Genomic diversity in Porphyromonas: evidence of Porphyromonas catoniae commensality in lungs

Abstract

Data on the genomics of Porphyromonas species other than Porphyromonas gingivalis (POTG), particularly within pulmonary environments, are scarce. In this study, we conducted whole-genome sequencing on pulmonary isolates of POTG, specifically Porphyromonas catoniae (n=3), Porphyromonas pasteri (n=1) and Porphyromonas uenonis (n=2), from people with cystic fibrosis. These genomic analyses were complemented with antimicrobial susceptibility tests for these strains. We compared the genomic sequences of these pulmonary isolates with those of previously characterized Porphyromonas species. Our study revealed a distinct clade differentiation between non-pigmented and pigmented Porphyromonas species. Unlike P. gingivalis, the pulmonary POTG strains lacked known virulence genes, with the exception of a putative haemolysin gene. Regarding antibiotic resistance, notable resistances were limited to vancomycin in P. catoniae and clindamycin in P. uenonis. These findings support the hypothesis that POTG species may predominantly behave as commensals in the lung environment rather than as pathogens.

Impact Statement

Porphyromonas constitutes a prevalent genus of anaerobic bacteria within the respiratory tract. Despite its ubiquity, the precise taxonomic delineation of Porphyromonas species implicated in respiratory conditions needs to be better defined, as most microbiome analyses report results only at the genus level. Consequently, data about these pulmonary species often default to associations with Porphyromonas gingivalis, a well-characterized oral pathogen. In this study, we sequenced the complete genomes of six Porphyromonas strains isolated from the airway microbiota of people with cystic fibrosis (CF) to enhance the representation and identification of Porphyromonas other than gingivalis (POTG) species. Our phylogenomic analysis aimed to elucidate the diversity within pulmonary POTG. Notably, genomes of the commensal species Porphyromonas catoniae isolated from CF patients did not harbour virulence genes typically associated with P. gingivalis. Additionally, phenotypic resistance profiling against nine clinically relevant antibiotics revealed low resistance levels, notwithstanding the frequent antibiotic treatments administered to CF patients. Our findings provide compelling evidence for the non-pathogenic role of these POTG species in the pulmonary environment.

Data Summary

Six Porphyromonas genomes generated in this study are available in the Sequence Read Archive and GenBank databases under BioProject accession PRJEB75658.

The raw dataset generated during the current study is available in the European Nucleotide Archive (repository with the project accession number).

The authors confirm that all supporting data, code and protocols have been provided within the article or through supplementary data files.

Introduction

In recent years, the composition of the pulmonary microbiota has garnered significant interest, particularly regarding its role in chronic respiratory diseases. The lung microbiota has a complex ecology rich in anaerobes, offering insights into disease progression [1,15]. Porphyromonas, ranking among the top ten anaerobic taxa in healthy lungs, exemplifies the metabolic diversity of lung anaerobes, spanning both commensal and pathogenic roles [16,17]. Notably, Porphyromonas gingivalis, a key pathogenic species, has been extensively studied for its virulence and adaptation strategies. However, other Porphyromonas species, such as Porphyromonas catoniae and Porphyromonas pasteri, typically associated with the healthy oral microbiota, have yet to be investigated, particularly from pulmonary sources [18,19].

Our study aims to fill this gap by performing whole-genome sequencing (WGS) on pulmonary Porphyromonas isolates previously collected [17]. We explore the genomic distinctions between oral and pulmonary strains and identify potential genes of interest within the genomes of pulmonary Porphyromonas other than gingivalis (POTG). Additionally, we examine the antibiotic resistance genes (ARGs) in conjunction with phenotypic resistance profiles to assess whether POTG could act as a commensal species within the respiratory tract. Our findings could significantly impact our understanding of the potential commensal or pathogenic roles of less-studied Porphyromonas species.

Methods

Isolation and culture of strains

The clinical strains of POTG employed in this study comprised three strains of P. catoniae (PC), one strain of P. pasteri (PP) and two strains of Porphyromonas uenonis (PU), all isolated from the sputum of people with cystic fibrosis (CF). These CF patients had not had an exacerbation or been on antibiotic therapy for at least a month at the time of sampling. Sample selection adhered to the Murray–Washington quality score to exclude contamination from saliva [20]. Collection was conducted using a device specifically designed to preserve an anaerobic environment as previously described [17]. This involved plating the samples within an anaerobic chamber (Bactron 300, Sheldon Manufacturing) that maintained a gas composition of 5% CO₂, 5% H₂ and 90% N₂.

