Convergence of hormones and autoinducers at the host/pathogen interface
- First Online:
- Cite this article as:
- Rumbaugh, K.P. Anal Bioanal Chem (2007) 387: 425. doi:10.1007/s00216-006-0694-9
Most living organisms possess sophisticated cell-signaling networks in which lipid-based signals modulate biological effects such as cell differentiation, reproduction and immune responses. Acyl homoserine lactone (AHL) autoinducers are fatty acid-based signaling molecules synthesized by several Gram-negative bacteria that are used to coordinate gene expression in a process termed “quorum sensing” (QS). Recent evidence shows that autoinducers not only control gene expression in bacterial cells, but also alter gene expression in mammalian cells. These alterations include modulation of proinflammatory cytokines and induction of apoptosis. Some of these responses may have deleterious effects on the host’s immune response, thereby leading to increased bacterial pathogenesis. Prokaryotes and eukaryotes have cohabited for approximately two billion years, during which time they have been exposed to each others’ soluble signaling molecules. We postulate that organisms from the different kingdoms of nature have acquired mechanisms to sense and respond to each others signaling molecules, and we have named this process interkingdom signaling. We further propose that autoinducers, which exhibit structural and functional similarities to mammalian lipid-based hormones, are excellent candidates for mediating this interkingdom communication. Here we will compare and contrast bacterial QS systems with eukaryotic endocrine systems, and discuss the mechanisms by which autoinducers may exploit mammalian signal transduction pathways.
KeywordsQuorum sensing Autoinducer P. aeruginosa Interkingdom signaling Endocrine disruption
Introduction: the universality of hormone signaling
There are diverse arrays of lipid or fatty acid-based signaling molecules that modulate hundreds of biological effects in eukaryotic organisms. These signaling compounds include the eicosinoids or prostanoids, which include the prostaglandins, leukotrienes, and thromboxanes, and the lipidic and steroid hormones such as retinoic acid, estrogen, and testosterone (Fig. 1). One of the primary outcomes of hormone signaling pathways is the modulation of the transcriptome of the responding cell and alteration of cellular behavior. Although some lipid-based hormones, such as the prostanoids, can bind to membrane-associated receptors, many of these hydrophobic signaling compounds diffuse through cell membranes and directly affect transcription by interacting with intracellular transcription factors of the nuclear hormone receptor (NHR) superfamily (Fig. 2) . Like eukaryotes, bacteria also synthesize lipid-based signaling compounds. The lipidic autoinducers synthesized by many Gram-negative bacteria bear a striking resemblance to eukaryotic hormones, both in structure and in mechanism of action. Considering the intimacy between eukaryotes and their bacterial denizens over billions of years coevolution, we hypothesize that autoinducers function as bacterial hormones that alter mammalian gene transcription through symbiotic communication and/or deleterious endocrine disruption. We have termed this cross-talk interkingdom signaling, which is defined as the exploitation of signal transduction pathways by the signaling compounds of one organism to alter the behavior, through changes in gene transcription, of an organism from a different kingdom of nature. In this review, we will consider the possibility that bacterial adaptation has resulted in the use of autoinducers to exploit mammalian signaling pathways. We will first examine several examples of interkingdom signaling via hormones between organisms from other kingdoms of nature. We will highlight the similarities between hormones produced by a variety of organisms, discuss how hormones may be used for interkingdom signaling, and how bacteria may use autoinducers to manipulate the host response through a form of endocrine disruption.
Hormonal signaling in nature
Prokaryotic endocrine-like signaling networks
The Monerans are the most numerous and ubiquitous organisms on Earth. Kingdom Monera is comprised of prokaryotes, which lack a nucleus or other membrane-bounded organelles. Bacteria are the most ancient form of life, appearing on Earth ~3.5 billion years ago . Over 300 bacterial genomes have now been sequenced and they illustrate the tremendous phenotypic and metabolic diversity found in this kingdom (see http://www.ncbi.nih.gov/genomes/lproks.cgi/). Although 40% of the genomes from known bacterial species have been sequenced or are targeted for sequencing, we are still extremely limited in our understanding of the genomic diversity of bacteria, considering that an estimated >99% of bacterial species have not yet been cultured in the laboratory .
