An oxidation resistant pediocin PA-1 derivative and penocin A display effective anti-Listeria activity in a model human gut environment

ABSTRACT Pediocin PA-1 is a class IIa bacteriocin that is particularly effective against the foodborne pathogen Listeria monocytogenes. The loss of activity of PA-1 pediocin due to methionine oxidation is one of the challenges that limit the wider application of the bacteriocin. In this study, we heterologously expressed an oxidation resistant form of pediocin PA-1, i.e., pediocin M31L, and compared its activity to that of native pediocin PA-1 and to penocin A, a pediocin-like bacteriocin that displays a narrower antimicrobial spectrum. Minimal inhibitory concentration assays revealed that pediocin M31L was as effective as PA-1 and more effective than synthetic penocin A against Listeria with negligible activity against a range of obligate anaerobic commensal gut bacterial species. The anti-Listeria activity of these pediocins was also assessed in a simulated human distal colon model assay using the L. monocytogenes, spiked at 6.5 ±  0.13 Log CFU/mL, as a bioindicator. At 24 h, pediocin M31L and penocin A (2.6 μM) reduced Listeria counts to 3.5 ± 0.4 and 3.64 ± 0.62 Log CFU/mL, respectively, whereas Listeria counts were considerably higher, i.e. 7.75 ±  0.43 Log CFU/mL, in the non-bacteriocin-containing control. Ultimately, it was established that synthetic penocin A and the stable pediocin M31L derivative, heterologously produced, display effective anti-Listeria activity in a human gut environment.


Introduction
Bacteriocins are a heterogeneous group of antimicrobial peptides ribosomally 1 produced by certain bacterial strains and are active against a wide range of microbial targets. 2 Despite their potential for use as food preservatives and decades of research by the food industry, nisin is the only bacteriocin authorized for human consumption by the Food and Drug Administration (FDA) and the European Food Safety Authority (EFSA). 3 Increasing demand for foods with fewer chemical preservatives combined with the appearance of multi-resistant bacterial strains has led to novel bacteriocins being proposed as an alternative natural food additive, postbiotic components in functional foods or even pharmaceutical therapeutics. [3][4][5] The risk of developing antimicrobial resistance can be reduced by using bacteriocins exiguously, and this can be achieved through combining them with other bacteriocins particularly those that use different modes of action, alternative antimicrobials, preservation additives or preservation methods. 4 Methicillin-resistant strains of Staphylococcus aureus, Streptococcus pneumoniae, Enterococcus spp. and Clostridioides difficile are, for example, sensitive to the bacteriocin nisin, and its use in combination with conventional antibiotics successfully inhibits an even wider range of Gram-positive pathogens and undesirable bacterial strains. 6,7 In addition, as certain bacteriocins have been reported as having anti-viral properties, they may have potential for SARS-CoV2 treatment. 8 Pediocin PA-1 is a class IIa bacteriocin, produced by lactic acid bacteria (LAB) such as Pediococcus spp. and Lactobacillus plantarum, which exhibits antimicrobial activity against the pathogen Listeria monocytogenes. [9][10][11] Listeriosis, caused by the Grampositive bacteria L. monocytogenes, is considered one of the most serious foodborne disease due to its high mortality rate, particularly among the elderly, people with compromised immune systems, pregnant women, unborn and neonatal babies. 12 Contamination is often through ready-to-eat foods where the pathogen can be present in high amounts due to its ability to grow under refrigeration temperatures. 13 The pathology varies from a noninvasive form (gastroenteritis, febrile diarrhea, headache) to invasive listeriosis, which manifests in more serious symptoms such as septicemia and meningitis. The effectiveness of pediocin PA-1, as an anti-Listeria preservative in food, has been studied for decades with many studies using ALTA 2341 (Quest International, Irvine, CA, USA), a commercially available pediocin-containing P. acidilactici syrup fermentate. In some studies, ALTA 2341 was incorporated into slurries prepared with Turkey breast 14 or in "queso blanco"-type cheese 15 and was found to decrease L. monocytogenes counts over time compared to untreated controls. Furthermore, studies have also established that use of pure pediocin led to a reduction of intestinal Listeria counts compared to the control in the gastrointestinal tract of BALB/c mice challenged with L. monocytogenes LSD 348. This occurred in the absence of major changes in the composition of mouse gut microbiota. 16 Further investigations using pediocin producers (ped+) or equivalent ped-studies have shown the absence of changes in the overall structure of the fecal microbiota in BALB/c female mice 17 and in rats with a human-associated microbiota. 18 This point is important as the bactericidal effect of new therapeutic compounds on the commensal microbiota has attracted more attention as a consequence of the increasing number of diseases and physiological imbalances that have been associated with disruption of the indigenous bacterial community present in the intestine.
An understanding of the dose response of an antimicrobial compound is required for clinical applications. In this regard, the use of pediocin PA-1-producing strains rather than the pure bacteriocin is problematic, given the variety of factors that can influence bacteriocin production in the human gut environment. Large-scale production of pediocin PA-1 in a highly pure form (>95%) is required for cytotoxicity and immunogenicity assessment to facilitate subsequent approval by regulatory authorities. Typically, at least three purification steps are required to produce a sufficiently pure form of a class IIa bacteriocin from a natural producing strain but, unfortunately, yields can often be low (from 350 to 1600 µg of class IIa bacteriocin per liter of culture). 19 A further complication in pediocin PA-1 production is that the oxidation of methionine at position 31 can significantly reduce its activity, and its substitution to leucine, isoleucine or alanine solve this problem without compromising its anti-Listeria activity. 20,21 To overcome these challenges, in this work, we optimized the production of an oxidation-resistant pediocin PA-1, pediocin PA-1 M31L, employing an expression vector system that controls plasmid copy number in E. coli cells previously described by Mesa-Pereira et al. 22 Chemical synthesis of penocin A, another bacteriocin that harbors a YGNGVX 1 CX 2 K/NX 3 X 4 C, (X 1-4 : polar uncharged or charged residues) pediocin box in its sequence was carried out in tandem to compare its antimicrobial efficacy against L. monocytogenes and the commensal gut microbiota. Penocin A is considered a silent bacteriocin as its mature form is not produced naturally due to the absence of a key gene (s) in the associated natural strains. Diep et al. 23 revealed that heterologous expression of the pediocin inducer factor led to penocin A expression in Pediococcus pentosaceus 25745 and its inhibition rate against 71 strains from a collection of 8 Gram positive bacteria genera (Carnobacterium, Clostridioides, Enterococci, Lactobacilli, Lactococci, Leuconostoc, Listeriae, Pediococci) was lower (48%) than that presented by pediocin PA-1 (58%). This reduced antimicrobial activity against some important bacterial groups (e.g Lactobacilli, Lactococci, Leuconostoc) may be a desirable feature when used as a therapeutic agent. Here the anti-Listeria effects of pure forms of recombinant pediocin M31L, natural pediocin PA-1 and synthetic penocin A were each evaluated in an ex vivo model that simulates the human gut environment for the first time. This study verifies the efficiency of the pediocin variants against L. monocytogenes and highlights the merits of progressing to in vivo experiments.

