.

Forty swabs, including 17 wound swabs from diabetes patients, 12 nasal swabs from patients hospitalized due to SARS-CoV-2 infection, and 11 nasal swabs from patients without current viral infection, were collected from the patients of the hospitals of the Pomeranian region (Table I). The samples were inoculated on Columbia Blood Agar (Graso, Poland) containing 5% (vol/vol) horse blood at 35–37°C for 18–24 h. Identification of S. aureus strains was performed using the advanced colorimetric method VITEK^® GP, VITEK 2^® Compact (bioMerieux, Poland). The isolates from SARS-CoV-2 positive patients were compared by protein profile with the isolates from patients without SARS-CoV-2 and symptoms of respiratory tract infection (negative reference). The spa typing method was used to distinguish between the selection and adaptation of strains.

Strain protein profiling. A colony material sample was transferred to the target steel plate (Bruker Daltonik GmbH, Germany) without a protein extraction procedure (direct method). Samples were overlaid with 1 μl of HCCA matrix (alpha-cyano-4-hydroxycinnamic acid) solution. MALDI-TOF MS spectra were obtained using a Microflex LT smart system with Bruker’s proprietary smartbeam^™ solid-state laser technology at 200 Hz repetition rate. The system was controlled by the flex-Control software Version 3.4 (Bruker Daltonik GmbH, Germany) and calibrated using a sample of Bruker bacterial test strain (BTS) containing Escherichia coli DH5. Spectra for individual specimens were compiled as an averaged result from 100 shots for m/z values in the range of 2,000–10,000. Spectra for the two replicates were compared to determine the within-isolate peak reproducibility, then summed to minimize random effects. Data analysis of protein profile was conducted by mas software (Strohalm et al. 2008). Classification and Regression Tree (C&RT) and ANOVA comparison of protein peaks in groups of strains was performed using the Statistica software package (ver. 13.0).

Spa typing. Spa typing was applied to evaluate the distribution and selection of the staphylococcal strains in SARS-CoV-2 infected patients. The polymorphic fragment X of the spa gene encoding protein A was amplified by PCR reaction using the primers proposed by Frénay et al. (1996). 5 μl of bacterial DNA (5 ng/μl) was mixed with 5 μl of reaction buffer (100 mM TRIS-HCl pH = 8, 500 mM KCl, 15 mM MgCl₂, 0.8% detergent; MBI Fermentas, Lithuania), 5 μl of dNTP mixture (dATP, dCTP, dGTP, dTTP) at a concentration of 2 mM each, 1 μl of each of the SPA-1 and SPA-2 primers (100 μM each concentration), 2 U Taq DNA polymerase (MBI Fermentas, Lithuania) and made up to a volume of 50 μl with water (Sigma, USA). Amplification was performed in a GeneAmp PCR System 2400 automatic thermocycler (PerkinElmer, USA) according to the scheme (Harmsen et al. 2003). Amplification results were read on a 4% agarose gel (Sigma, USA) stained with ethidium bromide (ICN Biomedicals Inc., Austria). For this purpose, 2 μl of the loading buffer (60% glycerol, 0.09% bromophenol blue, 0.09% xylene, 60 mM EDTA; MBI Fermentas, Lithuania) was added to 10 μl of the reaction mixture, and the whole was applied to the covered in TAE buffer (0.04 M Tris-acetate, 0.001 M EDTA; Sigma, USA) wells in an agarose gel (2% agarose in TAE buffer supplemented with ethidium bromide; Sigma, USA). The electrophoresis was performed for 90 minutes (at a current of 45 mA) using a STABNAP 300 power supply (Kucharczyk TE, Poland). Then the gel was photographed on a UV transilluminator (Sigma, USA). The amplified spa gene fragment’s size was compared with the molecular weight marker position (GeneRuler 50 bp DNA Ladder, MBI Fermentas, Lithuania). PCR product sequencing was performed using an ABI 377 apparatus (Applied Biosystems, USA). The resulting short repeated sequences (SSRs) were numbered and compiled using the Ridom SpaType program available at http://spaserver.ridom.de. The results were given as a numeric code identifying the spa types.

In our study, we identified 29 peaks ranging from 2,151 to 9,632 m/z. Out of them, 12 peaks were already described in the literature as characteristic of S. aureus (Table II). Five of the peaks were the most common (prevalence >87,5%), including 5,035, 6,820, and peaks 3,445, 4,307, and 6,892, which were detected in all isolates (prevalence 100%). What is important, one of the peaks (2,430) was described here for the first time and was unique for the isolates from patients infected with the SARS-CoV-2 virus.

The protein peaks expressed high diversity in prevalence among strains (5–100%) and relative amount (11–100%). Within all groups of isolates, we noticed the high diversity of spa types. In isolates from SARS-CoV-2 positive patients, six different spa types were found, and all of them had a unique peak of 2,430 m/z. A peak of 2,151 m/z, described for the first time in this study, was found in the t019 and t253 spa types, while the peak of 2,896 in the spa type t230. C&RT data analysis of the prevalence and intensity of protein peaks resulted in the identification of seven groups of strains (Fig. 1). Isolates from the wound were found in all of them, while isolates from SARS-CoV-2 positive patients were only in two protein groups.