Although they are single-celled organisms, many bacteria have the capacity to work essentially as multicellular organisms by coordinating the expression of sets of genes within a population using a process called quorum sensing (QS). QS is a cell-density-dependent signaling mechanism through which bacteria judge their overall density and regulate diverse activities [9, 10]. The main components of QS systems are the autoinducer compounds made by the bacteria and the receptors for these molecules. A number of structurally distinct AHL autoinducers are synthesized by dozens of Gram-negative bacteria. AHL autoinducers consist of a homoserine lactone ring attached to an acyl side-chain, which can vary in terms of the number of carbons and the presence of modifications within the side-chain (Fig. 1). Autoinducer receptors in Gram-negative bacteria are primarily transcriptional regulators whose activities are modulated upon binding to their cognate ligand . A great deal of research over the last three decades has focused on delineating the regulatory cascades and solving the structures of the ligands and receptors that mediate QS in many bacteria. Consequently, there are now several well-established models to account for the QS-based mechanisms by which bacteria communicate with each other, and a number of excellent reviews have been published on this topic (for example, see [11, 12, 13]). However, several groups have recognized that many bacterial signaling compounds, including AHLs, are functionally and structurally similar to eukaryotic lipid-based hormones (Fig. 1), and there is now growing evidence that AHLs can elicit biological effects in eukaryotic cells. The evidence that supports this concept and the implications of these signaling pathways for pathogenic and mutualistic relationships between bacteria and their hosts will be discussed in this review. Before delving deeper into bacterial–mammalian interactions, we will summarize evidence for interkingdom signaling between organisms from other kingdoms of nature.
Signal interference by fungal metabolites
Fungi produce hormone-like compounds that can cause severe damage to mammals. A subset of fungal metabolic byproducts, also called mycotoxins, are a diverse group of chemicals that are not necessary for fungal growth but are thought to function as virulence factors that facilitate host tissue invasion and/or antibiotics targeted at competing microbes. Mycotoxins can have detrimental effects on mammalian cells by interfering with steroid binding. For example, Fusarium spp. produce an estrogenic resorcylic acid lactone compound called zearalenone that induces sporulation. However zearalenone can also cause endocrine disruption in many mammals, such as domestic livestock that ingest contaminated feed . Binding of zearalenone to mammalian estrogen receptors can result in enlarged mammae, swelling of the uterus and vulva, atrophy of the ovaries, and prolapse of the vulva and rectum .
Like bacteria, fungi use QS systems to coordinate gene expression. Farnesol is an isoprenyl alcohol that has been identified as a QS compound synthesized by C. albicans that is responsible for inhibiting hyphal formation in stationary-phase cells  and biofilm formation . Tyrosol, a second QS compound made by C. albicans, abolishes the lag phase in stationary-phase cells . Interestingly, one of these compounds, farnesol, is also produced by plants, insects, and mammals (in addition to fungi). It regulates developmental processes in several kingdoms of nature and has anticancer effects in human cells [19, 20]. Farnesol exerts its effects on gene transcription through activation of nuclear hormone receptors, including the farnesoid X receptor (FXR) , the peroxisome proliferator-activated receptors-alpha (PPARα) and -gamma (PPARγ) , and thyroid hormone receptor (THR) . Since the farnesols produced by these different organisms are structurally identical, this ubiquitous chemical is an excellent candidate for an interkingdom signaling mediator.