Cloning and expression of pediocin M31L
Recombinant expression of pediocin M31L was evaluated indirectly through antimicrobial activity of culture supernatant against Listeria innocua DPC3572. It was established that pediocin M31L was produced a b c under low plasmid copy number following induction for 3 and 6 hours but not overnight as observed by the well-defined inhibition zones (Figure 1a). To quantify the highest anti-Listeria activity, samples with the largest inhibition diameters were serially diluted twofold to quantify Bacteriocin Units (BU) (Figure 1b and 1c). Higher antimicrobial activity within the cell-free supernatant (CFS) was achieved in the presence of 1000 μM of IPTG following 6 hours induction at 37°C, the only condition that achieved 1/128 BU ( Figure 1c-red circled). This assay was performed in two independent assays. In addition, it was previously demonstrated that no anti-Listeria activity was observed in the extracellular fraction of E. coli (DE3) Tuner TM transformed with modified pETcoco empty vector. 22 Pediocin M31L and native pediocin PA-1 were purified from cell-free supernatant of E. coli/ pETcoco-pedM31L and P. acidilactici LMG 2351, respectively, in a 5-step process using Amberlite XAD16N beads, SP sepharose, C18 SPE and two RP HPLC purification steps with a yield of 0.725 mg per liter of E. coli/pETcoco-pedM31L culture and 0.625 mg per liter of P. acidilactici LMG 2351 culture. Purity of >95% for pediocin PA-1 (˜4624 Da) and M31L (˜4007 Da) were confirmed by MALDI-TOF mass spectrometry (data not shown).