The bacterial colonies were initially cultivated on anaerobe basal agar supplemented with 5% sheep blood (ABA-SB) from ThermoFisher. ABA-SB media containing kanamycin–vancomycin and colistin–nalidixic acid were also employed for selective cultivation. The strains were identified using MALDI-TOF MS with the Biotyper MBT system (Bruker). Additional confirmation for some colonies was achieved by 16S rRNA sequencing performed on an Applied Biosystems 3130xl instrument (ThermoFisher) using the BigDye Terminator Cycle Sequencing Kit Ready Reaction v3.1 (Applied Biosystems) as outlined by Lamoureux et al. [17].

Porphyromonas cultures were grown on ABA-SB plates for 4–7 days, followed by subculturing in a homemade peptone yeast liquid medium for genomic DNA extraction. The cultures were incubated for an additional 3 days at 37 °C within the anaerobic chamber, ensuring optimal growth conditions for subsequent analyses.

DNA extraction and sequencing

DNA was extracted using the QIAamp DNA Mini Kit (Qiagen), following the manufacturer’s instructions, with elution in 100 µl. We employed two distinct sequencing platforms to ensure comprehensive genomic data coverage.

1. PacBio Sequel system. Library preparation was carried out using the Template Prep Kit V2 per the manufacturer’s specifications. Sequencing was conducted in a single run at the Faculty of Medicine, Brest, France, using the Sequel Sequencing Kit 3.0 and the Sequel SMRT Cell 1M v3 LR Tray. This system is noted for its long-read capabilities, crucial for assembling complex genomic regions.

2. Illumina MiSeq system. Libraries were prepared using the TruSeq Nano kit, following the manufacturer’s guidelines. WGS was performed in a single run using V3 chemistry (2×250 bp) at the Genotoul platform, Castanet-Tolosan-Auzeville, France. This system is favoured for its high-precision short-read sequencing.

This hybrid sequencing strategy was designed to maximize the coverage and accuracy of the genomic data, facilitating a deeper understanding of the genetic intricacies of the Porphyromonas species under study.

Genome assembly

Quality control. Raw sequence reads underwent rigorous quality control to assess and filter out low-quality sequences. Adapter sequences and substandard reads were removed using bbduk from the bbmap suite v38.92–0, setting a quality base threshold of 20 and a minimum length requirement of 200 bp. Furthermore, we assessed quality and potential contamination using kat v2.4.2 and kraken2 v2.1.2 to identify human and other non-target sequences.

Genome assembly and annotation. Hybrid genome assemblies were conducted using the Unicycler pipeline v0.4.8-beta, incorporating a suite of tools including spades.py v3.13.0, racon v1.3.3, makeblastdb v2.9.0+, tblastn v2.9.0+, bowtie2-build v2.3.5, bowtie2 v2.3.5, samtools v1.9, java v11.0.1 and pilon v1.23 [21]. Assembly quality statistics were generated using Quast v5.02, and genome completeness and contamination levels were evaluated with CheckM v1.0.18.

Scaffold profiling and gene identification. We utilized Anvi'o v7 for scaffold profiling, employing Prodigal v2.6.354 with default settings for gene identification and HMMER v3.355 to detect bacterial single-copy genes and rRNA-based hidden Markov model (HMM). Short reads from each strain were mapped back to their respective scaffolds using Bowtie2 v2.4.4–0, with a minimum identity threshold of 95%. The mapped reads were then stored as BAM files using samtools. Anvi’o was subsequently used to profile each BAM file, allowing for precise estimation of coverage and detection statistics for each scaffold.

Phylogenomics and pan-genome analysis

To determine the evolutionary relationships of our six pulmonary Porphyromonas strains, we downloaded all publicly available Porphyromonas and Alloprevotella genomes (used as an outgroup for comparison) from the NCBI database (accessed on 3 June 2021). The downloads were performed using the ncbi-genome-download tool (https://github.com/kblin/ncbi-genome-download). Genome quality was assessed by estimating completeness and contamination levels based on the presence of single-copy core genes using Anvi’o (https://anvio.org/). Genomes with >90% completeness and <10% contamination were selected for further analysis. We further evaluated the similarity between these genomes by calculating their pairwise average nucleotide identity (ANI) using FastANI v1.32. Genomes with >99% ANI were grouped together, and for each group, only the genome with the highest quality (based on completeness and contamination) was selected for detailed phylogenomic and pan-genomic analyses. As a result, we ultimately retained 59 Porphyromonas genomes and 15 Alloprevotella genomes for further analysis. Using HMMER v3.3, we extracted HMM hits for each genome, focusing on 36 ribosomal single-copy core genes. These genes were aligned, and gaps affecting more than 50% of nucleotide positions were removed using trimAl v1.4.1, ensuring alignment quality for subsequent analyses. The phylogenetic analysis of these high-quality ribosomal gene alignments was conducted on a concatenated dataset. Substitution models for each gene were automatically selected from the GTR+F+Γ+I+R family using model-testing algorithms to find the best-fitting partition scheme. Phylogenetic trees were inferred using maximum likelihood estimation with proportional variation in branch lengths across partition subsets, implemented in IQ-TREE2 v1.7, using the LG+F+R4 model. The robustness of the phylogenetic tree was supported by 5,000 bootstrap replicates, providing statistical confidence in the branching patterns observed.

