Current culture-based methods for known saprozoic bacterial pathogens (Legionella spp., nontuberculous mycobacteria etc.) take 10-days to weeks for confirmation (e.g., (ISO 11731, 2017; Pfaller et al., 2021)) – far too long for a timely response. Further, cells of these pathogens may largely be present in viable but non-culturable (VBNC) states (hence missed by traditional culture methods) yet still infectious to humans (Dietersdorfer et al., 2018). Therefore, specialized amoeba co-culture is recommended for resuscitation (García et al., 2007; Dey et al., 2019a; Dey et al., 2019b; Dey et al., 2020). While still disputed by some, molecular methods are preferred for the management of water systems, such as qPCR (Lee et al., 2011). However, determining the fraction of infectious cells by qPCR is still problematic, given uncertainties in infectivity status and the variable residual chlorine levels likely present in piped water systems (Casini et al., 2018; Donohue et al., 2019; Donohue, 2021). Flow cytometry using immunocapture (Füchslin et al., 2010) or with cell-sorting in combination with qPCR shows promise to also identify VBNC cells, particularly those missed by conventional culture (e.g., ISO 11731:2017-05 pre-treatment procedure) but capable of infecting host cells (Nisar et al., 2023).

Nonetheless, live or dead, high concentrations (in excess of 10³ cells/100 mL, (Hamilton et al., 2021)) in distal parts of water delivery systems infer growth in the system that needs to be addressed, else risk exposure to infectious aerosols at some stage. The major amplification site for L. pneumophila in water systems is within biofilm amoebae (NAS, 2020). Therefore, herein a prior conceptual model for legionellae growth within pipe biofilms (Shaheen and Ashbolt, 2021) is expanded upon to include EV and possible amoeba monitoring targets and identifies research gaps. Overall, targeting hosts prior to rapid, explosive growth of legionellae is hypothesized as a useful target within a system of checks (flushing, temp <25 or >50°C, disinfectant residual etc.) for a proactive early warning water management system.

The primary hosts for Legionella within biofilms are various free-living protozoa, principally free-living amoebae (FLA) (Thomas et al., 2010) and ciliates (Tsao et al., 2019). While ciliates may excrete viable legionellae within fecal pellets for subsequent re-ingestion (Hojo et al., 2012; Berk and Garduño, 2013), L. pneumophila released following their lysis of amoebal trophozoites are more likely to lead to explosive growth cycles of phagocytosis-growth-release (Shaheen et al., 2019), given several hundred of cells that can grow per amoeba (Buse and Ashbolt, 2012).

A recent review of FLA is provided by Scheid (2019), in which he describes their diversity and the predatory heterotrophic feeding by trophozoites on biofilm microbiota and extracellular polymeric substances. Indeed, FLA are ubiquitous and instrumental to both biofilm formation and ecological successions. Important amoebal hosts supporting Legionella and other amoeba-resisting bacterial (ARB) pathogens include members of Acanthamoeba, Naegleria, Vermamoeba and Willaertia (Thomas et al., 2010). Although the term FLA has no relevance to amoebal taxonomy or phylogeny, it does separate them from the parasitic intestinal amoebae (e.g., Entamoeba histolytica) that have significantly reduced genomes due to their host’s providing most needs (Shabardina et al., 2018). Critical to FLA success and dispersal (air, water, soils & by wildlife) is the environmentally robust cyst form, that may also encase/protect beneficial and pathogenic bacteria and their viruses (Thomas and Greub, 2010; Schulz et al., 2020; Shi et al., 2021), many of whom may carry virus-encoded auxiliary metabolic genes (vAMGs) (Yuan and Ju, 2023).

Pertinent to engineered water systems has been the increasing recognition of the importance of suspended flocs/particulates in amoebal planktonic dynamics (Anderson, 2018). While not considered biofilms of fixed surfaces, these planktonic niches may represent sloughed biofilm material or biofilm-like growth on suspended particulates that contribute to microbial activity in piped water systems (Liu et al., 2016) and ultimately the bulk of internalized pathogens transmitted via aerosols (Hamilton et al., 2018; Shaheen and Ashbolt, 2021).

Nearly all living cells may excrete EVs, which are described based on their site of origin (i.e., exosomes, ectosomes, cytonemes & nanotubes) (Vermeulen et al., 2020.). However, only recently have EVs been recognized as participating in cell-to-cell communication processes, including between predator and prey (de Souza and Barrias, 2020), and carry specific virulence factors (Cruz Camacho et al., 2023) with known immunomodulatory properties (Costa et al., 2021). Exosomes are EVs that originate from the endocytic pathway of a cell (de Souza and Barrias, 2020; Cruz Camacho et al., 2023) and FLA-EVs are of focus here because of their likely multiple connections across the life cycle of L. pneumophila in biofilms of engineered environments (Figure 1).

Various drinking water associated FLA have been shown to generate EVs containing tens to hundreds of L. pneumophila cells within the respiratory range (< 10 µm dia.) (Shaheen and Ashbolt, 2018). These EVs may provide packages of pathogens directly to the alveoli of our lungs, potentially influencing current quantitative microbial risk assessment (QMRA) models of Legionella risks and specifically the dose-response relationship (Hamilton et al., 2019). Hypothesized in this article are the environmental queues that may not only influence EV production but also their role leading to explosive growth of legionellae in their native biofilms (Figure 1).

