Chlamydial metabolism revisited: interspecies metabolic variability and developmental stage-specific physiologic activities

Abstract

Chlamydiae are a group of obligate intracellular bacteria comprising important human and animal pathogens as well as symbionts of ubiquitous protists. They are characterized by a developmental cycle including two main morphologically and physiologically distinct stages, the replicating reticulate body and the infectious non-dividing elementary body. In this review we reconstruct the history of studies that have led to our current perception of chlamydial physiology, focusing on their energy and central carbon metabolism. We then compare the metabolic capabilities of pathogenic and environmental chlamydiae highlighting interspecies variability among the metabolically more flexible environmental strains. We discuss recent findings suggesting that chlamydiae may not live as energy parasites throughout the developmental cycle and that elementary bodies are not metabolically inert but exhibit metabolic activity under appropriate axenic conditions. The observed host-free metabolic activity of elementary bodies may reflect adequate recapitulation of the intracellular environment, but there is evidence that this activity is biologically relevant and required for extracellular survival and maintenance of infectivity. The recent discoveries call for a reconsideration of chlamydial metabolism and future in-depth analyses to better understand how species- and stage-specific differences in chlamydial physiology may affect virulence, tissue tropism, and host adaptation.

Graphical Abstract

Introduction

Bacteria constitute a highly successful and diverse group of organisms, inhabiting almost all niches on Earth. Central to their success is their exceptional capacity to adapt to a wide range of physical and chemical conditions, their versatile physiology, and their potential to carry out a broad spectrum of metabolic conversions to enable energy generation and biomass production (Hoehler & Jorgensen, 2013, Simon & Klotz, 2013). Many bacteria live in intimate association with other prokaryotes or eukaryotes, and some – collectively referred to as intracellular bacteria – have specialized on survival and growth in a rather peculiar niche, the interior of eukaryotic cells (Moulder, 1974, Casadevall, 2008). Intracellular bacteria often have profound effects on their host and indeed include some of the most important pathogens of humans, animals, and plants, as well as essential symbionts on which their hosts rely for survival and reproduction (Dale & Moran, 2006, Palmer & Azad, 2012, Tan & Bavoil, 2012).

Facultative intracellular bacteria such as Salmonella or Listeria species retained the capacity to thrive in a host-cell free environment but also replicate within eukaryotic host cells. In contrast, extracellular growth has not yet been reported for several other intracellular bacteria, e.g. Chlamydia, Rickettsia, or Buchnera species. As a consequence of their adaptation to an intracellular lifestyle, these obligate intracellular bacteria display highly reduced metabolic capabilities when compared to their free-living relatives and are thus dependent on a broad range of host-derived metabolic precursors and cofactors that would not be readily available outside of eukaryotic host cells (Zientz, et al., 2004, Eisenreich, et al., 2010, Fuchs, et al., 2012). Growth and nutrient requirements are difficult to predict even if genome sequences are available, and axenic (i.e. host-free) cultivation is thus still not possible for the majority of these bacteria. This has significantly hampered the analysis of obligate intracellular bacteria, particularly with respect to their physiology. The Chlamydiae constitute one of the most successful groups of obligate intracellular bacteria. Their investigation has a long history, and many of the former mysteries indeed seemed to be resolved with the advent of modern genomics.

However, more recent investigations not only led to an increasing awareness of a previously unrecognized diversity within this group of bacteria, but also demonstrated a necessity to reconsider basic aspects of their physiology and to revise and partly revive former interpretations in the light of new discoveries. In this review, early findings and hypotheses, current knowledge, and new concepts of chlamydial metabolism are thus summarized and discussed with the aim to highlight important areas for future research. We focus mainly on energy and central carbon metabolism and discuss in more detail the differences between phylogenetically distinct members of the Chlamydiae as well as recent studies on host-free metabolic activity of chlamydiae. Finally, implications and future perspectives that arise from the recent discoveries will be addressed.

The Chlamydiaceae – major pathogens of humans and animals

The Chlamydiae exclusively comprise obligate intracellular bacterial pathogens and symbionts of eukaryotic cells (Horn, 2011). Perhaps best known within this phylum is the family Chlamydiaceae that includes the etiologic agents of a number of important diseases of humans and livestock (Kuo & Stephens, 2011). The major human pathogenic species, Chlamydia trachomatis, is comprised of over 15 serologically defined variants. Endemic blinding trachoma, caused by serovars A-C (Mabey, et al., 2003, Wright, et al., 2008), has been known since ancient times (Ghaliounghui, 1987, Taylor, 2008) and continues to be a major health concern in less developed countries, currently affecting about 21.4 million people, of whom about 2.2 millions are visually impaired and 1.2 million are blind (Burton & Mabey, 2009). (http://www.who.int/blindness/causes/priority/en/index.html). Serovars D-K are a leading cause of sexually transmitted disease (Schachter, 1999, Bébéar & de Barbeyrac, 2009) (http://www.who.int/vaccine_research/diseases/soa_std/en/index1.html). Chlamydial genital tract infections often remain asymptomatic and thus unrecognized. As a consequence, severe complications, such as ectopic pregnancy or infertility, can arise (Schachter, 1999, Bébéar & de Barbeyrac, 2009). C. trachomatis serovars L1, L2, and L3 are also transmitted through sexual contact, but cause lymphogranuloma venereum (LGV), a more systemic infection that disseminates to lymph nodes (Schachter, 1999, Bébéar & de Barbeyrac, 2009). A more recently identified human pathogen and member of the Chlamydiaceae, Chlamydia pneumoniae, is mainly considered to be an agent of respiratory disease accounting for approximately 10% of cases of community-acquired pneumonia (Grayston, et al., 1993, Burillo & Bouza, 2010). C. pneumoniae has also been associated with multiple chronic diseases, such as asthma, chronic obstructive pulmonary disease (COPD), atherosclerosis, and Alzheimer’s disease (Balin, et al., 1998, Campbell & Kuo, 2004, Blasi, et al., 2009, Shima, et al., 2010).

Other members of the Chlamydiaceae including Chlamydia psittaci, C. suis, C. felis, C. caviae, C. muridarum, C. percorum, and C. abortus are mainly known to infect animals, in particular mammals but also birds. Many of these species are of veterinary importance since they can cause abortions in livestock and consequently great economic damage. Furthermore, some of these species have the ability to cause severe zoonotic infections, including pneumonia and miscarriage, when transmitted to man (Longbottom & Coulter, 2003, Beeckman & Vanrompay, 2009, Wheelhouse & Longbottom, 2012).