To explore the genetic diversity and evolutionary relationships within the Porphyromonas genus, we employed the Anvi’o platform to compute and visualize the pan-genome of Porphyromonas, encompassing both our strains and reference genomes. We stored 65 genomes (59 reference genomes from NCBI and our 6 pulmonary strains) in an Anvi’o database using the ‘anvi-gen-genomes-storage’ command. For the pan-genomic analysis, we focused on closely related genomes available from NCBI. Creating the pan-genome involved passing the stored database to the ‘anvi-pan-genome’ command, which utilized NCBI’s blast to assess gene similarity within and between the genomes. To cluster these genes into functionally coherent groups, we applied the Markov cluster algorithm, facilitating the identification of core and accessory genomic components across the studied Porphyromonas species. This comprehensive analysis allows us to understand better the genomic architecture and variability among species, which is crucial for elucidating their evolutionary trajectories and potential functional roles in different environmental niches.

Genome annotation and characterization of pulmonary P. catoniae strains

The assembled genomes of P. catoniae isolated from pulmonary sources were comprehensively annotated and analysed using several advanced bioinformatics tools. We utilized Prokka v1.13 [22] for rapid annotation, the MicroScope Microbial Genome Annotation and Analysis Platform (MaGe) [23], the NCBI integrated Prokaryotic Genome Annotation Pipeline and the RAST server v2.0 (Overbeck, 2014). These tools were employed with default parameters to ensure a standardized approach across all annotations.

To delve deeper into the metabolic capabilities of our lung P. catoniae strains, we compared their profiles with those of P. catoniae ATCC 51270 and several strains of P. gingivalis (ATCC 33277, TDC60 and W83), using the KEGG and BioCyc databases. This comparison was facilitated by the analytical tools available on the MicroScope platform, which also enabled the detection of macromolecular systems via MacSyFinder v1.0.2, prediction of CRISPR arrays using CRISPRCasFinder v4.2.19, identification of ARGs through the CARD database v3.0.2 and exploration of secondary metabolite regions using antiSMASH v5.0.0. Additionally, we investigated the presence of mobile genetic elements in the genomes using Phaster, viralVerify [24] and VirSorter2 [25]. These analyses provided insights into the genomic flexibility and potential adaptive mechanisms of P. catoniae in the pulmonary environment.

Phenotypic tolerance and antibiotic resistance testing

Given the genomic identity established among the strains PC103, PC123 and PC135, phenotypic tests were conducted exclusively on strain PC103 as a representative to evaluate environmental tolerance.

Oxygen tolerance. The PC103 strain’s oxygen tolerance was assessed by suspending the bacteria in non-oxygen-reduced physiological serum at two initial concentrations: a high concentration of 10⁷ c.f.u. ml⁻¹ and a lower concentration of 10³ c.f.u. ml⁻¹. The suspensions were exposed to ambient oxygen, and samples were periodically collected, plated and incubated for 7 days at 37 °C in an anaerobic chamber (5% CO₂, 5% H₂, 90% N₂) to measure the residual inoculum.

Acid tolerance. PC103’s acid tolerance was tested using ABA-SB medium acidified with 1M HCl to achieve pH values of 6.86, 6.38, 6.0, 5.5 and 5.0. Actively growing cultures of P. catoniae PC103 on standard ABA-SB were transferred to the pH-adjusted ABA-SB. These plates were incubated under the same anaerobic conditions for 7 days, after which colony growth was evaluated.

Cadmium tolerance. Cadmium tolerance was examined using ABA-SB medium supplemented with increasing cadmium concentrations, ranging from 0 to 80 mg l⁻¹. Cultures of PC103 grown on standard ABA-SB were subcultured onto cadmium-enriched ABA-SB and incubated for 7 days in an anaerobic environment at 37 °C. Growth assessment was performed by examining the colonies’ presence. These tests provide valuable insights into the robustness of the PC103 strain under various stress conditions, which may contribute to understanding its survivability and adaptability in the pulmonary niche.