Plant interkingdom signaling via phytohormones
Interkingdom signaling by hormones between plants and bacteria, fungi and mammals has been described. These hormones can function as communication signals in symbiotic relationships, or more commonly as “jamming signals” in pathogenic relationships. Plants produce hormones, or phytohormones (including auxin, jasmonic acid, cytokinin, gibberellins, abscisic acid, ethylene, and brassinosteroids), that mediate endogenous functions including the regulation of growth, photosynthesis, senescence, ovulation, scent production, and coloring . However, phytohormones also appear to have exogenous functions. For example, phytohormones are one component of a large arsenal of chemicals that plants use to protect themselves from bacteria, fungi, insects and mammals. These chemical defenses may for example produce a bad taste to repel a predatory herbivore, or may attract carnivores which in turn eat the herbivores, thus preserving the plant [24, 25]. It has also been reported that plant phytohormones can inhibit the sexual reproduction of predatory herbivores . This disruption of normal hormonal signaling by exogenous ligands is referred to as endocrine disruption and will be discussed further below.
Phytohormone interkingdom signaling is not limited to interactions between plants and mammals, but can also occur between plants and fungi or bacteria, and this signaling can benefit the pathogen or the plant. For example, some plant pathogens such as Erwinia spp. produce phytohormone derivatives that their plant hosts cannot distinguish from their own endogenous chemicals. The resulting excess of phytohormones can result in disease states, such as tumor or gall formation, which alter the metabolism of the host and provide an attractive niche for bacteria . Interkingdom signaling via phytohormones can also serve to recruit symbiotes which provide growth advantages for plants, as in the relationship of Rhizobium soil bacteria to leguminous plants. In this symbiotic relationship, legumes produce flavanoids that serve as ligands for bacterial NodD receptor proteins. Liganded NodD activates transcription of genes encoding nod factors, which bind to receptors on plant roots and initiate nodulation. Nodules serve as a protected niche for the nitrogen-fixing symbiotes, which in turn enhance the soil microenvironment and promote growth of the plant .
Endocrine systems and disruption in Kingdom Animalia
Insect pheromone signaling
The Kingdom Animalia is comprised of vertebrates and invertebrates, and members of both groups synthesize and respond to a diverse range of hormones and hormone-like chemicals. In the insect world, these signaling compounds are more commonly referred to as pheromones and they control most aspects of development including moulting, pupation, and metamorphosis [1, 29]. In addition to physical development, insect pheromones influence behavior. For example, in a manner similar to QS bacteria living in a biofilm, ants live in large communities and produce chemical pheromones that alter the behavior of the group as a whole. Two well-known examples are the trailing and alarm pheromones, which send signals to other members of the colony to avoid or follow a chemical concentration gradient. Pheromone perception is mediated for the most part through odorant receptors, a superfamily of seven-transmembrane domain receptors found in high concentrations in insect chemosensory organs (Fig. 2) . The binding of an odorant to these receptors initiates a signaling pathway that relays a message about the environment to the neural network of the organism. Insect pheromones that act as odorants are typically volatile and soluble but are structurally diverse .
Autoinducer recognition and responses by invertebrates
The soil nematode Caenorhabditis elegans has been used extensively as a model invertebrate organism, and much information concerning the physiology of higher animals has been gathered through studies in C. elegans. Like insects, C. elegans monitors its environment through chemoperception and the odorant receptors involved have been studied intensely [31, 32]. C. elegans can sense a wide array of chemicals which act as attractants or repellents , and can use this sensory information to find bacterial food sources or to avoid bacterial pathogens [33, 34]. After exposure to pathogenic strains of bacteria, C. elegans “learns” to avoid these strains upon subsequent exposures . We recently discovered that this process appears to involve the perception of AHL autoinducers . In our study, C. elegans was attracted to AHLs produced by Pseudomonas aeruginosa and Agrobacterium tumefaciens, and chemotaxed towards them. We postulate that C. elegans uses chemosensory mechanisms to find bacterial food sources in nature, and that autoinducers, which are expected to be present in high concentrations around biofilms, are candidate attractive odorants. Some strains of bacteria, including P. aeruginosa, can kill C. elegans via QS-controlled virulence factors [36, 37]. Therefore, we also examined whether C. elegans could “learn” to avoid an autoinducer made by a pathogenic strain of bacteria. Remarkably, while naïve worms that had not been exposed to P. aeruginosa were attracted to autoinducers, worms that had been exposed to pathogenic, autoinducer-producing strains of P. aeruginosa were subsequently repelled by it . Furthermore, we determined that this learned aversion to autoinducers was mediated through serotonin signaling pathways. Therefore, perception of AHLs made by pathogenic bacteria and subsequent aversive learning may enhance the survival of C. elegans in nature. Thus, it is possible that C. elegans can be used as a model organism for investigating the mechanisms controlling interkingdom signaling by bacterial autoinducers, just as it has been used for bacterial pathogenesis.