Environmental chlamydiae – a hidden diversity

While the Chlamydiaceae undoubtedly remain by far the best studied representatives of the Chlamydiae, our perception of chlamydial diversity radically changed in the 1990s with the discovery of the first “Chlamydia-like” bacteria (Kahane, et al., 1995, Amann, et al., 1997, Birtles, et al., 1997, Fritsche, et al., 2000, Horn, et al., 2000); today also referred to as environmental chlamydiae (Horn, 2008). These bacteria represent the closest relatives of the pathogenic Chlamydiaceae and currently constitute eight additional families in the phylum Chlamydiae (Horn, 2008, Steigen, et al., 2013, Stride, et al., 2013) (Figure 1). Moreover, molecular evidence suggests an even greater family-level diversity (Ossewaarde & Meijer, 1999, Horn & Wagner, 2001, Corsaro, et al., 2003, Horn, 2008), which may exceed the known diversity by an order of magnitude (Lagkouvardos, et al., 2013). All characterized environmental chlamydiae display an obligate intracellular lifestyle similar to that of the Chlamydiaceae (Horn, 2011, Kuo & Stephens, 2011), yet these bacteria naturally thrive in a diverse range of host species. Indeed Chlamydiae are known to infect all major groups of vertebrates (mammals, birds, reptiles, amphibians, fish) and many groups of invertebrates (e.g. insects, crustaceans, mollusks), and protists may serve as hosts for many of them (Everett, 2000, Corsaro, et al., 2003, Horn, 2008).

Figure 1. Diversity of the Chlamydiae.

The phylogenetic tree shows the relationship of pathogenic chlamydiae (Chlamydiaceae) with other families collectively referred to as environmental chlamydiae. Representative species are shown, and their genome sizes relative to the 3.1 Mb genome of Parachlamydia acanthamoebae are indicated as pie charts. Note that the names Parilichlamydiaceae and Actinochlamydiaceae were independently proposed for what might represent a single family, here indicated as Parilichlamydiaceae (Horn, 2008, Steigen, et al., 2013, Stride, et al., 2013). The unrooted tree is a based on a recent survey of chlamydial diversity in metagenomic and amplicon datasets (Lagkouvardos, et al., 2013).

Although the term “environmental chlamydiae” is commonly used to collectively refer to all non-Chlamydiaceae species in the phylum Chlamydiae (opposed to the term “pathogenic chlamydiae” reserved for members of the Chlamydiaceae), it should be noted that environmental chlamydiae represent a diverse group of bacteria displaying major variability with respect to morphology, growth characteristics, genomic repertoire, host range, and ecology. In fact, a clear-cut distinction between environmental and pathogenic species becomes difficult if considering that some members of the Chlamydiaceae seem to be more widespread in the environment than previously recognized (Bodetti, et al., 2002, Horn, 2008).

Further, there is evidence for a potential pathogenicity of some environmental chlamydiae (Corsaro & Greub, 2006, Horn, 2008), such as Simkania negevensis (family Simkaniaceae) and Waddlia chondrophila (family Waddliaceae), which can thrive both in amoebae and in mammalian cells, and which were associated with respiratory disease and/or miscarriage in humans (Dilbeck, et al., 1990, Rurangirwa, et al., 1999, Kahane, et al., 2001, Henning, et al., 2002, Friedman, et al., 2003, Michel, et al., 2004, Michel, et al., 2005, Kahane, et al., 2007, Goy, et al., 2008, Haider, et al., 2008, Kahane, et al., 2008, Baud & Greub, 2011, Kebbi-Beghdadi, et al., 2011). In contrast, the Parachlamydiaceae, including Parachlamydia acanthamoebae and Protochlamydia amoebophila, are primarily considered symbionts of amoebae (Horn, 2008) and appear to have a limited capacity to grow in non-protozoan host cells (Maurin, et al., 2002, Greub, et al., 2003, Collingro, et al., 2005, Casson, et al., 2006, Hayashi, et al., 2010, Roger, et al., 2010, Kebbi-Beghdadi, et al., 2011, Sixt, et al., 2012). The study of these environmental counterparts of the medically important Chlamydiaceae represents a unique opportunity to further define variability and characteristic features of chlamydial biology as well as to identify factors contributing to pathogenicity and host adaptation.

The chlamydial developmental cycle

All chlamydiae share a similar developmental cycle that consists of transitions between two main cell types, the infectious non-dividing extracellular form, known as the elementary body (EB), and an intracellular replicative form, called the reticulate body (RB) (Matsumoto, 1988, Moulder, 1991, Abdelrahman & Belland, 2005). The concept of chlamydial development has major implications on chlamydial physiology. Following endocytosis, EBs remain within a membrane-bound compartment termed an inclusion where they begin to differentiate to RBs. RBs divide by binary fission for the duration of the intracellular developmental cycle but asynchronously begin to differentiate back into EBs which accumulate within the inclusion until release from the host cell by lysis or extrusion (Figure 2) (Todd & Caldwell, 1985, Hybiske & Stephens, 2007, Lutter, et al., 2013).

Figure 2. The chlamydial developmental cycle.

The Chlamydia trachomatis developmental cycle is depicted here with approximate times indicated for an LGV strain. Other strains or species grow at different rates. Infection is intiated by the attachment and internalization of an EB which remains within a non-fusogenic vesicle known as the inclusion. Once committed to development, Chlamydia EBs rapidly lose infectivity; one of the first indications of their differentiation to RBs. C. trachomatis EBs increase in size as they differentiate to RBs over the first 10 – 12 h before they initiate multiplication. After about 18 h (for the LGV strains, longer for ocular and genital strains of C. trachomatis and other chlamydial species), the RBs continue to multiply even as a sub-set asynchronously begins to differentiate back to EBs, which accumulate within the inclusion until release by lysis or extrusion at around 48 h or later.

In contrast to RBs, which are morphologically similar among all members of the Chlamydiae, EBs vary significantly in size and shape (Siegl & Horn, 2012). The infective stage of Chlamydiaceae is usually considerably smaller than the replicative stage (0.3 μm for EBs compared to 1.0 μm for RBs in the case of C. trachomatis) (Friis, 1972, Matsumoto, 1988, Moulder, 1991, Abdelrahman & Belland, 2005), yet a similarly pronounced size distinction between the two developmental forms is not commonly observed for environmental chlamydiae. While EBs of members of the Chlamydiaceae, Parachlamydiaceae, and Waddliaceae are usually coccoid, the infective stages of some environmental chlamydiae have been described as being rod-shaped, head-and-tail or star-shaped (Kahane, et al., 2001, Kahane, et al., 2002, Kostanjšek, et al., 2004, Michel, et al., 2005, Thomas, et al., 2006, Corsaro, et al., 2007, Karlsen, et al., 2008, Corsaro, et al., 2009). For Parachlamydiaceae, a third developmental stage, the crescent body, has been proposed (Greub & Raoult, 2002), which was, however, recently identified as artifact of conventional transmission electron microscopy (Pilhofer, et al., 2013). It remains to be determined whether other reported cell shapes of chlamydial EBs represent native states (Rusconi, et al., 2013).