Antibiotic resistance. The antibiotic resistance profiles of six Porphyromonas strains isolated from pulmonary sources were assessed following the 2020 guidelines provided by the European Committee on Antimicrobial Susceptibility Testing and the Committee for Antibiogram of the French Society of Microbiology [26]. Freshly isolated cultures from ABA-SB were suspended to a density equivalent to 1 McFarland standard in anaerobe basal broth (ThermoFisher). These suspensions were then plated onto ABA-SB agar to establish a uniform bacterial lawn. The antibiotics tested included amoxicillin (25 µg), amoxicillin/clavulanic acid (20 µg/10 µg), clindamycin (2 µg), imipenem (10 µg), metronidazole (5 µg), moxifloxacin (5 µg), piperacillin (30 µg) and piperacillin/tazobactam (30 µg/6 µg) (ThermoFisher). Additionally, a vancomycin disc (30 µg) was included due to the noted susceptibility of other Porphyromonas species, although P. catoniae typically shows resistance [27]. The inoculated agar plates were incubated in an anaerobic chamber maintained at 37 °C for at least 48 h to allow adequate growth and interaction between the bacterial isolates and the antibiotic agents.

Results

WGS of POTG (P. catoniae, P. pasteri and P. uenonis)

The genomic characteristics of pulmonary strains of the Porphyromonas genus, specifically P. catoniae, P. pasteri and P. uenonis, were investigated through WGS. The G+C content of these strains fell within the range of 51–56 mol%, consistent with prior findings for this genus. Utilizing de novo hybrid assembly methods, the genomes of the pulmonary POTG were effectively reconstructed, resulting in contig numbers ranging from 1 to 66, with N50 values spanning from 216,372 to 2,401,504 bp. The largest contig length observed for Porphyromonas ranged from 320,622 to 2,401,504 bp. Notably, the de novo assembly lengths agreed with previously reported median total lengths for each species. Assessment of genome completeness revealed values ranging from 95.05 to 99.92%. Detailed genomic statistics are provided in Table 1.

Table 1. Sequencing and assembly statistics of pulmonary strains of POTG.

Strain	POTG species	Unicycler assembly	N50	Largest contig	G+C (mol%)	Contigs no.	Completeness	Contamination	Predicted-coding sequence
PC103	P. catoniae	2,032,211	2,032,211	2,032,211	51.00	1	95.05	0.157	1,619
PC123	P. catoniae	2,032,214	1,601,551	1,601,551	51.00	3	95.05	0.157	1,622
PC135	P. catoniae	2,032,218	2,032,218	2,032,218	51.09	1	95.05	0.157	1,620
PP88	P. pasteri	2,225,553	2,205,487	2,205,487	55.75	2	98.87	0	1,703
PU103	P. uenonis	2,401,504	2,401,504	2,401,504	52.31	1	99.92	0	1,882
PU61	P. uenonis	2,325,647	216,372	320,622	52.49	66	99.92	0	1,876

Analysis of ANI in pulmonary Porphyromonas strains

ANI analysis of pulmonary strains revealed two distinct genetic clusters among the Porphyromonas species (Table 2). Non-pigmented strains of P. catoniae and P. pasteri formed one group, while all P. uenonis strains comprised the second group. Specifically, strains PC103, PC123 and PC135 were identified as belonging to the same species, adopting a 95% similarity threshold for species delineation, as established by Konstantinidis and Tiedje [28] and Konstantinidis et al. [29]. These strains exhibited ANI values of 96.6 and 96.7% with the reference strain of P. catoniae (F0037).

Table 2. ANI between Porphyromonas genomes.

Strain	PC103	PC123	PC	PC135	PP88	PP	PU103	PU61	PU	PA	PB	PE	PS	PG
PC103	100.0	99.9	96.6	99.9	79.0	78.5	<80	<80	<80	<80	<80	<80	<80	<80
PC123	99.9	100.0	96.7	99.9	78.8	78.4	<80	<80	<80	<80	<80	<80	<80	<80
PC135	99.9	99.9	96.6	100.0	79.0	78.6	<80	<80	<80	<80	<80	<80	<80	<80
PP88	79.0	78.8	78.7	79.0	100.0	96.4	<80	<80	<80	<80	<80	<80	<80	<80
PU103	<80	<80	<80	<80	<80	<80	100.0	97.6	97.4	88.8	<80	<80	<80	<80
PU61	<80	<80	<80	<80	<80	<80	97.6	100.0	97.4	88.9	<80	<80	<80	<80

*PC, P. catoniae F0037; PP, P. pasteri JCM 30531; PU, P. uenonis DSM 23387; PA, Porphyromonas asaccharolytica DSM 20707; PB, Porphyromonas bennonis DSM 23058; PE, Porphyromonas endodontalis 52278C01; PS, Porphyromonas somerae DSM 23386; PG, P. gingivalis ATCC 33277.

Conversely, strain PP88 showed only 78.7–79% similarity with strains PC103, PC123 and PC135, indicating significant genomic divergence. However, PP88 displayed a higher similarity of 96.4% with the reference strain of P. pasteri (JCM 30531), suggesting a closer genetic relationship with this species. Furthermore, strains PU61 and PU103 showed a high similarity of 97.6%. These strains and the reference strain of P. uenonis (DSM 23387), which was initially isolated from a human sacral decubitus ulcer, shared a similarity of 97.4% ANI.