Mammalian endocrinology and endocrine disruption
The mammalian endocrine system is a complex network composed of several glands or groups of specialized cells located throughout the body that function to secrete hormones. Hormones are transported through the bloodstream and typically act at sites distant from their origin. Mammals secrete a vast array of structurally diverse hormones, which are involved in just about every biological process including cell differentiation, reproduction, metabolism and immune responses. Hormone levels are held in a delicate balance and disruptions in hormone signaling can have dire consequences such as reproductive interference or cancer . Endocrine disruption occurs when environmental chemicals mimic hormones, bind to nuclear or membrane-bound receptors, and interfere with the binding of natural hormone ligands . For this to occur, the hormone mimic must have affinity for a receptor and either agonize the receptor by binding and initiating an abnormal response, or antagonize it by blocking the native hormone from binding or exerting its normal effect. A diverse group of compounds are known to function as disruptors of endocrine function in mammals, including the antioxidant flavanoids produced by some fruits and vegetables, and synthetically produced pesticides . In mammals, endocrine-disrupting signals can bind to many receptors including the estrogen receptor (ER), the thyroid hormone receptor, and other orphan nuclear receptors . However, endocrine disruption is not limited to mammals. Endocrine disruption has been reported in every phylum investigated to date, including crustaceans, fish and birds . In fact, even bacteria are subject to endocrine disruption by environmental chemicals. As mentioned previously, many leguminous plants rely on symbiotic nitrogen-fixing bacteria, such as Rhizobium spp., for growth. In this relationship, symbiotes are recruited to plant roots by phytoestogen signals. However, it has recently been shown that many classes of environmental endocrine disruptors such as insecticides, herbicides, and fungicides can disrupt this interkingdom signaling, presumably by binding to bacterial receptors [1, 39]. Interestingly, these same classes of endocrine disruptors can also cause health concerns in mammals by targeting hormone receptors such as estrogen receptors (for example, binding of zearalenone to mammalian estrogen receptors, as mentioned before). Interestingly, the Rhizobium NodD phytoestrogen receptor and the mammalian estrogen receptor share significant structural and functional homology, although they are not orthologously related . Thus, this may be an example of interkingdom signaling that emerged via convergent evolution .
Bacteria face the difficult challenge of synthesizing, receiving and interpreting their own signaling molecules amongst the cacophony of competing signals in the environment. Exogenous signals bacteria have to contend with may originate from other species of bacteria, insects, plants, fungi or animals. For example, bacteria colonizing a mammalian host are influenced by the hormonal environment, which affects the growth, metabolism and virulence of its flora . “Microbial endocrinology,” which is the consideration of the host’s hormonal environment on the growth, metabolism and virulence of pathogenic bacteria, has become an emerging field of study . Strikingly, these competing signals also appear to affect QS, as the mammalian hormones epinephrine and norepinephrine can mimic the native autoinducers made by enterohemorrhagic Escherichia coli, inducing its QS systems . The nature of this specific interaction is not clear at present, but host manipulation of bacterial QS systems by hormone mimics could benefit the host. For example, this “cross-talk” could force the invading bacteria to prematurely induce QS, resulting in the secretion of antigens that alert the host immune system and promote bacterial clearance. Conversely, it can also be imagined that hormone/autoinducer crosstalk promotes the persistence of normal flora. For example, commensals may produce “friendly” autoinducers that signal the cells lining the gut to tolerate them. Thus, endocrine signaling or disruption may be a mechanism by which evolutionarily diverse organisms interact.
