Precision modification of the human gut microbiota targeting surface-associated proteins

This work describes a new procedure that allows the targeted modification of the human gut microbiota by using antibodies raised against bacterial surface-associated proteins specific to the microorganism of interest. To this end, a polyclonal antibody recognising the surface-associated protein Surface Layer Protein A of Lactobacillus acidophilus DSM20079T was developed. By conjugating this antibody with fluorescent probes and magnetic particles, we were able to specifically identify this bacterium both in a synthetic, and in real gut microbiotas by means of a flow cytometry approach. Further, we demonstrated the applicability of this antibody to deplete complex human gut microbiotas from L. acidophilus in a single step. L. acidophilus was found to interact with other bacteria both in synthetic and in real microbiotas, as reflected by its concomitant depletion together with other species. Further optimization of the procedure including a trypsin step enabled to achieve the selective and complete isolation of this species. Depleting a single species from a gut microbiota, using antibodies recognizing specific cell surface elements of the target organism, will open up novel ways to tackle research on the specific immunomodulatory and metabolic contributions of a bacterium of interest in the context of a complex human gut microbiota, including the investigation into therapeutic applications by adding/depleting a key bacterium. This represents the first work in which an antibody/flow-cytometry based application enabled the targeted edition of human gut microbiotas, and represents the basis for the design of precision microbiome-based therapies.


Results
Detection of L. acidophilus using a polyclonal anti-SlpA antibody. For enrichment and depletion of L. acidophilus from a synthetic or real microbiota, we developed a protocol based on using antibodies and magnetic beads. The polyclonal anti-SlpA antibody was fluorescently labelled with fluorescein isothiocyanate (FITC) or allophycocyanin (APC). The availability of an anti-APC antibody conjugated to magnetic beads allows the targeted isolation of those L. acidophilus cells that have been bound to the antibodies (see Material and Methods for details).
Eight Lactobacillus strains, Lactobacillus casei ssp. rhamnosus GG, Lactobacillus acidophilus DSM20079 T , Lactobacillus amylovorus B13, Lactobacillus casei 393, Lactobacillus delbrueckii ssp. delbrueckii IPLAlb101, Lactobacillus reuteri DSM20016 T , Lactobacillus gasseri BM7/10, and Lactobacillus plantarum NCIMB 8826, were chosen as being representative of the Lactobacillus genus. Bifidobacterium longum subsp. longum and Escherichia coli were also considered as outgroups representing other bacterial phyla commonly present in the human gut microbiota. The labelling of the ten bacterial species with the anti-SlpA polyclonal antibody was evaluated by FC and analysis of dispersion diagrams. These were obtained by separating the bacteria in FC according to Side Scatter (SSC), Forward Scatter (FSC) and their fluorescence in the FITC channel, designating three regions in the plots corresponding to P1 (bacterial cells); LR2 (unlabelled bacteria) and UR2 (fluorescent bacteria). The anti-SlpA demonstrated a very specific recognition of L. acidophilus DSM20079 T cells, with virtually 100% of bacteria labelled after 15 min of incubation (Fig. 1A). The rest of the bacteria showed labelling percentages lower than 15% as deduced by FC, and showed no fluorescence at the confocal microscope ( Supplementary Fig. 1).

Specificity of polyclonal antibody against L. acidophilus surface layer protein A. As polyclonal
anti-SlpA antibody bound to L. acidophilus cells more efficiently than to the rest of species tested, even than to those containing S-layer proteins such as L. gasseri, we performed further analyses to determine which part of the cells was selectively recognized by the antibody and to confirm whether the antibody could recognize any structures in other bacterial species. Immunofluorescence microscopy of the ten bacterial species previously stained with FITC-anti SlpA antibody, showed that antibody binding was located on the surface of L. acidophilus cells, whereas fluorescence was undetectable in the rest of species assayed in this work. This contrasted with the 10-15% of fluorescent cells detected in the FC experiments for those species assayed, events that even while considered positive, presented fluorescence emission values drastically lower than L. acidophilus (Fig. 1A). The higher emission of L. acidophilus compared to the lower emissions of the rest of bacteria can be easily deduced comparing the positive events displayed in the dispersion plots of L. acidophilus DSM20079 T (98.29 ± 0.99%, Fig. 2A) to those of L. rhamnosus GG (11.47 ± 2.74%, Fig. 2B). The dispersion plots for other tested bacteria are included in Supplementary Fig. 1. The response of the bacteria to the pre-immune serum was also evaluated by FC, returning values lower than 8% in all bacteria ( Supplementary Fig. 2). The specificity of the polyclonal anti- www.nature.com/scientificreports/ body was also evaluated through Western blotting, using the extracellular protein fractions of the eight bacteria included in the synthetic microbiota, showing a sharp blot of the right molecular weight with the L. acidophilus fraction, but not with the other bacterial fractions ( Supplementary Fig. 3).