Notably, all analysed POTG lung strains demonstrated less than 80% similarity with the reference strain of P. gingivalis (ATCC 33277), further highlighting the genetic distinctiveness of the pulmonary strains within this group.

Phylogenomic analysis

In the phylogenomic analysis, ribosomal sequences from all publicly available Porphyromonas strains were concatenated to construct a comprehensive phylogenomic tree (Fig. 1). This analysis enabled the clear differentiation of species within the Porphyromonas genus. The non-pigmented strains of the POTG, including P. pasteri and P. catoniae, formed a distinct clade, demonstrating their phylogenetic separation from other pigmented Porphyromonas species.

Fig. 1. Phylogenomic tree of Porphyromonas pulmonary species. Tree based on concatenating ribosomal sequences using IQ-TREE natively with ModelFinder Plus [49].

In detail, lung strains PU61 and PU103 are closely aligned with all reference genomes of P. uenonis, indicating strong phylogenetic consistency within this species. Similarly, PC103, PC123 and PC135 clustered closely with all reference genomes of P. catoniae, confirming their classification within this species. Additionally, PP88 was phylogenetically aligned with all reference genomes of P. pasteri, underscoring its species-specific grouping.

Pan-genomic analysis of Porphyromonas strains

A comprehensive pan-genomic analysis was conducted on lung strains of Porphyromonas from our collection alongside reference genomes from the POTG group. This analysis revealed a core genome consistent across all Porphyromonas species, comprising 357 gene clusters encompassing 8,587 ORFs. A species-specific accessory genome was identified exclusively in P. catoniae, containing 255 gene clusters (419 ORFs).

The study aimed to elucidate potential virulence factors among the POTG, mainly focusing on non-pigmented Porphyromonas strains. The virulence genes examined were primarily those previously identified in P. gingivalis ATCC 33277 by Mendez et al. [30], who extensively searched for such genes in that strain. However, our current analysis did not detect genes associated with gingipains or haemagglutinin in the non-pigmented Porphyromonas genomes. Only one putative haemolysin gene was found, and an additional gene was identified as hyaluronidase/collagenase (K01197; EC: 3.2.1.35), indicating a reduced virulence gene profile in these strains. Regarding secretion systems, a type IX secretion system was detected in both PC103 and P. catoniae ATCC 51270 genomes, albeit incomplete, and a type IV secretion system was noted in P. catoniae ATCC 51270, also incomplete. Interestingly, specific metabolic pathways and transport mechanisms, such as the pentose phosphate pathway and iron transport genes, were explicitly identified in the lung strains of P. catoniae, suggesting adaptations to their niche. These screened genes and their potential implications are summarized in Table S1, available in the online Supplementary Material.

Genome comparison of Porphyromonas strains

The comprehensive genomic analysis of the lung-derived P. catoniae strain PC103 reveals a single circular chromosome measuring 2,032,211 bp with a genomic DNA G+C content of 51.09 mol%. Genome annotation, performed using the MaGe platform, predicted 1,669 genes, including 1,609 protein-coding sequences (Fig. S1). The analysis also identified one CRISPR locus between 805,723 and 806,843 bp downstream of a CAS-type II cluster spanning from 800,799 to 805,615 bp. Additionally, two incomplete prophage regions were detected, each ~27 and 26 kb in size, showing 40% completeness.

The metabolic profile of P. catoniae PC103 was compared against P. catoniae ATCC 51270 and several P. gingivalis strains (ATCC 33277, TDC60 and W83) using the MicroScope platform. This comparison highlighted significant metabolic distinctions between the P. catoniae and P. gingivalis strains, irrespective of their source (oral or pulmonary). Key pathway differences are summarized in Table 3, including a cadmium transport system unique to P. catoniae strains and absent in P. gingivalis strains. Conversely, genes involved in protein citrullination were exclusive to P. gingivalis strains.

Table 3. Metabolic comparison of P. catoniae PC103, P. catoniae ATCC 51270, P. gingivalis ATCC 33277, P. gingivalis TDC60 and P. gingivalis W83 with the MicroScope platform using KEGG and MicroCyc databases.