Host immune modulation by bacterial AHLs
To date, studies examining the effects of autoinducers on the host immune response have primarily focused on three autoinducers made by P. aeruginosa, two AHLs (N-(3-oxododecanoyl)-L-homoserine lactone (3OC12-HSL), N-butanoyl-L-homoserine lactone (C4-HSL)), and the Pseudomonas quinolone signal (PQS; 2-heptyl-3-hydroxy-4-(1H)-quinolone) (Fig. 1). These autoinducers have pleiotropic effects on the host response, which depend on the assay conditions and identity of the responding cell type. However, many of their effects are similar to mammalian hormonal responses and are likely to have significant physiological consequences. A summary of the studies that documented these effects is presented below.
AHL blunting of the immune system
While AHLs are known to regulate important bacterial virulence factors through QS, which in turn affect the host, the concept of AHLs directly influencing the host immune system and acting as independent virulence factors originated almost a decade ago in a landmark paper by Telford et al. . In this study, the P. aeruginosa AHL 3OC12-HSL inhibited the production of the cytokines interleukin-12 (IL-12) and tumor necrosis factor alpha (TNF-α) by LPS-stimulated macrophages. 3OC12-HSL also inhibited the proliferation of activated T-cells and the release of IL-2 in vitro [42, 43], leading the authors to speculate that AHLs may shift the host protective T-helper-1 (Th1) response towards the pathogen protective T-helper-2 (Th2) response. The PQS autoinducer also inhibited T-cell proliferation in these studies . Further investigations into the action of AHLs on T-cells revealed that 3OC12-HSL inhibited the differentiation of both Th1 and Th2 cells; however, the influence of 3OC12-HSL on cytokine production by either Th1 or Th2 cells varied depending on the antigen treatment and the background of the mice used . The effects of 3OC12-HSL on T-cells were also shown to occur rapidly within 1–2 hours of stimulation by 3OC12-HSL . These data suggested a scenario whereby, upon infection, P. aeruginosa reaches a high cell density and AHL production is initiated, which facilitates QS in the bacterial population, but also blunts the host’s adaptive immune response and leads to increased bacterial persistence. This phenomenon could help explain why P. aeruginosa produces chronic infections in diseased lungs or in diabetic wounds for instance. A somewhat related study showed that 3OC12-HSL effectively alleviated the autoimmune inflammation associated with diabetes in nonobese diabetic mice . Therefore, the potential exists to use AHLs as anti-inflammatory agents to treat autoimmune disorders in humans.
AHL-induced proinflammatory response
In another series of studies, Smith et al. showed that AHLs can also induce proinflammatory responses in fibroblasts, both in vitro and in vivo [47, 48, 49]. These studies showed that 3OC12-HSL induced the expression of genes encoding the chemokine interleukin-8 (IL-8), cyclooxygense 2 (Cox-2) and prostaglandin E2 synthase, and enhanced prostaglandin E2 production in human lung fibroblasts and epithelial cells [47, 48]. Similar responses were observed in vivo when purified 3OC12-HSL was injected into the skins of mice. Dermal injections of 3OC12-HSL induced Cox-2 as well as interleukin-1 (IL-1), IL-6 and the chemokines macrophage inflammatory protein 2 (MIP-2), monocyte chemotactic protein 1, MIP-1β, inducible protein 10, and T-cell activation gene 3 . These studies identified NFκB and AP-2 as two potential transcription factors that mediate 3OC12-HSL activity [47, 48]. We recently confirmed these data and detected a 3OC12-HSL-induced proinflammatory response in mouse fibroblasts and human vascular endothelial cells . Thus, 3OC12-HSL may contribute to another hallmark of many acute P. aeruginosa infections, which is uncontrolled inflammation and subsequent destruction of host tissue, which in turn leads to dissemination of the bacteria and consequent sepsis.