Detection of L. acidophilus in a synthetic gut microbiota using flow cytometry. Given the high specificity of the anti-SlpA antibodies, we tried to specifically label the L. acidophilus DSM20079 T strain when included in a defined bacterial consortium containing representatives of the major human intestinal bacterial  When all bacteria were added at 1:1 proportion, and therefore L. acidophilus represented around 12.5% of bacteria present in the mix, L. acidophilus population was evident in the dispersion plots and, in those mixes including percentages of L. acidophilus from 12.5% to 58.82%, the proportion of events displaying fluorescence in the FITC channel increased with the L. acidophilus representation in the mix (Fig. 3A). The dispersion diagrams of the bacterial community mixes tested are included in Enrichment and depletion of L. acidophilus from a synthetic microbiota. In the next step, we combined the use of the anti-SlpA IgG polyclonal antibody conjugated to APC, with an anti-APC antibody conjugated to magnetic beads to specifically enrich L. acidophilus in a synthetic microbiota including the bacterial strains previously listed as representatives of B. longum, E. coli, Bl. obeum, F. praustnizii, L. gasseri, B. thetaiotaomicron, A. muciniphila and L. acidophilus, each bacterium representing a 12.5% of the mix. Using this approach, we were able to deplete the synthetic microbiota from L. acidophilus DSM20079 T cells in a single step (Fig. 4A, left panel), although the positive fraction retained in the magnetic column contained L. acidophilus together with L. gasseri and A. muciniphila cells, as evidenced by culture-dependent techniques and further 16S rRNA gene-based identification (Fig. 4B). Cell counts evidenced that 8.5 ± 2.4 × 10 5 A. muciniphila and 1.4 ± 1.2 × 10 6 L. gasseri cells were recovered after positive L. acidophilus enrichment, which indeed represented two logs lower than the recovered L. acidophilus concentration (1.0 × 10 8 ). The presence of these bacteria in the positive fraction was evident during FC analysis as they appeared as events that did not emit fluorescence when analysed in the APC emission channel (Fig. 4C). In a previous work, we showed how bacterial clumps characteristic of stationary-phase cultures of L. rhamnosus GG were efficiently disaggregated by trypsin 26 . In the same way, herein the inclusion of a single 15 min trypsin treatment prior to the magnetic selection protocol, resulted in the selective and exclusive isolation of L. acidophilus cells in a single step, as revealed by a very limited presence of non-fluorescence events in the positive fraction, and confirmed by culture-dependent and 16S rRNA analyses (Fig. 4D).