A distinctive feature of EBs observed for all investigated chlamydial species is their highly condensed nucleoid, which is readily detectable in transmission electron micrographs (Matsumoto, 1988, Moulder, 1991, Siegl & Horn, 2012). Chromatin condensation is caused by bacterial histone-like proteins, such as the histone H1-like protein Hc1 that is present in high abundance in Chlamydia spp. EBs (approx. 20 000 copies/EB) (Hackstadt, et al., 1991, Barry, et al., 1992, Perara, et al., 1992, Christiansen, et al., 1993) and which is also encoded in the genomes of environmental chlamydiae (Horn, et al., 2004, Greub, et al., 2009, Bertelli, et al., 2010, Collingro, et al., 2011). DNA compaction in the infective stage has been proposed to confer a complete transcriptional shut-down (Barry, et al., 1993, Pedersen, et al., 1996), which is consistent with the reduced RNA to DNA ratio observed in EBs compared to RBs (Tamura, et al., 1967, Haider, et al., 2010) and with the opinion of chlamydial EBs as being a spore-like, metabolically dormant stage, whose primary function is to promote extracellular survival until they can gain access to a suitable intracellular habitat for replication (Moulder, 1991, Stephens, et al., 1998, Hatch, 1999, Iliffe-Lee & McClarty, 1999).

Another characteristic of Chlamydia spp. EBs that contributed to their consideration as spore-like is their relatively high resistance to harsh environmental conditions such as osmotic or physical stress (Hackstadt, et al., 1985). This has been attributed to the unique cell wall of Chlamydia in which structural stability is conferred by disulfide cross-linking of outer membrane proteins, including the major outer membrane protein (MOMP) and two cysteine-rich proteins OmcA and OmcB (Newhall & Jones, 1983, Bavoil, et al., 1984, Hatch, et al., 1984, Hackstadt, et al., 1985, Hatch, et al., 1986). In C. trachomatis and C. psittaci, the extent of disulfide cross-linking differs dramatically between EBs and RBs (Newhall & Jones, 1983, Bavoil, et al., 1984, Hatch, et al., 1984, Hackstadt, et al., 1985, Hatch, et al., 1986). In EBs, the cell wall complex is extensively cross-linked into a macromolecular structure that appears essentially as a spherical sacculus after detergent extraction (Caldwell, et al., 1981) thus explaining their increased stability. In contrast, in RBs, the proteins of the outer membrane complex are in a reduced state, thus the membrane is more fluid and porin activity of MOMP is increased (Bavoil, et al., 1984). MOMP, the cysteine-rich proteins, and several other proteins believed to compose the outer membrane of Chlamydiaceae are variably present in the environmental chlamydiae (Horn, et al., 2004, Greub, et al., 2009, Bertelli, et al., 2010, Collingro, et al., 2011), suggesting significant differences in the outer membrane structure among the Chlamydiae. However, although the amoeba symbiont P. amoebophila does not encode a homologue of the porin MOMP (Horn, et al., 2004), these bacteria were recently shown to possess a non-related family of porins, which together with conserved chlamydial cysteine-rich proteins could be detected in sarkosyl-insoluble outer membrane complexes of EBs, indicating that protochlamydial EBs have a similar rigid cross-linked outer membrane structure than Chlamydiaceae EBs (Heinz, et al., 2009, Heinz, et al., 2010, Aistleitner, et al., 2013).

The chlamydial inclusion

All Chlamydiae replicate within the confines of the inclusion. The inclusion of Chlamydia spp. is distinct from the pathogen-containing vacuoles inhabited by all other known vacuolar intracellular pathogens (Moulder, 1985, Heinzen, et al., 1996, Taraska, et al., 1996, van Ooij, et al., 1997, Hackstadt, 1998). It is initially derived from the host plasma membrane during the entry process (Scidmore, et al., 2003). Once bacteria have been internalized, the nascent inclusion does not detectably interact with cellular endocytic pathways (Heinzen, et al., 1996, Scidmore, et al., 1996, Taraska, et al., 1996, van Ooij, et al., 1997, Scidmore, et al., 2003). The inclusion is non-acidified (Heinzen, et al., 1996, Taraska, et al., 1996, Hackstadt, et al., 1997, van Ooij, et al., 1997, Fields & Hackstadt, 2002) and remains dissociated from the endosomal/lysosomal pathways for the duration of its existence, but is believed to intercept Golgi-derived vesicles and multivesicular bodies from which the chlamydiae obtain sphingomyelin and cholesterol (Hackstadt, et al., 1995, Hackstadt, et al., 1996, Rockey, et al., 1996, Wolf & Hackstadt, 2001, Carabeo, et al., 2003). The bacteria also actively recruit lipid droplets and host lipid transfer proteins to satisfy their needs for host lipids (Hackstadt, et al., 1996, Hatch & McClarty, 1998, Hatch & McClarty, 1998, Carabeo, et al., 2003, Su, et al., 2004, Beatty, 2008, Cocchiaro, et al., 2008, Heuer, et al., 2009, Derre, et al., 2011, Elwell, et al., 2011, Cox, et al., 2012, Elwell & Engel, 2012).

Conditions encountered by the bacteria within the inclusion are largely unknown, although it has been shown for C. trachomatis that pH and sodium, potassium, and calcium content of its vacuole appear to mimic that of the host cell cytosol (Grieshaber, et al., 2002). The inclusion membrane thus appears to be permeable to ions, although it is impermeable to microinjected fluorophores greater than 520 Da (Heinzen & Hackstadt, 1997). A potential permeability to metabolites smaller than 520 Da, such as certain sugars, sugar-phosphates, amino acids, and nucleotides has, however, not been further evaluated. It is thus not known yet, whether these nutrients can enter the chlamydial inclusion passively, or whether they have to be actively transported across the inclusion membrane by either host cell-derived or bacterial transporter proteins. To date, only a single host-derived transporter for metabolite uptake into the inclusion, the mammalian sodium multivitamin transporter (SMVT), has been identified (Fisher, et al, 2012). Nutrient acquisition could potentially also be facilitated by a group of bacterial proteins that are inserted in the inclusion membrane shortly after invasion of a host cell by Chlamydia spp. (Shaw, et al., 2000). These so-called inclusions membrane proteins (Incs) constitute a diverse family of proteins (Rockey, et al., 1995, Scidmore-Carlson, et al., 1999, Bannantine, et al., 2000, Rockey, et al., 2002), which are not at all well-conserved between species. They show a low degree of similarity or are absent all together from some species (Dehoux, et al., 2011, Lutter, et al., 2012), with few predicted Incs being conserved throughout all known Chlamydiae (Collingro, et al., 2011). While functions or interactions of a small number of specific Incs have been described, none of these interactions are common to all Chlamydia spp. and thus appear to be nonessential (Hackstadt, et al., 1999, Suchland, et al., 2000, Scidmore & Hackstadt, 2001, Rzomp, et al., 2006, Verbeke, et al., 2006, Cortes, et al., 2007, Derre, et al., 2011). Alternatively, it has recently been proposed that chlamydial Inc proteins may play a role in inclusion membrane structure and biogenesis (Mital, et al., 2013).