Comparison of pediocin PA-1, pediocin M31L and penocin A antimicrobial activity
The effect of pure pediocin PA-1, pediocin M31L and penocin A on L. monocytogenes 10403S was evaluated in BHI soft agar (0.8% agar) (Figure 2d,e,f,) and BHI broth over 24 h at 37 °C (Figure 2d,e,f,). Figure  penocin A (Figure 2d,e,f,, respectively). These conditions showed significantly different OD 600nm compared to the bacteriocin-free control at OD 600nM of 0.4 (one-way ANOVA followed by Tukey's post hoc test p < .05). The antimicrobial activity of these bacteriocins against selected human gut commensal strains was also evaluated. Faecalibacterium prausnitzii A2-165, Eubacterium rectale A1-86, Roseburia inulinivorans A2-194, Akkermansia muciniphila MucT and Ruminococcus bromii VPI 6883, originally isolated from human feces, 24 have been used as indicators to elucidate if bacteriocins effective against L. monocytogenes have a detrimental effect on the commensal microbes. These bacteria are abundant species in a normal healthy gut and have also gained special interest as next generation health-promoting species. 25 No inhibition was observed against these strains when 1 µM of pediocin PA-1, 1 µM pediocin M31L and 1 µM penocin A were tested ( Figure 3). Antimicrobial activity was observed against the control strain, L. innocua DPC3572, in the presence of 1 µM of pediocin PA-1, 1 µM pediocin M31L and 1 µM penocin A. In addition, ampicillin (10 µg) or chloramphenicol (30 µg) resulted in a zone of inhibition against all strains assessed. This assay was performed in two independent assays.

Ex vivo model assays
The ability of natural pediocin PA-1, recombinant M31L and synthetic penocin A to inhibit L. monocytogenes 10403S in a simulated human gut environment was assessed using 6.5 Log CFU/mL of the bioindicator strain inoculated in fecal media and slurry and fermented in a bench top bioreactor. Listeria numbers increased from 6.5 � 0.13 Log CFU/mL (T0h) to 8.56 � 0.16 Log CFU/mL and qPCR using primers to amplify a variable region of iap gene (Figure 4b) were also used to quantify Listeria. This gene encodes p60 extracellular protein produced by all Listeria species, but a specific region of the gene can be used to differentiate between L. innocua and L. monocytogenes species.- 26 Figure 4b shows that the Log of iap copy number/mL increased from 7.58 � 0.126 Log iap copy/mL (T0h) to a mean of 8.67 � 0.17 Log iap copy/mL after 5 h of fermentation and subsequently reduced to 8.45 � 0.18 Log iap copy/mL (T24h) in the bacteriocin-free control. In the presence of pediocin PA-1, the recombinant M31L and penocin A, iap amplification was not detected. It is important to highlight that the detection limit of this qPCR assay is 4.5 Log iap copy/mL.