Death by AHL: induction of apoptosis
A third mechanism of 3OC12-HSL action on host cells is the induction of programmed cell death. 3OC12-HSL-induced apoptosis has been observed in several cell types, including neutrophils and macrophages [51, 52], breast carcinoma cells , and fibroblasts and vascular endothelial cells . Apoptotic induction appears to be dose-dependent and specific, as other autoinducers, such as C4-HSL, do not cause apoptosis. In fact, analysis of the 3OC12-HSL structure revealed that the presence of the 3-oxo group in the acyl side-chain and the homoserine lactone moiety of the L-isoform were both essential for apoptosis in macrophages . 3OC12-HSL-induced apoptosis is also cell-type-specific, as epithelial cells appear to be resistant to 3OC12-HSL-induced apoptosis . Apoptosis in neutrophils and macrophages was associated with elevated caspase 3 and 8 activities , indicating that 3OC12-HSL-induced apoptosis may occur via the extrinsic or death receptor pathway, rather than the intrinsic or mitochondrion pathway. The extrinsic pathway is initiated by the binding of activators such as Fas ligand and tumor necrosis factor alpha (TNFα) to their respective transmembrane receptors. This binding sets into motion a signaling cascade that results in apoptosis .
Effective in vivo concentrations of AHLs
As described above, 3OC12-HSL has been reported to both promote inflammation and to suppress inflammation in host cells. The primary difference between these two sets of studies was the AHL dose used. In general, apoptosis and proinflammatory responses have been observed at higher concentrations (50–100 μM), while anti-inflammatory effects have been seen at lower concentrations (<10 μM) of 3OC12-HSL. We previously suggested that these different effects could result from the existence of a continuum of responses of mammalian cells to different concentrations of AHLs . Unfortunately, the concentrations of AHLs to which cells are exposed to in vivo in different infection scenarios are unknown. To date, only a handful of studies have attempted to detect autoinducers in vivo, all of which have focused on P. aeruginosa lung infections [55, 56, 57, 58]. In these studies, 3OC12-HSL was detected in the nanomolar range in the sputum , and even lower (fmol/gram) in the lung tissue . These measurements presumably represent significant underestimations of the local AHL concentrations in the vicinity of biofilms within the infected lungs. AHLs clearly become diluted in lung fluid and sputum, although the specific dilution factor is unknown. Another potential confounding factor is that the current bioassays used to measure AHLs  suffer from sensitivity limitations, such as significant post-synthesis handling during which AHLs may be subject to degradation. In addition, most studies have been performed on cells in culture and different cell lines may respond differently to autoinducer. For example, we recently showed that nonimmortalized cells appear to be more sensitive to AHLs than immortalized cell lines .
In contrast to the existing “in vivo” measurements, AHL concentrations as high as 600 μM have been measured in the supernatant of biofilms grown in vitro . Therefore, it is quite likely that there are “local” and systemic concentrations in vivo, where cells in the immediate vicinity of a high concentration of bacteria (i.e., biofilm) are exposed to high concentrations of AHLs, which are progressively diluted out further away from the biofilm. Importantly, we have recently shown that P. aeruginosa forms biofilms preferentially around blood vessels in a mouse acute wound model . Thus, it is likely that cells located immediately adjacent to biofilms in vivo are exposed to high local concentrations of AHLs; however, the “systemic” concentrations in the wound fluid may be substantially less. More sensitive, real-time assays are needed to determine accurate in vivo AHL concentrations, and future in vitro studies investigating the mechanisms of AHL immune modulation should thoroughly consider the applicability of the cell line being tested, its relevance to human disease, as well as whether the observed effects are dose-responsive.