The parasitophorous vacuole inhabited by environmental chlamydiae has not been extensively characterized. These bacteria encode putative Incs that contain the characteristic bi-lobed hydrophobic domain (Heinz, et al., 2010, Collingro, et al., 2011). Yet, while localization to the inclusion membrane could recently be demonstrated for four Inc proteins of P. amoebophila (Heinz, et al., 2010), subcellular localization of other candidates, as well as their function in host cell interaction remains to be explored. Likewise, little is known about intracellular trafficking of inclusions of environmental chlamydiae. In monocyte-derived macrophages the replicative vacuoles of W. chondrophila were shown to acquire the early endosomal marker EEA1, while co-localization with late endosomal and lysosomal markers could not be observed (Croxatto & Greub, 2010). These findings suggest that evasion of the endocytic pathway is a more general feature linked to the interaction of chlamydiae with their host cells, although exceptions may exist (Greub, et al., 2005).

The study of chlamydial metabolism – a historical perspective

The study of chlamydial physiology is complicated by the obligate intracellular lifestyle of chlamydiae and the consequent difficulty in distinguishing host and parasite metabolic processes. Therefore, this branch of research has progressed with the development of techniques that facilitate separation of pathogen and host - enabling the analysis of host-free purified bacteria - as well as with the understanding of chlamydial developmental biology. In this section, the study of chlamydial physiology and metabolism is reviewed in a historical perspective according to prominent developments in the field. This approach illustrates the importance of biochemical information from early studies and how this can serve as a benchmark for refined reconstructions of metabolic pathways based upon more recent genomic data and experimental evidence.

The pre-genomic era

The earliest cultivation of chlamydiae, other than via animal passage, involved cultivation in embryonated eggs (Burnet & Rountree, 1935, Rake, et al., 1940, Rake & Jones, 1942) and facilitated the first investigations of chlamydial physiology.

Due to methodological constrains, initial studies on the ability of Chlamydia spp. to metabolize amino acids or intermediates of central metabolic pathways, including glycolysis and the tricarboxylic acid (TCA) cycle, failed to reveal an ability of the organisms to oxidize these substrates (Moulder & Weiss, 1951, Perrin, 1952, Allen & Bovarnick, 1957, Allen & Bovarnick, 1962). This apparent lack of systems to support de novo energy metabolism supported the view that Chlamydia spp. might scavenge ATP and other high-energy compounds from the host cell to enable intracellular bacterial replication a concept presented by Moulder as the “energy parasite” hypothesis (Moulder, 1962).

For many years, the relatively small size, filterable nature, obligate intracellular replication, combined with the initial failure to detect a capacity for host-independent generation of metabolic energy, led to chlamydiae being considered viruses or somehow intermediate between viruses and bacteria. However, in 1966 Moulder convincingly argued that chlamydiae exhibit several properties that are unusual for viruses and that undoubtedly demonstrate their bacterial nature (Moulder, 1966). These features include the presence of both DNA and RNA, the presence of ribosomes, retention of morphological integrity during intracellular replication, reproduction by binary fission, and a Gram-negative type cell wall. In addition, pioneering studies by Weiss and colleagues, had clearly demonstrated that Chlamydia spp. possess metabolic capacities that by far exceeded those that were known for viruses.

By studying bacteria that were grown and subsequently purified from embryonated chicken eggs, Weiss and colleagues could demonstrate that C. trachomatis and C. psittaci were capable of utilizing D-glucose, as inferred from the observed release of radioactive CO₂ from ¹⁴C-labeled D-glucose (Ormsbee & Weiss, 1963, Weiss, et al., 1964). However, this substrate did not stimulate increased oxygen consumption and while production of labeled CO₂ was readily observed in the presence of D-glucose labeled at either carbon atom 1, or carbon atoms 3 and 4, or randomly, liberation of carbon 6 could not be observed (Ormsbee & Weiss, 1963, Weiss, et al., 1964).

Thus, Weiss and colleagues concluded that aerobic respiration and TCA cycle activity were absent in these bacteria. Activity of the pentose phosphate pathway, combined with additional decarboxylation reactions, was suggested to explain their findings. The subsequent detection of enzyme activities related to the pentose phosphate pathway in lysates of purified C. psittaci supported this view (Moulder, et al., 1965). Further studies exploring TCA cycle activity confirmed that this major catabolic and energy generating pathway is not fully operational in host-free Chlamydia spp., although TCA-cycle related enzyme activities were shown to be partially present. Thus, while C. psittaci produced labeled CO₂ from pyruvate labeled at position 1, consistent with its conversion to acetyl-CoA by pyruvate dehydrogenase, no other carbon atoms were liberated from pyruvate (Weiss, 1967), making a further progression through the TCA cycle unlikely. Moreover, while L-glutamate could be converted to oxoglutarate (α-ketoglutarate) by a transaminase reaction, labeled CO₂ could again only be released from position 1 (Weiss, 1967), probably by activity of the TCA cycle enzyme oxoglutarate dehydrogenase.

Despite the apparent lack of a complete TCA cycle or aerobic respiration, activity of the glycolytic pathway could be demonstrated in host-free Chlamydia spp., based on the observation that isotopically-labeled pyruvate could be formed from D-glucose even when this substrate was labeled only at position 1, a carbon atom that would have been released as CO₂ earlier if the pyruvate would have been formed solely via the alternative pentose phosphate pathway (Weiss, et al., 1964). Interestingly, although glycolytic activity should enable the bacteria to produce a certain amount of ATP by substrate-level phosphorylation, utilization of D-glucose by Chlamydia spp. appeared to be absolutely dependent on external supply of ATP (Weiss, 1965). This paradox could be solved by subsequent studies that demonstrated that externally added ATP was required to fuel the conversion of D-glucose to D-glucose-6-phosphate by a host-derived co-purified hexokinase activity (Vender & Moulder, 1967). Indeed liberation of labeled CO₂ from radio-labeled D-glucose-6-phosphate by host-free Chlamydia spp. was also detected in the absence of externally provided ATP, which indicated that host-independent glucose catabolism in Chlamydia spp. starts with the phosphorylated compound and that Chlamydia spp. lack hexokinase (Weiss, 1965, Weiss & Wilson, 1969).