Detection, enrichment and depletion of L. acidophilus in a real gut microbiota. As a proof-ofprinciple validation, the application of the developed antibody and above described procedure to detect, enrich and deplete L. acidophilus on synthetic microbiotas was validated on real human gut microbiota samples. For this purpose, five samples of real human gut microbiotas obtained from healthy adult donors were supplemented in an approximate proportion of one L. acidophilus per ten microbial cells, as explained in the Material and Methods section. By using this approach, the population of L. acidophilus cells were perfectly discriminated from the rest (and non-fluorescent) microbiota cells ( Fig. 5; Supplementary Fig. 5). www.nature.com/scientificreports/ Enrichment and depletion of L. acidophilus from the fecal microbiotas (n = 5) also followed the same trend as the synthetic microbiota ( Fig. 5; Supplementary Fig. 5). Fecal microbiotas were depleted from L. acidophilus in a single step ( Fig. 5C and Supplementary Fig. 6), although plating of these positive fractions evidenced the co-isolation (around 2 logs less in terms of colony numbers) of different isolates showing 16S rDNA sequence homology to Streptococcus salivarius (92%), St. gallolyticus (91%) and Tumebacillus flagelatus (94%) (data not shown). In order to disaggregate the L. acidophilus DSM20079 T cells from the rest of associated bacteria, a single 15-min trypsin pre-treatment was applied prior to the magnetic separation as previously described. With this technique, positive fractions contained around 95% of L. acidophilus cells as determined by FC analysis. However, a minimal number of fluorescent events were also retrieved in the negative fraction (Fig. 5D). The positive and negative controls were L. acidophilus DSM20079 T labelled and unlabeled (Fig. 5E). As additional validation, confocal microscopy corroborated that trypsin treatment did not disrupt the antibody interaction (Fig. 5F).

Microbial profile of positive and negative fractions.
To further check the relative abundance of L. acidophilus in the negative (or depleted) and positive (or enriched) fractions of gut microbiota, a 16S rRNA gene sequencing analysis was performed on fractions obtained following the procedure described above, on five different microbiotas supplemented with L. acidophilus DSM20079 T . The results show that, in all the microbiotas analysed, the genus Lactobacillus was significantly enriched in the positive fraction, and depleted in the negative

Discussion
The identification of specific surface-associated proteins and development of antibodies specifically recognizing them, may contribute to the conception of FC approaches in the field of microbiology. As a proof of concept, we applied this methodology to the L. acidophilus species, showing that is not only possible to detect a specific species in a complex microbiota, but also to perform a targeted depletion, and even a further enrichment, in a single step.
To this end, we have analyzed how a polyclonal antibody raised against the surface-associated S-layer protein A of L. acidophilus can be applied in its detection through a FC approach. Besides, confocal laser microscopy analysis evidenced that, from the bacterial species tested in this study, only L. acidophilus cells were efficiently labelled with this serum. Remarkably, some positive events were identified through FC for those strains that evidenced no labelling through immunofluorescence microscopy. This is important as the polyclonal antibody presented in this work may contain other antibodies recognizing other bacterial targets that are usually present in the surface of the gut bacteria. The higher sensitivity of FC with respect to the fluorescence microscopy represents www.nature.com/scientificreports/ one of its main advantages, and generation of a monoclonal antibody would decrease the false positive rates obtained with our polyclonal anti-SlpA antibody 27 . The polyclonal antibody against the surface associated S-layer protein A also allowed for the detection of L. acidophilus cells when present in a synthetic microbiota comprised of 8 representative gut bacterial species. Furthermore, the same antibody enabled the specific depletion of this bacterium from the community in a single step. Applying the same methodology over fecal microbiotas, we showed that it is possible to deplete a complex gut microbial population from L. acidophilus cells in a single step and, that by using a pre-treatment with trypsin, we could prevent the concomitant isolation of other species likely associated with the target microorganism within the community. These findings are very relevant as, for the first time, a procedure for obtaining a kind of "Knock-out" microbiota depleted of one specific bacterial species, is described. This may represent a novel methodology that will aid to understand the specific contribution of a single bacterium to a given phenotype or  www.nature.com/scientificreports/ effect associated with a microbial community. However, such approaches would benefit from the development of more specific and monoclonal antibodies. Many diseases are associated with alterations in the gut microbiome, including diabetes, obesity, inflammatory bowel disease, atherosclerosis, alcoholic liver disease, non-alcoholic fatty liver disease, cirrhosis cancer or metabolic syndrome [28][29][30] . Changes in the diet, use of probiotics, prebiotics or antibiotics are common means to achieve microbiota modifications, although they are not very selective, at least at species level 31 . In a recent work, tungstate was used to selectively limit the expansion of the members of the Enterobacteriaceae family, which is characteristic from active inflammatory bowel disease 32 . Also, the work of Hsu and colleagues provides an insight into the ecological importance of phages as modulators of bacterial colonization, and it additionally suggests the potential impact of gut phages on the mammalian host with implications for their therapeutic use to precisely modulate the microbiome 33 . The technique presented in this work represents a novel methodology for the precise modification of a microbiota, in this case targeting the L. acidophilus species. Our results showed that it is possible to deplete a microbiota from L. acidophilus cells in a single step using a pre-treatment with trypsin. This method could be applied to the majority of L. acidophilus strains due to the high similarity and conservation of the SplA between the strains of this species (Supplementary Fig. 7). Other Lactobacillus strains possess an S-layer such a, L. crispatus, L. helveticus or and L. gasseri, and these proteins are also present on the surface of other Bacteria and Archaea representatives 34 . The results of the 16S rRNA gene sequencing analysis show that L. acidophilus is enriched in the positive fraction, whereas the percentage of L. acidophilus that remains in the negative fraction is minimal. The described procedure could likely be further optimized to achieve a percentage of L. acidophilus in the negative fraction closer to 0% by increasing the anti-APC used, since the concentration of antibody and beads influences the depletion capacity.