Discussion
Heterologous expression of pediocin PA-1 has been achieved previously in different E. coli strains (BL21 (DE3), M15/pRep4, Origami (DE3), V850) but, to our knowledge, significant quantities of peptide were not obtained in a pure form from E. coli. 27,28 Despite the ability to obtain recombinant pediocin using affinity chromatography, a further purification step to remove the fusing protein or HIS-tag results in a considerable loss of bacteriocin yield typically, i.e. 0.06 mg per 100 mL of culture. 28 Essential genes, necessary for the heterologous expression of mature pediocin in the extracellular fraction in E. coli, were identified by Mesa-Pereira et al. 22 However, pediocin PA-1 production was considerably lower compared to that obtained by the natural producer P. acidilactici LMG2351, probably due to oxidation of recombinant pediocin PA-1 as evidenced by a second peptide mass of ~4640 Da.
Pediocin PA-1 instability due to methionine oxidation has been widely reported. 20  peptide. 20 Therefore, in this study, a pediocin M31L derivative was expressed for the first time in E. coli. This microorganism is widely used for heterologous expression due to the availability of well-established molecular tools and the ability to grow E. coli in numerous inexpensive and/or mineral media, thereby minimizing costs. 30 The yield of pediocin M31L (0.725 mg/Liter of culture supernatant) obtained in this study with >95% of purity is reasonable when compared with other bacteriocin studies, 19 and there is potential to further improve yield. It is worth noting that scale up of pediocin M31L production in E. coli requires further evaluation taking into account cost benefit analysis and other variables such as substrate selection. Indeed, mineral media is often used for E. coli heterologous expression and could lead to a reduction in the number of purification steps, potentially increasing yield as mature pediocin is present in the supernatant fraction.
In this study, pediocin M31L displayed similar anti-Listeria activity to native pediocin PA-1. When the antimicrobial activity of the three pediocins was assessed against Listeria in broth media, a bacteriostatic effect was observed even at higher bacteriocin concentrations. Mechanisms of bacteriocin resistance and bacteriocin tolerance are not fully studied but have been described within all classes of bacteriocin. 31 Bacteriocin resistance can be both acquired and innate, and the main resistance mechanism for class IIa bacteriocins is the downregulation of the expression of the mannose phosphotransferase system (Man-PTS), which has been described for E. faecalis and L. monocytogenes. [32][33][34] The regulatory gene rpoN also influences the mpt expression and consequently influences the development of resistance. 35 Metabolism shifts have also been observed when pediocin PA-1-sensitive and -resistant E. faecalis were compared 36 and with L. monocytogenes bacteriocinresistant strains. 37,38 Upregulation of genes involved in the usage of sugars may play a role in bacteriocin resistance, and recently, it has shown that cellobiose and sucrose increased the generation of resistant L. monocytogenes when compared to other carbohydrates tested. 39 Additionally, the composition of the food matrix should be considered as it could favor the development of bacteriocin resistance and at the same time modify the sensory properties of the product by inducing the growth of the spoilage bacteria.
Undoubtedly, studies carried out with the pure form of these bacteriocins lead to more accurate assessments of the dose needed to observe an anti-Listeria effect than those carried out with bacteriocin containing fermentates or bacteriocin-producing cultures. The greater precision in the amount of antimicrobial compound required for efficacy is essential to design more effective and reproductive protocols, reducing the chances of bacteriocin-resistant strain appearance. Certainly, more controlled use of emerging food additives will be increasingly requested for regulatory agencies.
The antimicrobial activity of bacteriocins on gut commensal bacteria is also of critical importance as strains from species such as A. muciniphila, F. prauznitzii and Eubacterium spp. have gained attention due to their health-promoting effects and the potential to serve as next-generation probiotics or biotherapeutics. 40 In the present study, none of the pediocin variants tested displayed activity against commensal bacteria, though undoubtedly further experiments should be carried out using encapsulated pediocins in controlled clinical studies. Previous investigation whereby CFS from LAB strains containing nisin A, nisin Z or pediocin PA-1 41 were assessed against different commensal bacteria. While both nisin variants displayed activity against known commensal bacteria such as Eubacterium biforme DSM 20477, Ruminococcus productus DSM 2950 and many species of Bifidobacterium and Lactobacillus, none of the tested strains were inhibited by pediocin PA-1, indicating that this bacteriocin might be a more appropriate treatment for gut infections as no disturbance of the intestinal bacterial balance is likely to occur.
Currently, no bacteriocin is used to treat gut bacterial infections despite the unquestionable evidence of their efficacy. Simulated gut model assays performed, at a concentration of 2.6 µM of pediocin PA-1, pediocin M31L or penocin A, resulted in total inhibition of Listeria after 5 hours of fermentation, and, even after 24 h of fermentation, Listeria counts were considerably lower (>4 Log CFU/mL) than the control. In this regard, it is noteworthy that Dabour et al. 16 observed in a mice model challenged with L. monocytogenes LSD348 that three intragastric doses of pure pediocin PA-1 (250 μg) led to almost a 2 log reduction in Listeria counts in fecal samples, while only a slight anti-Listeria effect was observed when the pediocin producer strain P. acidilactici UL5 was administrated, reinforcing the importance of antimicrobial concentrations and potential challenges associated with the use of producing strains in treating a specific pathogen.
Undoubtedly, there are a number of hurdles to overcome with respect to the use of these pediocins for clinical applications. In this work, the use of a microbioreactor to simulate colon conditions proved to be a very reproducible and reliable platform to verify the listericidal effect of the studied bacteriocins. Encapsulation employed can protect the bacteriocins from protease breakdown in the gastrointestinal tract, while combining pediocins with other antimicrobial compounds also merits consideration. The evaluation of different pediocin carriers (nanocrystals and nanofibers of cellulose or nanostructured lipid) with different dosages of pediocin still needs to be tested prior to in vivo and clinical trials. The use of 24-well microbioreactors has been widely applied 42,43 and would accelerate the optimization of pediocin as a therapeutic product. Regulating the use of a new drug is a very complex and laborious process, and optimizing the production of a more stable form of pediocin and applying it to initial ex vivo model studies was a notable achievement in this study and forms the basis of future more targeted clinical studies.