Additional studies conducted by Weiss and colleagues further defined media, cofactor, and atmospheric requirements for metabolic activity in host-free Chlamydia spp. (Weiss, 1965, Weiss, 1967, Weiss, et al., 1968) and demonstrated incorporation of ¹⁴C-labeled carbon from amino acids and carbohydrates into bacterial biomass (Weiss, et al., 1964, Weiss & Wilson, 1969). While most incorporated radioactivity was observed in the methanol-chloroform soluble fraction, indicating incorporation into lipids, with the limited technological means available at the time the authors did not observe evidence for host-free protein synthesis (Weiss & Wilson, 1969).

The discovery of developmental stage-specific metabolic features

The propagation of chlamydiae in continuous cell culture, which was first introduced in the 1960s (Gordon, et al., 1960, Pollard, et al., 1960, Bernkopf, et al., 1962, Gordon & Quan, 1965, Gordon, et al., 1969) as an alternative to the more cumbersome cultivation in egg yolk sacs and primary/explant cultures, greatly facilitated isolation of chlamydiae for use in physiological studies. The introduction of isopycnic (density gradient) centrifugation as a means to obtain highly pure preparations of EBs and RBs moreover represented a key technological achievement that promoted characterization of the physiological features of distinct developmental forms (Tamura & Higashi, 1963, Allen, 1967, Tamura & Manire, 1967, Tamura, et al., 1967, Tamura, et al., 1971). In particular the ability of purified EBs and RBs to synthesize macromolecules - such as proteins, DNA, and RNA - in host-free media, as well as their potential to import and hydrolyze ATP was analyzed in some detail. While it had been shown relatively early that C. psittaci utilizes the host cell nucleoside triphosphate pool for bacterial RNA synthesis, suggesting an inability to de novo synthesize nucleotides and occurrence of net import of host-derived nucleotides (Hatch, 1975), in 1982 Hatch et al. directly demonstrated that density gradient purified RBs of C. psittaci also transported ATP and ADP by an ATP-ADP exchange mechanism (Hatch, et al., 1982). In addition, the bacteria were shown to hydrolyze ATP by an oligomycin-sensitive ATPase and it was concluded that ATP hydrolysis, in an opposite manner to ATP synthesis during oxidative phosphorylation, was coupled to the generation of a membrane potential required to drive transport processes, such as import of lysine (Hatch, et al., 1982). However, ATP uptake was not detected in EBs (Hatch, et al., 1982) although Sarov and Becker reported ribonucleic acid biosynthesis in density gradient purified EBs of C. trachomatis (Sarov & Becker, 1971). Yet while Hatch and colleagues convincingly demonstrated protein, RNA, and DNA synthesis in Chlamydia RBs, they did not detect activity in EBs (Hatch, et al., 1985, Hatch, 1988, Plaunt & Hatch, 1988, Crenshaw, et al., 1990). Taken together, these findings contributed to the perception of the chlamydial EB as a spore-like metabolically inactive stage, the basis of which was suggested to be related to the highly cross-linked rigid outer membrane that has been proposed to represent a permeability barrier (Bavoil, et al., 1984). Indeed, a disulfide bond reduction of the EB outer membrane complex by addition of dithiothreitol (DTT) has been proposed to initiate a partial differentiation of EBs into RBs and has been shown to strongly enhance detectable metabolic activity (Hackstadt, et al., 1985). In addition, it was reported that ATPase activity could also be detected in EBs of C. trachomatis, yet only after treatment with reducing agents such as β-mercaptoethanol or glutathione (Peeling, et al., 1989). Consistent with their structural fragility, synthesis of macromolecules could be detected only for short periods of time in host-free RBs (Hatch, et al., 1985). RBs, moreover, displayed a strict requirement for molecular ATP to support host-free protein synthesis (Hatch, et al., 1985), which is consistent with the proposed function of ATP hydrolysis in fueling membrane potential-driven transport processes such as the import of amino acids (Hatch, et al., 1982).

In summary, pre-genomic investigations of the metabolic features of purified Chlamydiaceae contributed to the recognition of chlamydiae as being of bacterial nature and indicated major differences in the metabolic potential of the developmental forms, with EBs being described as metabolically inert. The validity of the energy parasite hypothesis remained elusive, as these early studies suggested a capacity for energy-generating activities, while at the same time indicating a requirement for external ATP for RB activity.

Chlamydia genomics reveal unexpected metabolic capacity

The sequence of the C. trachomatis serovar D genome published in 1998 (Stephens, et al., 1998) provided the first comprehensive insights into the metabolic capacity of chlamydiae and enabled more targeted functional characterizations of metabolic proteins via heterologous expression in Escherichia coli, which together helped to resolve some of the earlier controversies regarding this pathogen. Most notably, consistent with earlier observations by Weiss and colleagues (discussed above), metabolic pathway reconstruction questioned the dogma that chlamydiae were strict energy parasites (McClarty, 1999). In the years following the release of the first chlamydial genome, great efforts were undertaken to investigate genomic differences between chlamydial species and strains that display different host ranges and cause a diverse spectrum of diseases. Indeed, genome sequences of all members of the Chlamydiaceae, except from C. suis, are available today (Stephens, et al., 1998, Kalman, et al., 1999, McClarty, 1999, Read, et al., 2000, Read, et al., 2003, Carlson, et al., 2005, Thomson, et al., 2005, Azuma, et al., 2006, Thomson, et al., 2008, Mojica, et al., 2011, Voigt, et al., 2012). Comparative genomics revealed a surprisingly high conservation in gene content and gene order and consequently also highly similar metabolic capacities among representatives of this family. This section summarizes some of the most important findings from the genomic analysis of the Chlamydiaceae, with particular emphasis on insights into their central carbon and energy metabolism.