Theoretically, any bacteria presenting specific targets on their surface, could be precisely depleted from a gut microbiota following the protocol described in this paper, by employing antibodies raised specifically against selected particular protein markers from the target bacterial species. For any bacteria of interest, the application of a methodology like the one described in this work would firstly need to detect major proteins present on the bacterial surface by using, for instance, SDS-PAGE or surface shaving methodologies 35 . The procedure herein described is clearly a different and more specific technique than other procedures for targeted microbiota modification described to date, such as metagenomic engineering through conjugative transfer 36 . The implications of this technique in microbiome-based therapies may be of great importance. For instance, the absence of A. muciniphila has been associated with resistance to PD-1 chemotherapy in melanoma 37,38 , lower abundances of F. prausnitzii are associated with Inflammatory Bowel Disease, and presence of specific bacteria such as E. coli Pks + , Bacteroides fragilis Btf + and Fusobacterium nucleatum is related to colo-rectal cancer 39,40 . Targeted modification of a gut microbiota will enable depletion and/or enrichment of native microbiotas with these bacteria, for instance eliminating specific E. coli strains associated with certain conditions, or enriching commensal beneficial bacteria such as F. prausnitzii on individual fecal microbiotas, with a clear potential application in gut microbiota auto-transplant.
To date, most studies use probiotics or prebiotics to restore the composition of the gut microbial community and modify the functions of the microbiome and the phenotype of the host. Such approaches, however, do not allow to study the biological effect derived from the removal of a particular bacterium from the microbiota 41 . The methodology described in this work represents a novel technique providing proof of concept for the detection, enrichment and depletion of specific bacteria in a bacterial consortium, or in a more complex microbiota, without the requirement of possessing a FACS (Fluorescence-Activated Cell Sorting) sorter in the research facilities. By facilitating precision modification of a microbiota, this new approach will be of great relevance to those research groups interested in the study of the contribution of single species to different diseases, physiological, metabolic or immune processes, including fecal microbiota transplantation. We are currently using our methodology to deplete or eradicate intestinal microbiota from strains/species and applying this model to the study of how they contribute to a certain disease.

Material and methods
Ethics statement. Ethics approval for this study (reference code AGL2016-78311-R) was obtained from the Bioethics Committee of CSIC (Consejo Superior de Investigaciones Científicas) and from the Regional Ethics Committee for Clinical Research (Servicio de Salud del Principado de Asturias; no. 127/16) in compliance with the Declaration of Helsinki. All determinations were performed with fully informed written consent from all participants involved in the study.