Reaction	Database	No. of reaction	Completeness of pathway
			P. catoniae PC103	P. catoniae ATCC51270	P. gingivalis ATCC 33277	P. gingivalis TDC60	P. gingivalis W83
Cadmium transport	MicrocCyc	1	1	1	0	0	0
Protein citrullination	MicrocCyc	1	0	0	1	1	1
Sulphide oxidation II	MicrocCyc	1	1	1	0	0	0
Sulphur reduction II	MicrocCyc	1	1	1	0	0	0
Oxidized GTP and dGTP detoxification	MicrocCyc	2	0,5	0,5	0	0	0
Superoxide radical degradation	MicrocCyc	2	0,5	0,5	0	0	0
Arginine degradation IV	MicrocCyc	3	1	1	0,67	0,67	0,67
Arginine-dependent acid resistance	MicrocCyc	1	1	1	0	0	0
Glutamate degradation	MicrocCyc	1	0	0	1	1	1
Glutamine degradation I	MicrocCyc	1	0	0	1	1	1
Glutamine degradation II	MicrocCyc	1	0	0	1	1	1
Histidine degradation I	MicrocCyc	4	0	0	0,75	0,75	0,75
l-Serine degradation	MicrocCyc	1	0	0	1	1	1
Methionine degradation II	MicrocCyc	1	0	0	1	1	1
Tryptophan degradation II	MicrocCyc	1	0	0	1	1	1
Melibiose degradation	MicrocCyc	1	0	0	1	1	1
Fatty acid biosynthesis, elongation	KEGG	35	1	1	0,8	0,8	0,8
Fatty acid biosynthesis, initiation	KEGG	5	0,6	0,6	0,6	0,6	0,6
l-threo-Tetrahydrobiopterin biosynthesis, GTP to l-threo-BH4 (Folate biosynthesis)	KEGG	4	0,75	0,75	0,75	0,75	0,75
Tetrahydrobiopterine biosynthesis, GTP to BH4 (Folate biosynthesis)	KEGG	4	0,75	0,75	0,75	0,75	0,75
Tetrahydrofolate biosynthesis, GTP to THF (Folate Biosynthesis)	KEGG	12	0,75	0,75	0,83	0,83	0,83
Tetrahydrofolate biosynthesis, mediated by PTPS, GTP to THF (Folate biosynthesis)	KEGG	9	0,89	0,89	0,89	0,89	0,89
Tetrahydrofolate biosynthesis, mediated by ribA and trpF, GTP to THF (Folate biosynthesis)	KEGG	10	0,8	0,8	0,8	0,8	0,8
Lipoic acid biosynthesis, animals and bacteria, octanoyl-ACP to dihydrolipoyl-H to dihydrolipoyl-E2 (Lipoic acid biosynthesis)	KEGG	3	0,67	0,67	0,67	0,67	0,67
Lipoic acid biosynthesis, eukaryotes, octanoyl-ACP to dihydrolipoyl-H (Lipoic acid biosynthesis)	KEGG	2	1	1	1	1	1
Lipoic acid biosynthesis, octanoyl-CoA to dihydrolipoyl-E2 (Lipoic acid biosynthesis)	KEGG	2	0,5	0,5	0,5	0,5	0,5
Lipoic acid biosynthesis, plants and bacteria, octanoyl-ACP to dihydrolipoyl-E2/H (Lipoic acid biosynthesis)	KEGG	4	1	1	0	0	0
Pantothenate biosynthesis, valine/l-aspartate to pantothenate	KEGG	5	1	1	1	1	1
Riboflavin biosynthesis, plants and bacteria, GTP to riboflavin/FMN/FAD	KEGG	9	1	1	1	1	1
Pyridoxal-P biosynthesis, erythrose-4P to pyridoxal-P (Vit B6 metabolism)	KEGG	6	0,83	0,83	0,83	0,83	0,83

The pathway’s completeness is indicated as 1=100% and 0=0%. Reactions are named as they are in the KEGG or MicroCyc databases.

P. catoniae strains exhibited unique sulphur metabolism pathways (sulphide oxidation and sulphur reduction) and an arginine-dependent acid resistance gene. The genomes also showed a partial presence of pathways for detoxifying oxidized GTP and dGTP and the degradation of superoxide radicals. Furthermore, the fatty acid biosynthesis pathways showed 60% completion for initiation and 80% completion for elongation processes. Additionally, various vitamin biosynthesis pathways were identified in P. catoniae, including those for folate, lipoic acid, pantothenate, riboflavin and pyridoxal, contrasting with the degradation pathways for amino acids and sugars found in P. gingivalis genomes.

In vitro confirmation of genotypic characteristics of Porphyromonas strains

Oxygen tolerance. The presence of genes related to the detoxification of oxidized GTP and dGTP and the degradation of superoxide radicals in P. catoniae PC103 suggested potential oxygen tolerance. To verify this, we exposed a suspension of PC103 to atmospheric oxygen concentrations, monitoring the bacterial survival over time. The results showed only a 1.86 log decrease in c.f.u. per millilitre after 8 h at an initial concentration of 10⁷ c.f.u. ml⁻¹ and a 1.25 log decrease after 6 h at 10³ c.f.u. ml⁻¹.

pH tolerance. The in silico-identified arginine-dependent acid resistance gene prompted an in vitro assessment of PC103’s acid tolerance. Cultures were grown on ABA-SB medium at various pH levels (6.86, 6.38, 6.0, 5.5 and 5.0) and incubated for 7 days. PC103 was able to grow at pH levels down to 5.5, but no growth was observed at pH 5.0, even with extended incubation.