Fig. 1.

Among proteins designated as a split in C&RT, the abundance of protein with m/z 3,410 was the highest among S. aureus isolates from the nose swab, as the abundance of protein 2,652 – was the highest in isolates from wounds. In the strains from nose swabs from SARS-CoV-2 positive patients, the abundance of proteins 3,445 and 5,306 was lower than in other groups (Fig. 2).

Fig. 2.

The protein profile of bacteria is used for identifying species by the matrix-assisted laser desorption ionisation time-of-flight mass spectrometry (MALDI-TOF MS). Such identification of organisms is based on the generation of mass spectra obtained from a colony, which are compared to the spectra of known species in a reference library (Clark et al. 2013). However, until now, few studies have described using MALDI-TOF MS as a method of discrimination between different strains of S. aureus (Bernardo et al. 2002; Jackson et al. 2005; Wolters et al. 2011; Lasch et al. 2014). In our study, we have analysed the relative signal intensity as the amplitude of peaks in MALDI-TOF. Even though spectra are known to show variability due to several factors (Jackson et al. 2005; Wiśniewska et al. 2014; Mirzaei et al. 2020), we have identified 12 protein peaks already described in the literature. These results correspond to the previous reports and support the opinion on the described protein significance. Similarly to Østergaard and Møller (2018), we have found that peaks 6,892 and 3,445 were present in all strains, pointing to their essentiality for the species. However, one should consider that comparison could be limited by variations in spectra resulting from differences in material distribution on the target plate and as a result of different abundance of ionised protein (Østergaard and Møller 2018).

Apart from already described peptides, we have identified the protein unique for SARS-CoV-2 isolates. According to the most common opinion, viral infection facilitates bacterial adherence by modification of epithelial cell properties (Mirzaei et al. 2020). However, this study confirms the previous finding that some proteins, likely adhesins, may have a different expression when the strains are isolated from patients with SARS-CoV-2 infection. For example, changes in bacterial properties have been described in the CMV and enterococcal co-infection (Jarzembowski et al. 2013).

C&RT data analysis showed that the protein profile is related to the prevalence of the strains as commensals in the host or etiological agents of the infection. Furthermore, the significant difference between the intensity of peaks 3,410 and 3,445 between the nose and wound isolates may suggest that the above protein contributes to the infection development and could be a candidate for the marker in the estimation of co-infection risk. In the current study, we failed to find a clear correlation between a spa type and the protein profile. The high diversity of spa types in the group of different origin makes the hypothesis of selection of strains during the SARS-CoV-2 infection not likely. Considering the above, the results obtained here support the hypothesis of bacterial adaptation do new conditions caused by viral infection.

The analysis of polymorphism in the X region of the protein A gene (spa typing) has proven to be an effective typing tool to distinguish strains within a heterogeneous species such as S. aureus. With spa typing, a significant number of different types were obtained. Some correlations between spa types and sources of isolation could be found. The spa type t127 related to MLST CC1 was frequently isolated from bloodstream isolates in the Gdansk area (Wiśniewska et al. 2014). In Polish laboratories, the most prevalent was t127/ST-1 among MSSA (Methicillin Sensitive Staphylococcus aureus) strains (Grundmann et al. 2010). Spa type t127 was present among hospital and community patients, but spa t159 – was only among hospital isolates in Malaysia (Ghasemzadeh-Moghaddam et al. 2011). MSSA spa type t127 was isolated from the nasal cavity of handless in China, but MRSA (Methicillin Resistant Staphylococcus aureus) of spa type t127 was found in holdings of breeding pigs in Italy (Franco et al. 2011). An enterotoxin type A producing S. aureus of spa type t127 was described in human cases and different ice-cream types (Fetsch et al. 2014). Most of the PVL-positive MSSA isolates were obtained from wound infections in Nigeria and classified in clonal complexes CC1 (t127) and CC121 (t159) (Shittu et al. 2011). PVL-producing MSSA affiliated with CC121 are common worldwide (Wiese-Posselt et al. 2007). Kwapisz et al. (2020) found that the predominant clonal complex in dental isolates from infections was CC45, with the most common spa type being spaCC t015. CC45 are widely distributed among both the strains colonizing nares and bloodstream infections in Europe (Deasy et al. 2019; Roe et al. 2020), and recent results suggest that the nasal isolates carry the potential to cause an invasive disease (Bonnet et al. 2018). However, no reports of severe infections caused by oral CC45 strains have been published. Bonnet et al. (2018) observed a predominance of spaCC t015 among S. aureus strains associated with infective endocarditis, and Deasy et al. (2019) pointed to this spa type as an emerging etiological factor of bloodstream infections.

The fact that only local isolates were collected and analysed could be a potential weakness of our study. On the other hand, the strength of the present study is that the MALDI-TOF peaks reported in the published papers were confirmed in our isolates, as well as that some unique proteins were identified in isolates from co-infected patients.