Bacterial strains, culture conditions and plasmid constructions
The pediocin PA-1 structural gene (pedA), containing a codon change to introduce a methionine to leucine substitution at position 31 (M31L), and the associated accessory (pedC) and transport genes (pedD) were synthesized (Eurofins MWG Operon Inc., Ebersberg, MU, Germany) with codons optimized for expression in E. coli. This DNA fragment was amplified by polymerase chain reaction (PCR) using Platinum® Taq DNA Polymerase (Thermo Fisher Scientific) and primers containing SphI and AvrII restriction sites (5ʹ-3ʹCCTGCATGCAAATGAAGAAAATCGAA AAGC and CATACCTAGGCTAGGTCACTCCT GATTATGA). PCR-generated fragments and the modified pETcoco-2 expression vector 22 (Novagen, EMD Millipore, Billerica, MA, USA) were digested with SphI and AvrII, and the resulting fragments were ligated with T4 DNA ligase (Thermo Fisher Scientific). One-shot TOP10 cells (derivatives of E. coli DH10β) were used for standard cloning procedures, and transformants containing pETcoco-pedM 31L were confirmed by PCR amplification and sequencing. PCR clean up steps and plasmid extractions were performed with NucleoSpin® Extract II Kit and NucleoSpin® plasmid kits (Macherey-Nagel, Duren, CO, Germany), respectively.
E. coli (DE3) Tuner TM cells (Novagen, EMD Millipore) were transformed with pETcoco-pedM 31L, and the recombinant expression was carried out in Luria-Bertani (Difco Laboratories, Detroit, MI, USA) medium supplemented with ampicillin (50 μg/mL). Cells were grown to OD 600nm 0.5-0.6 and were subjected to different temperatures (25 °C or 37 °C), isopropyl β-d-1-thiogalactopyranoside (IPTG) (Sigma-Aldrich Inc., Arklow, Ireland) concentrations (100 or 1000 μM) and periods of induction (3 h, 6 h or overnight). Plasmid copy number was controlled by the addition of 0.2% of D(+) glucose (Sigma-Aldrich Inc.) for low copy state or 0.01% L-+-Arabinose (Sigma-Aldrich Inc.) for medium copy state. To determine antimicrobial activity, E. coli/pETcoco-pedM31L supernatants were collected by centrifugation at 10,000 x g for 15 min at 10°C and assessed against Listeria innocua DPC3572 via the well-diffusion method. Quantification of pediocin antimicrobial activity was measured according to Mesa-Pereira et al. 22 The minimal inhibitory concentrations of pure (>95%) pediocin PA-1, pediocin M31L or penocin A were assessed in soft media (0.8% agar) or broth Brain Heart Infusion (BHI) media using Listeria monocytogenes serotype 1/2a 10403S as a bioindicator. Initial bacteriocin concentrations were determined by QuantiPro™ BCA Assay Kit (Sigma-Aldrich Inc.) and then serially two-fold diluted in 0.1% trifluoroacetic acid (TFA). The resulting samples were assessed in BHI soft media or broth containing ~10 6 CFU 10403S cells (~10 6 CFU). In soft media, 50 μL bacteriocin-containing samples were added to each well (5 mm diameter) and incubated at 37 °C for 24 h, whereas in broth, Listeria growth was monitored at 600 nm every 30 min at 37 °C for 24 h using a Synergy HT plate reader (BioTek Instruments Inc, Winooski, VT, USA). The antimicrobial activity of these pediocins against specific human gut commensal strains was evaluated in anaerobic conditions (80% N 2 , 10% H 2 , 10% CO 2 gas mixture) in pre-reduced agar form of YCFAGSC medium. 44