Cadmium resistance. The WGS of PC103 had revealed a cadmium transport gene, suggesting possible resistance to cadmium. This was tested by culturing PC103 on an ABA-SB medium supplemented with increasing cadmium concentrations (0–80 mg l⁻¹). The results demonstrated that PC103 could tolerate cadmium concentrations up to 20 mg l⁻¹, but growth was inhibited at 40 mg l⁻¹ and above.

Antibiotic resistance. Investigations into the antibiotic susceptibility of pulmonary Porphyromonas species revealed sensitivity to a broad spectrum of antibiotics, with consistent antibiogram profiles across the species, except for specific resistances. Strains of P. uenonis (n=2) displayed resistance to clindamycin. Strains of P. catoniae, including PC103, were resistant to vancomycin. Intriguingly, no vancomycin resistance genes were identified in the reconstructed genomes of these strains using the CARD database, indicating an unidentified resistance gene or a potential non-genomic resistance mechanism. Indeed, P. catoniae, as a Gram-negative bacillus, may be intrinsically resistant to vancomycin due to the inability of this antibiotic to cross the outer membrane.

Discussion

In this study, we conducted WGS of non-pigmented pulmonary Porphyromonas strains (P. catoniae, P. pasteri and P. uenonis) isolated from CF patients. Unlike the well-known pathogen P. gingivalis, these POTG isolates lacked key virulence genes commonly associated with pathogenicity, possessing only a putative haemolysin gene. This genomic profile strongly suggests a commensal role in the lung environment rather than pathogenicity.

As reported in several studies, P. gingivalis is a well-documented pathogen extensively associated with periodontitis [18,31,33]. Recent research has also implicated this bacterium in systemic diseases, including lung cancer [34], rheumatoid arthritis (RA) [35] and Alzheimer’s disease [36], highlighting its potential role beyond oral health. In contrast, the genus Porphyromonas has been detected in both diseased and healthy pulmonary niches using molecular and cultural methods [2,5, 13,15, 17, 37]. Specifically, targeted metagenomic analyses have frequently identified POTG strains, predominantly non-pigmented Porphyromonas, in pulmonary samples [2,16, 17]. Unlike P. gingivalis, species such as P. catoniae and P. pasteri are primarily associated with healthy individuals, suggesting potentially different roles in the microbial ecology of the lung [18,38,40]. Our findings align with previous studies indicating non-pigmented Porphyromonas species frequently associate with healthy rather than diseased oral and pulmonary microbiota [40]. Similar observations were reported in patients suffering from less severe lung disease. Indeed, authors have shown that POTG species, as well as other anaerobes, are frequently present in the lungs of CF patients and may be linked to a less severe disease phenotype (less impairment of respiratory function and less infection with the pathogen Pseudomonas aeruginosa) [2,41]. It should be noted that all our strains came from patients without exacerbation at the time of collection of the respiratory samples. Beyond CF, Porphyromonas species have also been detected in other respiratory conditions, such as chronic obstructive pulmonary disease and smoker lungs, but were not associated with increased inflammatory markers unlike other anaerobes [5,42]. This suggests that POTG could serve a commensal or even beneficial role in preserving microbial balance within the diseased airway microbiome. Given the limited sample size of our study (n=6 strains), future investigations with larger datasets are needed to gain deeper insights and reliably establish correlations between POTG strains and clinical outcomes in patients.

Historically, genomic data on non-pigmented Porphyromonas species have been sparse. A notable study by O'Flynn et al. [43] included a genome of P. catoniae isolated from a human gingival crevice, revealing significant genomic distinctions between P. catoniae and P. gingivalis. Key differences highlighted were the absence of haemagglutinin and gingipains, virulence factors in P. gingivalis, whereas genes for haemolysins and iron transport were detected in P. catoniae. Consistent with these findings, our current analysis of pulmonary non-pigmented Porphyromonas strains identified similar absences and noted the presence of a putative haemolysin (cds 1638831 1640078 – HANLCC_06880 haemolysin SO:0001217, UniRef: UniRef50_A0A0A2G0A3, UniRef: UniRef90_L1N9Y2). Although known P. gingivalis virulence genes were absent in our pulmonary Porphyromonas isolates, there could be other virulence mechanisms at play. Future studies should incorporate in vivo and in vitro testing to comprehensively assess these bacteria’s biological roles and uncover any potential virulence factors that were not identified in silico.

