Figure 1

Figure 1. Geographic spread of 12 Legionella pneumophilasequence type 1905 isolates by probable municipality of exposure (clinical) or sampling (environmental), Lisbon region, Portugal, 2014–2022. Location of 2014 Legionnaires’ disease outbreak...

Figure 2

Figure 2. Core SNP-based phylogeny of whole-genome sequencing data from 12 Legionella pneumophilaisolates obtained from samples collected in 3 interconnected municipalities in the Lisbon region, Portugal. Phylogeny was generated using...

Since the large 2014 LD outbreak, the ST1905 genotype has been identified several times in Portugal. For our study, we analyzed all L. pneumophila serogroup 1 ST1905 isolates (6 clinical and 6 environmental) obtained from samples collected in 3 interconnected municipalities in the Lisbon region (Lisbon, Loures, and VFX) in the context of clinical case investigations (outbreaks or as sporadic cases) (Figures 1, 2; Appendix 1; Appendix 2). We also included 2 culture-negative clinical samples genotyped as ST1905 for analysis of the temporal and geographic distribution of this strain after 2014 (Appendices 1, 2).

Single-nucleotide polymorphism (SNP)–based diversity analysis confirmed that all ST1905 isolates were genetically similar to PtVFX/2014, differing by 3–6 SNPs. Overall, the observed microevolution across the 12 isolates was marked by 10 SNPs (9 nonsynonymous and 1 synonymous mutations), 4 insertion/deletions, and 1 small recombination event (Figure 2). The low number of SNPs during the 8-year period supports the notion that L. pneumophila evolves at a very slow rate, resulting in substantial temporal and spatial conservation, as previously reported (9,10). When compared with PtVFX/2014, all isolates presented a recombination event in an ≈2.5-kb region (contig 8, PtVFX/2014_08985-08995) belonging to the type IVA secretion system, which is associated with the survival of the bacteria in the environment (11). Of note, Lpn_PT_E259_y2021, isolated from a sporadic LD case, presented mutational events, including 2 deletions and 1 insertion, not found in any other isolates included in this study. The first deletion encompasses a 45,479-bp segment (contig 8, PtVFX/2014_09080-09280) that includes an ≈37.5-Kb genomic island containing an lvh/lvr type IVA secretion system cluster. That cluster was first identified within the L. pneumophila species in the PtVFX/2014 strain (PtVFX/2014_09115-09280), possibly because of interspecies transfer from L. oakridgensis (3). The second deletion spanned 142 bp and occurred in a CRISPR-associated intergenic region (PtVFX/2014_08935/08940). Moreover, we observed a frameshift 25-bp insertion at the PtVFX/2014_13190 locus, coding for a protein with unknown function. That locus seems to be specific for L. pneumophila subsp. fraseri and is located within a larger genomic region that contains known Dot/Icm substrates (3). Although we cannot make conclusions about a potential adaptive role of the observed mutations, it has been hypothesized that mutations found exclusively in clinical isolates, as in our study, might reflect human-specific adaptation (10). L. pneumophila can infect and replicate in human alveolar macrophages, but human-to-human transmission is assumed to be rare; thus, fixation of those mutations into L. pneumophila circulating in the human population is unlikely (10,12). Still, it has been proposed that the recent expansion of L. pneumophila in manmade water systems, together with the widespread distribution of specific clones at global scale, aligns with the potential dissemination between humans or from humans to the environment (13).

In-depth phylogenetic and microevolutionary analysis showed that ST1905 isolates did not cluster by year of isolation (Figure 2). Still, it is noteworthy that all clinical and environmental isolates associated with a particular location clustered apart, supporting persistence of the strain in that location and linkage between different events in some specific settings. Indeed, in one facility, located in the Lisbon municipality, we observed a genetic and epidemiologic correlation between the isolates collected in the outbreak investigation in 2015 and those from 2018 (Figures 1, 2; Appendix 1). In that facility, when both clinical and environmental isolates collected in the same investigation context were available, they differed by <3 SNPs, further supporting their epidemiologic linkage. Local microevolution is expected (13) and might contribute to fitness changes, such as increased tolerance to copper (14).

Three mutation events (2 SNPs and 1 recombination) were shared by all ST1905 isolates collected after the 2014 outbreak (Figure 2) We retrospectively inspected isolates from the 2014 outbreak for evidence of any of those mutations (3) and found that 1 nonsynonymous mutation (in PtVFX/2014_02480, coding for a hypothetical protein) had already been detected in 1 of the 2014 clinical strains (data not shown). That observation provides further evidence that the recurrent ST1905 detections derived from the large bacterial dispersion that occurred in 2014, probably resulting from atypical atmospheric conditions (6,7,15). It also supports the knowledge that L. pneumophila outbreaks can be caused by multiple same-strain subpopulations (present simultaneously in a source of infection and diversified over time) or even by different co-existing strains (9).

Our results strongly indicate that the atypical L. pneumophila sg1 ST1905 strain is potentially persistent in diverse and geographically dispersed environments, far beyond the epicenter of the large 2014 outbreak in Portugal (VFX region). The recurrence of isolated cases or outbreaks in other regions with susceptible populations is thus of public health concern. Moreover, the observed microevolutionary traits and the potential of genetic recombination raise additional uncertainties regarding the evolutionary landscape of the ST1905 strain and its ability to further adapt and persist in the environment and, ultimately, cause human disease. Our study highlights the need for targeted public health preparedness and control strategies, emphasizing the added value of molecular epidemiology in the surveillance and management of LD.