QS autoinducers as mammalian receptor ligands
To date, no eukaryotic proteins have been identified as bona fide receptors for bacterial autoinducers. However, considering the variety of effects of AHLs on eukaryotic cells, it is likely that AHLs act as ligands for one or more eukaryotic receptors. Although autoinducer-regulated signal transduction pathways have been investigated, they have primarily identified intermediates or terminal targets of 3OC12-HSL signaling. For example, some of the proinflammatory effects of 3OC12-HSL may be mediated by NFκB and AP-2 [47, 48]; however, it is unlikely that 3OC12-HSL directly influences the activity of these mammalian transcription factors, as they do not appear to be regulated by direct binding of lipid-based ligands. Two members of the MAP kinase family of serine/threonine kinases, ERK-1 and ERK-2, are activated by 3OC12-HSL in lung epithelial cells , and pharmacological inhibitors of these MAP kinases blocked 3OC12-HSL-mediated activation and placed ERKs upstream of NFκB. Again, MAP kinases are unlikely to be direct targets of 3O-C12-HSL, and they are not associated with 3O-C12-HSL signaling in all cell types, as 3OC12-HSL did not activate MAPK cascades in breast cancer cells . Another transcription factor whose activity was regulated by 3O-C12-HSL was signal transducer and activator of transcription (STAT) 3, whose activity was inhibited in breast cancer cells exposed to this autoinducer. The Akt/PKB pathway was also found to be partially inhibited in this same study, and these autoinducer effects were associated with apoptotic induction . Finally, we have recently shown that treatment with 3OC12-HSL increased intracellular calcium levels in mouse fibroblast and human vascular endothelial cells, and that calcium signaling was associated with apoptosis  (Fig. 3).
Based primarily on our own studies, we speculate that there are at least two distinct AHL receptors in mammalian cells. One putative receptor resides at or close to the host cell membrane and appears to mediate apoptotic induction via inositol triphosphate (IP3) and Ca2+ second messengers. Some of the most likely candidates for this protein are members of the G-protein-coupled receptor (GPCR) superfamily (Fig. 2). GPCRs bind many hormones and 30–40% of all medicinal drugs and mediate their primary signaling effects through phospholipase C. There are over 800 GPCR-encoding genes in the human genome, which represents the largest and most versatile group of membrane receptors . The serotonin receptors, of which there are up to thirteen in mammals, are GPCRs and have been implicated in mediating autoinducer signaling in C. elegans . However, serotonin receptors are primarily localized in neural tissue and are thus unlikely candidates for the apoptosis-associated receptor in fibroblasts and other cell types. Thus, identification of a bona fide AHL receptor in this superfamily is a major challenge for investigators in this area.
The rationale for proposing the existence of a second AHL receptor is based on our observation that the inhibition of calcium signaling pathways in fibroblasts only blocked apoptotic induction and did not affect pro-inflammatory gene induction by 3OC12-HSL. Based on the existing data, there are at least two possible locations for this second putative receptor. We have previously shown that 3OC12-HSL can enter and retain its function inside mammalian cells . Thus, it is possible that AHLs can diffuse into mammalian cells and interact with an intracellular AHL receptor. NHRs, which are ligand-activated proteins that usually reside in inactive complexes in the cytosol or nucleus of eukaryotic cells, are attractive candidates for such a receptor (Fig. 2). Classic NHRs, which include the steroid hormone receptors, are maintained in inactive states by binding to chaperones or transcriptional corepressors in the absence of their cognate ligands. These proteins undergo conformational changes and cofactor exchange upon ligand binding, resulting in their conversion to active transcription factors. This model is particularly appealing, as NHRs are frequent targets of endocrine disruptors, and there are clear functional similarities between AHL signaling in bacteria and steroid hormone signaling in mammals. Mammalian genomes contain up to 50 different NHR-encoding genes, and our preliminary data indicate that a subset of NHRs do respond to AHLs in transfection assays. Experiments are currently underway to determine whether these NHRs are AHL targets in vivo.