While speculative for FLA, in bacterial predators, such as Myxococcus xanthus, outer membrane vesicles (OMVs) appear to specifically carry a ‘personalized’ subset of proteins that only in-part are dictated by the predator’s genome (Zwarycz et al., 2020). In other words, Zwarycz et al. (2020) hypothesize that the non-genome-derived proteome fraction is there to target its prey. Hence, is L. pneumophila really the predator not amoebae, utilizing a large fraction of its genome to not only manipulate the hosts into providing a growth niche in the form of Legionella-containing vacuoles (LCV) but potentially also EVs production to increase its persistence/transmission? Furthermore, as seen with other EVs, those containing legionellae may also modulate/avoid human host immune responses (e.g., triggered by phosphatidylserine lipids on the surface of EVs (Cole and Nizet, 2016)).

In bacteria, outer membrane vesicles (OMVs) have also been shown to provide a defense mechanism to phage infection (Reyes-Robles et al., 2018). Hence, next genomic then viruses specific to FLA are discussed and how they may impact on the evolution of these important hosts and ARB.

As with their FLA hosts suited to the diversity of conditions in environmental biofilms, ARB maintain a larger genome than their relatives – so contradicting the expected genome reduction theory accepted for most intracellular pathogens (Moliner et al., 2010). This is likely due to the importance of gene transfer of pathogenicity factors, and the near 800 million year old arms-race between these predators and prey (Dupont et al., 2011; Shames, 2023) along with the role played by their respective viruses (Kaján et al., 2020). Of particular focus in this article is the remodeling of the Legionella-containing vesicle (LCV), with legionellae acquiring effector protein genes, such as Sar1/CopII that impact early secretory vesicle production genes (Robinson and Roy, 2006). As introduced in Figure 1, it is unknown how ARB may influence the release of EVs to enhance predator infection downstream, and provide a hot-spot for horizontal gene transfer (HGT). Recent work has pointed to FLA preferentially preying on non-ARB free bacteria over L. pneumophila (Shaheen and Ashbolt, 2021). What is unknown, but hypothesized here is that legionellae within EVs may not only avoid environmental stressors, but also overcome negative selective feeding on L. pneumophila by FLA. Indeed, are EVs preferential prey for FLA and an enhanced mechanism for prey to counter predators’ sensing of preferred food?

The role played by quorum sensing molecules (such as LAI-1 [3-hydroxypentadecane-4-one]) involved in legionellae changing from a reproductive to a transmissive phase (Schell et al., 2016), could also be part of the evolutionary arms race between predator and prey (Simon et al., 2015). It is currently unknown if LAI-like molecules are activated within EVs, as known to occur within trophozoites. Other potential changes within trophozoites and EVs that HGT may also impact could include environmental antimicrobial resistance (AMR) (Mortensen et al., 2021). AMR has generally not been viewed as important with pathogens that do not spread zoonotically or by person-to-person (such as with L. pneumophila). However, AMR within amoebal ecosystems increasingly impacted by AMR (e.g., cooling towers or irrigation systems receiving AMR-laden wastewater) could present a ‘perfect storm’ for enhance AMR within L. pneumophila-FLA ecosystems.

FLA evolved some 800-850 My ago (Cavalier-Smith, 2009), but have and continue to have exchange with bacteria/archaea and related viruses. A particular feature of FLA are their nucleocytoplasmic large DNA viruses (NCLDV), first described in 2003 (Abrahao et al., 2014) as Acanthamoeba polyphaga mimivirus (APMV) (now in the genus Mimivirus). In a recent whole genome analysis of A. polyphaga various NCLDVs (Marseillevirus, Mimivirus, Mollivirus, Pandoravirus, Pithovirus and a yet to be described family) were identified - inferring that NCLDV may have been domesticated during the evolution of amoebae (Schulz et al., 2020). Hence, though considered rather unusual, NCLDVs are not just a result of recent horizontal gene transfer (HGT) but in fact are associated with most major eukaryotic lineages, so impacting all ecosystems on earth (Schulz et al., 2020).

Another feature, potentially relevant to L. pneumophila’s human infectivity, has been described for human enteric viruses. A good exemplar is seen with human Norovirus, which change commensal gut bacteria’s EV production, likely impacting human host responses to infection (Mosby et al., 2022). While speculative for human infection by amoeba-legionellae material, it is intriguing that various enveloped and non-enveloped viruses generate a cytopathic effect in a range of amoebae and are released in amoebal EVs (Folkins et al., 2020; Dey et al., 2021). Hence, what roles may viruses play within amoebae (possibly in concert with internalized bacteria) regarding their expression of EVs and in modulating/avoiding human host immune responses? Important first insights to such mechanisms have been reported by Dey et al. (2022), who first described amoebic mitochondrial impacts by RNA viruses that appear to initiate and regulate apoptotic cell death. Overall this work points to research gaps that could be important to understand from an evolutionally perspective as well as to better describe human dose-response relationships for ARB pathogens.

Taking the above findings of NCLDV evolution, their likely (but not documented) presence within EVs and that specific genes from Legionella may be packaged within EVs – illustrates the is much to understand in the role that EVs may play in predator-prey interactions of ARB pathogens – and possibly much more to learn as to what actually leads to Legionella infection in human macrophages and alveolar epithelial cells.