Fecal sample collection and microbiota extraction. The study sample comprised 5 fecal samples from healthy donors. No donors reported any severe diseases in the last 6 months. Fresh fecal material was collected in a sterile container and immediately manipulated and homogenized within a maximum time span of 2 h. Nine millilitres of sterile NaCl 0.9% (w/v) was added to 1 g of sample, and the mixture was homogenized in a sterile bag, using a laboratory paddle blender (Stomacher Lab Blender 400, Seward Ltd. UK) for 1 min. Microbiota extraction was then performed following the protocol described by Hevia and coworkers 42   Purification of the surface-layer protein A. Purification of the Surface-Layer Protein A was performed after separation in SDS-PAGE from the extracellular protein fraction of the L. acidophilus DSM20079 T strain 43 . Fifty mL aliquots of fresh MRSC broth were inoculated (1% v/v) from an overnight culture. Cultures were allowed to enter the stationary phase of growth and cells were harvested by centrifugation (9300×g, 4 ºC, 10 min). Supernatants were then filtered (0.45 μm) and extracellular proteins were precipitated by adding a final concentration of 6% (w/v) trichloroacetic acid (TCA; SIGMA). Extracellular proteins were allowed to precipitate at 4 °C for 2 h. Proteins were recovered by centrifugation (9300×g, 4 °C, 10 min); pellets were washed twice with 2 mL of chilled acetone (SIGMA) and dried at room temperature. Proteins were resolubilized by ultrasonication (Ultrasonic bath; Deltasonic, Meaux, France) in 200 μL of 1 × Laemmli buffer for 10 min 44 . Bands corresponding to SlpA were excised from the gels, pooled and used for polyclonal antibody generation.

Polyclonal antibody generation.
A polyclonal serum against the purified L. acidophilus DSM20079 T SlpA was generated at the Central Facilities of the University of Oviedo (Spain) as already described 45 . A rabbit was immunized five times, with an interval of 15 days between immunizations, with around 500 µg of protein dissolved in 1 mL of PBS, and mixed with 1 mL of Freund's incomplete Adjuvant. The rabbit was finally sacrificed by intracardiac puncture and blood was let to coagulate at 37 ºC for 4 h and subsequent overnight incubation at 4 ºC. Serum was separated by centrifugation (30 min, 2000 × g), and used for purifying the IgG. Firstly, ammonium sulphate was added to a final concentration of 45% (w/v), and the mix was incubated overnight at 4 ºC. After centrifugation (1 h, 10,000 × g, 4 ºC), the pellet was resuspended in 30 mL of PBS. This was extensively dialyzed against PBS, and loaded in a ProteinA Sepharose 4 Fast Flow, previously equilibrated with 10 column volumes of PBS (50 mL). The column was washed with 6 column volumes of PBS, and five fractions of 5 mL were eluted with citric acid 100 mM pH 3.0. pH was corrected in each aliquot by adding 1 mL of 1 M Tris-HCl pH 9.0. Fractions were mixed, centrifuged in a Vivaspin 20 device (3000 × g, molecular weight cut-off of 10 kDa) and washed with 20 mL of PBS. Protein concentration was estimated by measuring the A 280 of the sample, aliquoted and stored at − 80 ºC. The response of the bacteria used in this work to the pre-immune serum was evaluated by direct labelling of the serum with FITC or APC. Labelled and unlabelled bacteria, bacterial mixes and microbiotas were acquired and analysed by MACSQuant flow cytometer (MILTENYI BIOTEC). Antibody conjugation. The Polyclonal anti-SlpA serum IgG fraction was conjugated with FITC or APC using specific conjugation kits (Abcam Cambridge, MA, USA) and following manufacturer's instructions. FITC is a green fluorescent dye that has an excitation wavelength maximum of 495 nm and an emission wavelength maximum of 519 nm. APC is an intensely bright phycobiliprotein isolated from red algae that exhibits far-red fluorescence with high quantum yields, with an absorbance maximum at 650 nm and a fluorescence emission peak at 660 nm. For FITC and APC conjugation, the anti-SlpA IgG polyclonal antibody fraction was reconstituted in phosphate buffered saline (PBS). One hundred microliters of a 1.5 mg/mL antibody dilution were added to the reactive dye for each conjugation event. Ten µL of FITC-Modifier reagent or APC-Modifier reagent was added to the antibody before adding the FITC or APC. The antibody-dye mixtures were incubated in the dark at room temperature for 3 h. After incubation, 10 µL of FITC-Quencher or APC-Quencher reagent was added and mixed gently. The concentration of conjugated antibody in the final sample was 20 µg/mL.