Similar to other obligate intracellular bacteria (Zientz, et al., 2004, Eisenreich, et al., 2010), Chlamydiaceae have relatively small genomes (about 1 Mb) that are highly reduced in gene content (Stephens, et al., 1998, Kalman, et al., 1999, McClarty, 1999, Read, et al., 2000, Read, et al., 2003, Carlson, et al., 2005, Thomson, et al., 2005, Azuma, et al., 2006, Thomson, et al., 2008, Mojica, et al., 2011, Voigt, et al., 2012). Nevertheless, these bacteria encode the enzymatic capacity for metabolism of glucose-6-phosphate to pyruvate via glycolysis, which allows them to generate ATP via substrate-level phosphorylation by the enzymes phosphoglycerate kinase and pyruvate kinase. Moreover, heterologous expression and characterization of respective C. trachomatis enzymes in E. coli confirmed that they are functional (Iliffe-Lee & McClarty, 1999). Consistent with the earlier experimental observations (Weiss, 1965, Vender & Moulder, 1967, Weiss & Wilson, 1969), genomic data confirmed that Chlamydiaceae lack a hexokinase gene, as well as substrate-specific components of the phosphotransferase system (PTS) (Stephens, et al., 1998, Kalman, et al., 1999, McClarty, 1999, Read, et al., 2000, Read, et al., 2003, Carlson, et al., 2005, Thomson, et al., 2005, Azuma, et al., 2006, Thomson, et al., 2008, Mojica, et al., 2011, Voigt, et al., 2012). The bacteria thus depend on the import of phosphorylated sugar (D-glucose-6-phosphate) from the host cell cytosol, which is likely accomplished by a sugar-phosphate/inorganic-phosphate antiporter (UhpC) (McClarty, 1999), the C. pneumoniae homolog of which has been functionally characterized in E. coli (Schwöppe, et al., 2002). The direct utilization of phosphorylated glucose, combined with the use of a pyrophosphate-dependent phosphofructokinase, increases the yield of ATP chlamydiae may gain during glycolysis and was thus proposed to represent an energetic advantage (McClarty, 1999).

The TCA cycle is incomplete in all Chlamydiaceae due to the absence of three enzymes, citrate synthase (GltA), aconitase (Acn) and isocitrate dehydrogenase (Icd) (Stephens, et al., 1998, Kalman, et al., 1999, McClarty, 1999, Read, et al., 2000, Read, et al., 2003, Carlson, et al., 2005, Thomson, et al., 2005, Azuma, et al., 2006, Thomson, et al., 2008, Mojica, et al., 2011, Voigt, et al., 2012). Thus genome sequences are fully consistent with the earlier observation that acetyl-CoA derived from pyruvate, the end product of glycolysis, cannot enter the TCA cycle (Weiss, 1967), nor can acetyl-CoA derived from fatty acid degradation. Instead, a constant metabolite exchange with the host is required for TCA cycle operation (Stephens, et al., 1998). This has been proposed to involve import of oxoglutarate and export of malate, both of which may be accomplished by a dicarboxylate translocator (SodTi) (McClarty, 1999). Alternatively, Chlamydiaceae may import L-glutamate, which may be converted to oxoglutarate either by a glutamate dehydrogenase or by a transaminase reaction (McClarty, 1999), the latter of which has already been experimentally demonstrated in purified C. psittaci (Weiss, 1967). Interestingly, there is evidence for ongoing reduction in TCA cycle genes in the Chlamydiaceae. All C. trachomatis strains sequenced so far carry a frameshift mutation in the gene encoding succinate dehydrogenase subunit C (SdhC), leading to a truncated gene product (Thomson, et al., 2008). Since C. trachomatis SdhC thus lacks functionally important residues required for the binding of hemes and the interaction with menaquinol (Cecchini, 2003, Lancaster, 2013), the enzyme may not be functional. This may have extensive consequences for the metabolism in C. trachomatis, since succinate dehydrogenase, also known as complex II of the respiratory chain (see below), is uniquely positioned to act as a regulator of the TCA cycle, as well as of oxidative phosphorylation (Cecchini, 2003). In C. trachomatis strains L2/434/Bu and L2/UCH-1/proctitis two frameshift mutations in the gene encoding fumarate hydratase (FumC) could reduce the function of the TCA cycle in these strains even more, to the mere synthesis of succinate and succinyl-CoA from oxoglutarate (Thomson, et al., 2008).

Genome analysis also revealed that Chlamydiaceae possess the capacity to produce ATP by oxidative phosphorylation using their complete, though minimal, respiratory chain, which consists of a Na⁺-translocating NADH dehydrogenase (Nqr, complex I), succinate dehydrogenase (SdhA-C, complex II), cytochrome bd oxidase (CydAB, complex IV), and a V-type ATPase (complex V) (Stephens, et al., 1998, Kalman, et al., 1999, McClarty, 1999, Read, et al., 2000, Read, et al., 2003, Carlson, et al., 2005, Thomson, et al., 2005, Azuma, et al., 2006, Thomson, et al., 2008, Mojica, et al., 2011, Voigt, et al., 2012). The presence of a cytochrome bd-like oxidase, usually exhibiting high oxygen affinity, was proposed to indicate that chlamydiae may face reduced oxygen conditions during intracellular growth and thus display a microaerophilic lifestyle (Juul, et al., 2007, Roth, et al., 2010, Dietz, et al., 2012, Omsland, et al., 2012). The respiratory chain may enable regeneration of the oxidized forms of NADH and FADH₂ (reduced during glycolysis and the TCA cycle), and may additionally allow the bacteria to build up an electrochemical membrane potential across their plasma membrane, which may be utilized to drive transport processes (McClarty, 1999). However, it is currently not clear whether Chlamydiaceae can indeed use their V-type ATPase also for ATP generation, by exploiting the membrane potential, or whether the ATPase is instead predominately used for the opposite purpose, aiding in the generation of a membrane potential by ATP-dependent extrusion of ions, as has been proposed previously based on experimental findings (Hatch, et al., 1982).

The NADH dehydrogenase of Chlamydiaceae is most closely related to Na⁺-translocating NADH dehyrogenases (Verkhovsky & Bogachev, 2010, Juarez & Barquera, 2012), indicating that in these bacteria a sodium motif force rather than a proton motif force, may be the driving membrane potential. While the respiratory chain of C. trachomatis may, however, also translocate protons via the cytochrome bd oxidase (McClarty, 1999), V-type ATPases are indeed known to use predominately sodium as coupling ion (Dzioba, et al., 2003, von Ballmoos & Dimroth, 2007). Because the level of sodium in the host cell and in the C. trachomatis inclusion is low and sodium-dependent ATP synthesis would require a relatively strong sodium gradient, it appears unlikely that C. trachomatis uses its ATPase for synthesis of ATP (Hase, et al., 2001, Grieshaber, et al., 2002, Dibrov, et al., 2004).

Apart from metabolic pathways involved in energy generation, genomic analysis revealed that Chlamydiaceae possess a complete pentose phosphate pathway, required for regeneration of NADPH and synthesis of pentose-phosphates, a complete pathway for gluconeogenesis, and the capacity to synthesize and degrade the energy storage compound glycogen (Stephens, et al., 1998, Kalman, et al., 1999, McClarty, 1999, Read, et al., 2000, Read, et al., 2003, Carlson, et al., 2005, Thomson, et al., 2005, Azuma, et al., 2006, Thomson, et al., 2008, Mojica, et al., 2011, Voigt, et al., 2012). Moreover, all Chlamydiaceae are able to synthesize long chain and branched chain fatty acids and some lipids, such as phosphatidylethanolamine and phosphatidylserine.