Purification of pediocin PA-1 from P. acidilactici LMG 2351 and pediocin M31L from recombinant E. coli
Pediocin PA-1 and pediocin M31L were purified as follows: 800 mL of P. acidilactici LMG 2351, a natural pediocin PA-1 producer, 45 or E. coli/pETcoco-pedM31L cells were harvested by centrifugation at 10,000 x g for 15 min at 10 °C and the supernatant loaded onto an Econo column containing 60 g Amberlite XAD16N beads (Phenomenex, Macclesfield, CH, UK) prewashed with Milli Q water. The column was washed with 250 ml 30% ethanol and antimicrobial activity eluted with 250 ml 70% propan-2-ol (IPA) 0.1% TFA. To perform the ion exchange purification step, the IPA was removed from the eluent sample via rotary evaporation and the pH adjusted to approximately 4.4. The resulting sample was applied to an Econo column containing 60 ml SP-Sepharose beads (GE Healthcare, Bio-sciences AB, Uppsala, UP, SWE) pre-equilibrated with 250 ml 20 mM sodium acetate pH 4.4. The column was washed with 60 ml of 20 mM sodium acetate pH 4.4 and antimicrobial activity eluted with 250-300 ml 20 mM sodium acetate pH 4.4 containing 1 M NaCl. The saltcontaining eluent from the SP-sepharose column was applied to a Strata-E C18 SPE column (Phenomenex) pre-equilibrated with methanol and water. The column was washed with 25% ethanol and antimicrobial activity eluted with IPA.
The sample was further purified by two Reversed Phase High Performance Liquid Chromatography (RP-HPLC) steps. The IPA was removed from the C18 SPE eluent by rotary evaporation and applied to a semi-preparative Jupiter Proteo (10 x 250 mm, 90 Å, 4 µm) RP-HPLC column (Phenomenex) running a 25-40% acetonitrile 0.1% TFA gradient. The eluent was monitored at 214 nm, and fractions were collected at 1-minute intervals. HPLC fractions were tested against L. innocua DPC3572, and the fractions with the highest antimicrobial activity were submitted to a final purification step on the same semi-preparative RP-HPLC column with a shallower 30-40% acetonitrile gradient. Fractions deemed pure by MALDI-TOF mass spectrometry (Shimadzu Biotech, Manchester, CO, UK) were pooled and lyophilized (Genevac, Ipswich, SU, UK).

Human distal colon model assays
Distal colon model assays were approved by the Clinical Research Ethics Committee of the Cork Teaching Hospitals (protocol no. APC 104). Human fecal inoculum was prepared as previously described by O´Donnell et al. 42 Briefly, each standardized fecal inoculum consisted of eight healthy donors who had not taken an antibiotic for 6 months prior to donating a fecal sample and did not harbor any significant acute or chronic illness. Ex vivo fermentation was carried out in bench top bioreactors micro-Matrix (Applikon Biotechnology, Heertjeslaan, DE, NE) according to the protocol described by O´Donnell et al. 42 for each well: 400 µL of slurry, 6.5 log CFU/mL of L. monocytogenes 10403S, 2.6 µM of purified pediocin PA-1, pediocin M31L or penocin A were inoculated in fecal media 46 up to a final volume of 5 mL. A well containing the bioindicator strain alone (no bacteriocin) was used as a control. Each treatment was performed in two independent assays. Samples were collected during the fermentation at time 0 (T0h), 5 (T5h) and 24 hours (T24h) and were serially diluted in maximum recovery diluent (MRD) (Oxoid Ltd., Basingstoke, HA, UK) and plated on Listeria Chromogenic Media supplemented with Polymyxin and Ceftazidime (Cruinn Diagnostics Ltd., Dublin, DU, IRE) at 30 °C for 48 h.
For total DNA extraction, 1 mL of fermentation sample was extracted using QIAmp Power Fecal DNA Kit (QIAGEN, Crawley, UK) according to the manufacturer's recommendation. Presence of Listeria was quantified through qPCR using iap as a target as described by Hein et al. 26 For statistical significance analysis, two-way ANOVA followed by Tukey's post hoc tests were performed using Minitab 17 Statistical Software (Minitab Inc., State College, PA, USA) (www.minitab.com).

Acknowledgments
This work was funded by São Paulo Research Foundation (FAPESP) and APC Microbiome Ireland, a research centre funded by Science Foundation Ireland (SFI), through the Irish Government's National Development Plan. We are also grateful to BioRender.com where Graph Abstract was created.

Data availability
The data that support the findings of this study are openly available in figshare.com at http://doi.org/10.6084/m9.figshare. 16908805.

Disclosure statement
No potential conflict of interest was reported by the author(s).