Flow cytometry analysis. Labelled cells/microbiotas with anti-SlpA polyclonal antibodies were acquired
and analysed in a MACS Quant Flow Cytometer device (MILTENYI BIOTEC) using the following acquisition parameters: flow rate set to "low", uptake volume of 10 µL, FSC set to hyperlogarithmic amplification (438 V), SSC set to hyperlogarithmic amplification (535 V), channel B1 corresponding to the FITC detection set to hyperlogarithmic amplification (331 V) and channel R1 corresponding to the APC detection set to hyperlogarithmic amplification (360 V). At least 10,000 events were acquired in each run. At least eight independent experiments were conducted with the exception of the five intestinal microbiotas. For each species, one millilitre of cultures in exponential phase of growth (OD 600 ≈ 0.7) was centrifuged at 10,000×g for 5 min. After that, bacterial suspensions were adjusted to an OD 600 = 0.2 (around 1E 8 CFU/mL) using FC buffer. Twenty-five µL of bacterial suspensions were mixed with 25 µL of the FITC-conjugated antibody at a final concentration of 20 µg/mL. The samples were incubated for 15 min at room temperature and were then washed with bacterial FC buffer. Finally, bacteria were resuspended in 150 µL of FC buffer and aliquots of 10 µL were further analysed by FC.

Scientific Reports
Detection of L. acidophilus in a gut microbiota. The detection of L. acidophilus in a mix of bacteria was investigated over a synthetic microbiota and over five samples of real gut microbiotas isolated from healthy donors.
Finally, both synthetic and real microbiotas were labelled with anti-SlpA polyclonal antibody conjugated to FITC, and data was acquired and analysed by MACS Quant flow cytometer, as indicated previously.
Specific enrichment and depletion of L. acidophilus from a synthetic or a real microbiota. The enrichment and depletion of L. acidophilus from a mix of bacteria supplemented with L. acidophilus DSM20079 T cells in a volumetric 1:1 ratio (12.5% of each bacteria) configured with cell suspensions at an OD 600 ≈ 0.2, or real gut microbiotas adjusted to OD 600 ≈ 5.0, were tested. A schematic diagram of the enrichment and depletion method of L. acidophilus from a synthetic microbiota is represented in Fig. 1B. Mixes were labelled with the polyclonal anti-SlpA serum IgG fraction conjugated to APC and incubated for 15 min at room temperature. Then, mixtures were washed at 13,000 × g for 5 min and the supernatants removed. For positive selection of L. acidophilus, bacterial mixes were resuspended in 100 µL of resuspension buffer [RB: PBS supplemented with 2 mM EDTA and 3% (v/v) of de-complemented bovine foetal serum]. To this mix, 10 µL of magnetic anti-APC particles were added (BD BIOSCIENCES, San José, USA). The mixture was incubated for 15 min at 4 °C. Then, bacteria were washed at 13,000×g for 5 min with 1X BD IMag buffer (BD BIOSCIENCES). Magnetic MS columns were placed on a MiniMACS Separator (MILTENYI BIOTEC), and conditioned with 500 µL 1X BD IMag buffer (BD BIOSCIENCES). Bacteria were either resuspended in 500 µL RB or in 0.25% (w/v) Trypsin-EDTA solution (SIGMA). For the latter, mixes were incubated for 15 min at 37 ºC in order to eliminate all potential protein-mediated interactions among bacteria. Positive fraction was retained in the magnetic columns after three washes with 1X BD IMag buffer (BD BIOSCIENCES). Finally, the positive fraction was eluted by separation of the magnetic columns from the MiniMACS separator and after addition of 500 µL 1X BD IMag buffer. Both positive and negative fractions were analysed by FC and 16S RNA gene sequencing. Serial dilutions of the positive fractions were plated onto the surface of MRS agar plates, and the resulting colonies were identified by 16S rRNA gene amplification and sequencing (i.e. three colonies of intestinal synthetic microbiota and seven colonies of complex intestinal microbiota).
Immunofluorescence microscopy. The binding specificity of the anti-SlpA antibodies contained in the polyclonal serum was also determined by confocal scanning laser microscope equipped with a Leica DFC365FX digital camera (DMi8, Leica Microsystems).