Chlamydiaceae are auxotrophic for most amino acids, cofactors, and for purine and pyrimidine nucleotides (Stephens, et al., 1998, Kalman, et al., 1999, McClarty, 1999, Read, et al., 2000, Read, et al., 2003, Carlson, et al., 2005, Thomson, et al., 2005, Azuma, et al., 2006, Thomson, et al., 2008, Mojica, et al., 2011, Voigt, et al., 2012); and consequently rely on the import of host-derived compounds. Some chlamydial transport proteins have already been characterized functionally in E. coli. These include C. trachomatis BioY and C. trachomatis CTL843, which enable import of host-derived biotin and S-adenosylmethionine, respectively (Binet, et al., 2011, Fisher, et al., 2012). Moreover, functional characterization of C. trachomatis nucleotide transporters Npt1Ct and Npt2Ct established the biochemical basis for chlamydial ATP scavenging and for the net import of nucleotides and nicotinamide adenine dinucleotide (NAD) (Tjaden, et al., 1999, Fisher, et al., 2013). Genomic analyses of Chlamydiaceae also provided evidence for species- and even strain-specific differences in their metabolic capacities (Stephens, et al., 1998, Kalman, et al., 1999, McClarty, 1999, Read, et al., 2000, Read, et al., 2003, Carlson, et al., 2005, Thomson, et al., 2005, Azuma, et al., 2006, Thomson, et al., 2008, Mojica, et al., 2011, Voigt, et al., 2012). Beside variable capabilities to synthesize or salvage nucleotides, this mainly affects their ability to synthesize the amino acid L-tryptophan from metabolic intermediates (Kalman, et al., 1999, Fehlner-Gardiner, et al., 2002, McClarty, 2006, Voigt, et al., 2012), which was proposed to account for serovar-specific differences in tissue tropism of C. trachomatis (Fehlner-Gardiner, et al., 2002, Nelson, et al., 2005). Enhanced tryptophan biosynthetic capabilities observed in genital strains of C. trachomatis may enable continued bacterial growth under conditions in which chlamydial species with reduced capabilities (e.g. ocular strains of C. trachomatis) may be inhibited or may enter a non-replicating “persistent” state in response to interferon-gamma-induced tryptophan depletion (Byrne, et al., 1986, Beatty, et al., 1993, Beatty, et al., 1994, Beatty, et al., 1994, Carlson, et al., 2005, McClarty, 2006).

Taken together, the availability of complete genome sequences of Chlamydiaceae greatly advanced our knowledge of chlamydial metabolism and its potential interconnections with host cell metabolism. It clearly demonstrated that pathogenic chlamydiae, beside their ability to withdraw ATP from the host cell cytosol, indeed have the capacity for independent ATP synthesis. Yet the actual contribution of bacteria-driven ATP generation to the overall energy demands of chlamydiae at different time points during development remains to be elucidated.

The extended metabolic repertoire of environmental chlamydiae

Genome sequences of four environmental chlamydiae, P. amoebophila (Horn, et al., 2004), Pa. acanthamoebae (Greub, et al., 2009, Collingro, et al., 2011), S. negevensis (Collingro, et al., 2011), and W. chondrophila (Bertelli, et al., 2010, Collingro, et al., 2011) have been published. Comparative analyses indicated that the genomes of these species are significantly divergent, displaying little or no synteny to genomes of the Chlamydiaceae nor to each other (Collingro, et al., 2011). Most notably, they are about 2–3-fold larger in size than genomes of the Chlamydiaceae and encode a comparably larger number of proteins (Figure 1) (Horn, et al., 2004, Greub, et al., 2009, Bertelli, et al., 2010, Collingro, et al., 2011).

While genes encoding essential house-keeping proteins and proteins involved in fundamental aspects of chlamydial biology, such as the regulation of the chlamydial developmental cycle, are expected to be part of the chlamydial core gene set (i.e. genes shared by all sequenced members of the Chlamydiae), those genes found only in a subset of species may represent more specific adaptations (Collingro, et al., 2011). Among the genes that were found only in environmental chlamydiae, many are of unknown function (Collingro, et al., 2011). However, a number of these genes were predicted to encode proteins involved in metabolic processes, indicating that these environmental chlamydiae, while still being fully dependent on a eukaryotic host cell, have retained a higher metabolic potential (Horn, et al., 2004, Horn, et al., 2006, Bertelli, et al., 2010, Collingro, et al., 2011). A comparison of selected metabolic capacities between pathogenic chlamydiae and their more recently discovered environmental relatives is illustrated in Figure 3 and is briefly summarized below.

Figure 3. The extended metabolic repertoire of environmental chlamydiae.

Selected metabolic features of environmental chlamydiae compared to the Chlamydiaceae as inferred from genome sequences are shown. Pathway reconstructions are based on data from (Horn, et al., 2004, Bertelli, et al., 2010, Collingro, et al., 2011).

In contrast to the Chlamydiaceae, all environmental chlamydiae encode a glucokinase, which enables these bacteria to phosphorylate and thus activate D-glucose on their own (Horn, et al., 2004, Horn, et al., 2006, Bertelli, et al., 2010, Collingro, et al., 2011). Environmental chlamydiae, moreover, encode a complete TCA cycle and are thus able to oxidize acetyl-CoA, derived from carbohydrate or fatty acid degradation, independently from their host to CO₂, accompanied by regeneration of the energy-rich molecules NADH, FADH₂ and GTP (Horn, et al., 2004, Horn, et al., 2006, Bertelli, et al., 2010, Collingro, et al., 2011). W. chondrophila and Pa. acanthamoebae, in contrast to all other chlamydiae, also encode a AMP-forming acetyl-CoA synthetase (Acs), which should enable them to activate acetate by conversion to acetyl-CoA (Kumari, et al., 1995, Starai & Escalante-Semerena, 2004). In addition, these two species also encode a glyoxylate bypass (Bertelli, et al., 2010, Collingro, et al., 2011), enabling them to build up glucose from fatty acids and thus to potentially grow on fatty acids as main carbon source. The importance of this capability for their lifestyle is currently unknown.