Moreover, O'Flynn et al. reported the absence of glutamate enzymes in P. catoniae [43], except for an NADP-specific glutamate dehydrogenase (EC 1.4.1.4), which we also identified in our study, indicating its conservation across different isolates of this species. Interestingly, our research further established the presence of genes encoding pentose phosphate pathway enzymes in P. catoniae, contrasting with their absence in P. gingivalis genomes. This study also uncovered genes involved in synthesizing glutamate from ammonia and maintaining the glutamate pool in P. catoniae, elements missing in P. gingivalis. Conversely, genes related to the synthesis of ammonia from nitrate, the synthesis of vitamin B12 from uroporphyrinogen and the release of iron from blood cells, typically found in P. gingivalis, were absent in P. catoniae [43].

Significantly, none of the P. catoniae genomes, including lung clinical strains and P. catoniae ATCC 51270, exhibited genes involved in protein citrullination, a process linked to the production of citrullinated protein antigens, which are specific markers for RA [44]. This absence is crucial because P. gingivalis has been implicated in RA due to its ability to produce such antigens [45]. The lack of citrullination genes in non-pigmented Porphyromonas suggests that these species play no role in RA and possibly other autoimmune diseases. This genetic profile supports the hypothesis that non-pigmented Porphyromonas, particularly P. catoniae, are less likely to contribute to autoimmune pathogenesis than P. gingivalis.

In addition to the absence of specific P. gingivalis virulence factors such as haemagglutinin and gingipains, our genomic analysis of P. catoniae strains revealed a notable lack of amino acid degradation pathways for glutamate, glutamine, histidine, serine, methionine, tryptophan and melibiose. Unlike P. gingivalis, which utilizes these pathways as carbon sources to support its pathogenic lifestyle, the absence of such metabolic capabilities in P. catoniae may suggest a non-pathogenic nature. This potential for commensalism is further supported by the lack of other virulence genes commonly associated with P. gingivalis, aligning with findings that characterize non-pigmented Porphyromonas species, notably P. catoniae, as less likely to contribute to disease states [16,46]. Future studies should incorporate in vivo and in vitro testing to comprehensively assess these bacteria’s biological roles and impacts.

Moreover, our study identified genes associated with acid resistance in P. catoniae strains. Experimentally, these strains demonstrated the ability to grow at a low pH of 5.5 on ABA-SB agar, indicating a robust capacity for acid tolerance. This finding aligns with previous descriptions of P. catoniae as a lactic acid producer. This attribute contributes to its survival in acidic environments, much like other lactic acid bacteria, including species of Lactobacillus [27]. This acid resistance capability not only underscores the adaptive strategies of P. catoniae to various environmental stresses but also suggests its potential role in maintaining microbial balance within its niche. The ability to produce and withstand acidic conditions may facilitate the colonization and persistence of P. catoniae in the human microbiome, particularly in acid-prone sites, which is a characteristic typical of commensal bacteria.

Our investigation into the antibiotic resistance profiles of pulmonary Porphyromonas strains highlights key findings. Consistent with prior studies [27,47], Porphyromonas generally exhibits high susceptibility to antibiotics targeting anaerobes, including β-lactams. However, notable exceptions were observed: P. uenonis demonstrated resistance to clindamycin, a relatively new observation given that this species was only recently described [48], and P. catoniae displayed resistance to vancomycin, as previously reported [27]. In the strains of P. catoniae, the search for glycopeptide resistance genes in their genome was negative. The resistance to vancomycin of P. catoniae is linked to structural particularities of its outer membrane, reducing antibiotic permeability and thus preventing access to peptidoglycan precursors targeted by the antibiotic [27]. In contrast, other Porphyromonas species remain sensitive, reflecting intra-genus variations in membrane composition and permeability. This distinct resistance pattern highlights significant structural and biochemical diversity within the Porphyromonas genus, influencing their differential antibiotic susceptibilities [27]. While the present dataset includes only six pulmonary strains, expanding this dataset could provide deeper insights into glycopeptide resistance mechanisms within the POTG group. The isolation of these strains from lung samples remains challenging due to their sensitivity to most antibiotics and oxygen and their slow growth rate.

In conclusion, our study did not detect virulence genes typically associated with P. gingivalis in non-pigmented Porphyromonas strains, except for a putative haemolysin. These findings suggest that non-pigmented POTG, particularly P. catoniae and P. pasteri, are endogenous inhabitants of the lung, contributing to homeostasis rather than disease. This distinction between POTG and the pathogenic P. gingivalis is crucial for understanding the lung niche’s microbial dynamics and guiding clinical approaches to lung health and disease.

Abbreviations

ABA-SB

anaerobe basal agar supplemented with 5% sheep blood

ANI

average nucleotide identity

ARGs

antibiotic resistance genes

CF

cystic fibrosis

HMM

hidden Markov model

MaGe

MicroScope Microbial Genome Annotation and Analysis Platform

POTG

Porphyromonas other than gingivalis

RA

rheumatoid arthritis

WGS

whole-genome sequencing