Alternatively, the identification of the ERK1 and ERK2 MAP kinases as AHL-dependent signaling intermediates in proinflammatory responses suggests that this second putative receptor might also be located at or close to the membrane. MAP kinase pathways are activated by a wide variety of external stimuli, including many growth factors and other mitogens. MAP kinase cascades are commonly initiated by the binding of a ligand to a tyrosine kinase receptor and the subsequent activation of the small GTP binding protein, Ras. In addition, the activation of some GPCRs may also lead to MAP kinase cascade activation. Thus, it is possible that a second membrane-associated AHL receptor mediates the proinflammatory effects of AHLs, although it is formally possible that apoptosis and inflammatory effects of AHLs are mediated through the differential activation of two separate signaling pathways by the same receptor. Clearly, these questions must await the elucidation of the specific receptors for these two biological responses.
The examples described in this review highlight the fact that hormone-like signaling molecules are produced by organisms in every kingdom of nature. Hormones are defined as “products of living cells that circulate in body fluids or sap and produce a specific effect on the activity of cells remote from their point of origin.” Bearing this definition in mind, it would appear that bacterial AHL autoinducers share several properties with bona fide hormones. For example, bacterial autoinducers are secreted factors that enter a variety of environmental milieux. Furthermore, they influence cellular activity, and can be found in the bodily fluids of humans infected with AHL-producing bacteria [56, 58]. Their effects are likely to be primarily paracrine in nature, affecting cells in the immediate vicinity of bacterial biofilms, as it is unclear whether AHL concentrations in serum would ever be high enough to affect distant cells. Further biochemical and molecular analysis of putative AHL receptors and their signaling pathways will be needed to confirm the physiological relevance of these concepts.
As a final thought, it is intriguing to consider some potential broader implications of AHL-based interkingdom signaling in interactions between bacteria and their eukaryotic hosts. The effects of the P. aeruginosa autoinducer 3OC12-HSL on host cells described here have been examined primarily within the context of pathogenesis. To obtain additional data on the effects of autoinducers on mammalian cells, we performed microarray analysis of mouse fibroblasts treated with 3OC12-HSL or C4-HSL compared to vehicle controls. We found that 3OC12-HSL and C4-HSL elicited significant changes in mRNAs derived from 10% and 8%, respectively, of murine genes represented on Affymetrix chips (K.P. Rumbaugh, unpublished results). A small subset of these genes were affected by both 3OC12-HSL and C4-HSL, while the majority exhibited autoinducer specificity. Although many of the differentially expressed genes fall into categories of expected or known targets of AHL action, such as genes associated with inflammation and apoptosis, the majority of the affected genes are novel autoinducer targets that fall into numerous functional gene ontology categories. Based on these results indicating the presence of a large number of autoinducer-responsive genes in a single cell type, we speculate that autoinducer-dependent interkingdom signaling may target a variety of additional cellular processes. For example, it is possible that many “friendly” interactions between commensals and their hosts might also rely on hormonal signaling to maintain symbiotic relationships. Given that there are approximately ten times as many bacterial cells than human cells in the human body , it is not unreasonable to speculate that small molecule signaling systems such as those described here might represent a form of communication across the prokaryote/eukaryote divide that is ultimately beneficial to all parties involved. Identification of the components of these communication networks and the unraveling of their evolutionary origins promises to be a fascinating story for researchers interested in interkingdom signaling.
Work in K.P. Rumbaugh’s lab is supported by the American Lung Association. Many thanks to Simon Williams for his conceptual and editorial contributions.