Firstly, the different bacterial species were grown individually until exponential phase OD 600 ≈ 0.7, were then washed with FC buffer and incubated for 15 min with the polyclonal antibody anti-SlpA conjugated with FITC (200 µg/mL). In the case of positive fractions enriched in L. acidophilus, bacterial mixes were incubated for 15 min with the polyclonal antibody anti-SlpA conjugated to APC (200 µg/mL). Individual bacteria and positive fractions were resuspended in 10 µL of flow cytometer buffer and analysed with fluorescence microscopy using a 100 × oil objective. The images were acquired with software LasX (LEICA Microsystems). The FITC filter cube (excitation 480/40, emission 527/30) was used for labelled bacteria species, and for the positive fraction of the supplemented microbiota the APC filter cube (excitation 620/60, emission 700/75) was selected. www.nature.com/scientificreports/ R-250 (THERMO FISHER SCIENTIFIC). Western blotting was performed with the objective of checking the signal of polyclonal antibody on the S layer protein A of L. acidophilus DSM20079 T . Extracellular proteins were obtained following a previously described protocol 46 . Proteins separated by SDS-PAGE were transferred and immobilized onto polyvinylidene fluoride (PVDF) membranes (GE HEALTHCARE, Madrid, Spain) for 30 min under a constant voltage of 30 V for one hour in a mini gel tank/blot module set (THERMO FISHER SCIENTIFIC). PVDF membranes were blocked with PBS (phosphate buffered saline, OXOID) supplemented with 0.1% (v/v) Tween-20 (PBST) (SIGMA) and with 5% (w/v) skimmed milk (PBST-L) (OXOID) for 3 h at RT under constant shacking. Membranes were washed twice and monoclonal/polyclonal antibodies were diluted to 1:2000 in PBST-L and incubated over the membranes (O/N, 4 °C) with slight agitation. Membranes were then washed 4 times with PBST and incubated for 1 h with a secondary antibody conjugated to horseradish peroxidase (horseradish peroxidase-conjugated anti-rabbit or anti-mouse IgG (SIGMA)), diluted to 1:1000 in PBST-L. For detection of immunoreactive bands, a commercial solution containing the chromogenic reagents chloronaphthol and diaminobenzidine (CN/DAB Substrate Kit, THERMO FISHER SCIENTIFIC) was used.

Specificity of polyclonal antibody by Western
16S rRNA sequencing of positive and negative fractions. Partial 16S rRNA gene sequences were amplified from extracted DNA using primer pair Probio_Uni and/Probio_Rev, targeting the V3 region of the 16S rRNA gene sequence 47 . 16S rRNA gene amplification and amplicon checks were carried out as previously described 47 . 16S rRNA gene sequencing was performed using a MiSeq (ILLUMINA) at the DNA sequencing facility of GenProbio srl (https ://www.genpr obio.com) according to the protocol previously reported 47 . Following sequencing, the .fastq files were processed using a custom script based on the QIIME software suite 48 . Pairedend read pairs were assembled to reconstruct the complete Probio_Uni/Probio_Rev amplicons. Quality control retained sequences with a length between 140 and 400 bp and mean sequence quality score > 20 while sequences with homopolymers > 7 bp and mismatched primers were omitted. In order to calculate downstream diversity measures (alpha and beta diversity indices, Unifrac analysis), 16S rRNA Operational Taxonomic Units (OTUs) were defined at 100% sequence homology using DADA2 49 ; OTUs not encompassing at least 2 sequences of the same sample were removed. Notably, this approach allows highly distinctive taxonomic classification at single nucleotide accuracy 49 . All reads were classified to the lowest possible taxonomic rank using QIIME2 48,50 and a reference dataset from the SILVA database 51 .
The bacterial profile of each sample was represented through Bar plots. Only genera with relative abundance > 0.5% (or 1%) are shown. Moreover, the bacterial profile at species level has been predicted, which is to be considered as approximate. The 16S rRNA gene sequences have been deposited in the NCBI database with the Bioproject Accession number PRJNA655374 (biosamples from SAMN15733960 to SAMN15733969). Statistical analysis. Data were represented by mean ± SEM and differences between the bacterial labelled conditions were assessed at least in triplicate.