The electron transport chain of environmental chlamydiae is branched and thus more versatile compared to the minimal repertoire found in Chlamydiaceae (Horn, et al., 2004, Horn, et al., 2006, Bertelli, et al., 2010, Collingro, et al., 2011). Though differences can also be observed among environmental chlamydiae, their genomes in general encode additional components, such as a H⁺ translocating NADH dehydrogenase, additional terminal oxidases, and cytochrome c oxidases, which altogether were suggested to enable these bacteria to build up a proton gradient more efficiently and potentially allow them to respond to changing conditions such as fluctuating oxygen tensions by differential expression of specific respiratory chain components (Horn, et al., 2004, Bertelli, et al., 2010, Collingro, et al., 2011). W. chondrophila and the Parachlamydiaceae, moreover, encode an additional F-type ATPase (Horn, et al., 2004, Horn, et al., 2006, Bertelli, et al., 2010, Collingro, et al., 2011), which, in contrast to the V-type ATPase that can be found in all chlamydiae, is predicted to use H⁺ as coupling ion and may thus use a proton gradient for ATP generation (Dibrov, et al., 2004). For F-type ATPases the electrochemical gradient of the coupling ion needed for ATP synthesis is, furthermore, half that of V-type ATPases, making them more efficient in energy generation (Murata, et al., 2001). Thus, although the environmental chlamydiae also encode an ATP/ADP translocase which has been functionally characterized in some species (Trentmann, et al., 2007, Knab, et al., 2011), they appear to have increased capacities for host-independent production of ATP via oxidative phosphorylation compared to the Chlamydiaceae.

Although environmental chlamydiae also appear to have in general enhanced biosynthetic capabilities compared to the Chlamydiaceae, pathways associated with the synthesis of amino acids, cofactors, vitamins and nucleotides are still truncated, consistent with their obligate intracellular lifestyle (Horn, et al., 2004, Horn, et al., 2006, Bertelli, et al., 2010, Collingro, et al., 2011). In contrast to the Chlamydiaceae, all environmental chlamydiae should be capable of synthesizing the amino acids L-proline, L-alanine, and L-glutamate. Yet the trp operon, which includes genes involved in L-tryptophan synthesis and which is partially present in some Chlamydiaceae (Stephens, et al., 1998, Read, et al., 2003, Azuma, et al., 2006, Mojica, et al., 2011) and complete in S. negevensis (Collingro, et al., 2011), is absent from other environmental chlamydiae (Horn, et al., 2004, Bertelli, et al., 2010, Collingro, et al., 2011). In addition, whereas Chlamydiaceae were proposed to depend on the import of both NAD and NADP, in C. trachomatis accomplished by the ATP/ADP translocase Npt1Ct (Fisher, et al., 2013) and a yet to be identified transporter, respectively, P. amoebophila was shown to import NAD by a more specialized nucleotide transporter (NTT4) and was predicted to convert NAD to NADP on its own using its NAD kinase (Haferkamp, et al., 2004). All other environmental chlamydiae appear to have an even higher biosynthetic potential, including the capacity to salvage nicotinamide to produce NAD (Bertelli, et al., 2010). S. negevensis may even be able to synthesize NAD de novo from the amino acids L-asparagine (Collingro, et al., 2011, Knab, et al., 2011). While all chlamydiae encode nucleotide transporters and enzymes required for net uptake of nucleotides and synthesis of deoxynucleotides for DNA synthesis, respectively (Stephens, et al., 1998, Kalman, et al., 1999, Horn, et al., 2004, Greub, et al., 2009, Bertelli, et al., 2010, Collingro, et al., 2011) the genome of W. chondrophila suggests a potential to synthesize pyrimidines de novo from L-glutamine and to synthesize AMP and GMP from adenine (Bertelli, et al., 2010).

Additional major differences in the biosynthetic potential among the Chlamydiae affect the synthesis of menaquinone, isoprenoids, and glycogen. While all environmental chlamydiae are able to synthesize menaquinone by the chorismate pathway (Collingro, et al., 2011), Chlamydiaceae were formerly believed to be unable to produce menaquinone and to rely on the import of ubiquinone from the host cell instead (Stephens, et al., 1998, McClarty, 1999). However, a new metabolic pathway (futalosine pathway) has recently been described (Hiratsuka, et al., 2008), which enables the synthesis of menaquinone from chorismate via futalosine and which appears to be present in all Chlamydiaceae genomes. For isoprenoid synthesis, chlamydiae use the more recently discovered methylerythritol phosphate (MEP) pathway (non-mevalonate pathway) (Grieshaber, et al., 2004, Hunter, 2007). An exception appears to be W. chondrophila that uses the mevalonate pathway instead (Bertelli, et al., 2010). The genetic prerequisites for the formation and degradation of glycogen are present in all chlamydiae except from W. chondrophila (Stephens, et al., 1998, Kalman, et al., 1999, Horn, et al., 2004, Bertelli, et al., 2010, Collingro, et al., 2011), which seems to lack glucose-1-phosphate adenylyltransferase (GlgC) or may encode a distantly related protein that fulfills this function.

Taken together, the complete genome sequences of divergent representatives of the environmental chlamydiae indicate that these species exhibit a - compared to the pathogenic chlamydiae - significantly increased, though at the same time highly variable, metabolic potential. This finding was proposed to reflect their potentially less homeostatic growth niche (Horn, et al., 2004, Collingro, et al., 2011). An increased versatility in responding to fluctuating environmental conditions may be required by chlamydiae naturally or occasionally infecting protozoan hosts.

Future studies using mass spectroscopy- and NMR-based metabolomics methods and novel techniques for imaging of bacterial and host cell metabolism such as two-photon laser scanning microscopy, Raman microspectroscopy, or secondary ion mass spectrometry will help to better understand the metabolism of both Chlamydiaceae and environmental chlamydiae and of their host cells during infection and under different environmental conditions (Ojcius, et al., 1998, Wagner, 2009, Haider, et al., 2010, Szaszak, et al., 2011, Müller, et al., 2013, Sixt, et al., 2013). When combined with novel genetic approaches, which are currently being developed (Tam, et al., 1994, Binet & Maurelli, 2009, Kari, et al., 2011, Wang, et al., 2011, Mishra, et al., 2012, Nguyen & Valdivia, 2012, Wang, et al., 2013), such studies permit further testing of genome-based hypotheses.

EBs as metabolically inert stage – time for reconsideration

Although genomic investigations greatly advanced our knowledge of the metabolic capacity of chlamydiae and unraveled significant species- and even strain-specific differences, we are still far from understanding how chlamydiae interact with their intracellular and extracellular chemical environment and which nutrients they require to sustain growth and infectivity. This is partly due to the fact that genomic analyses revealed a large number of genes of unknown function (Stephens, et al., 1998, Kalman, et al., 1999, Horn, et al., 2004, Bertelli, et al., 2010, Collingro, et al., 2011), some of which may be involved in unknown metabolic pathways. In any case, genomics does not provide information about how bacterial physiology changes during the course of the chlamydial developmental cycle. As described above, the majority of early studies on chlamydial structure and metabolism suggested that EBs are metabolically dormant. Consequently, all metabolic capabilities inferred from genome sequences were attributed to the intracellular replicative stage. However, a critical review of earlier investigations, as well as more recent experimental findings challenge this concept and suggest that the chlamydial EB can be metabolically active, yet displays a unique physiology and unique nutrient requirements.

Weiss and colleagues have analyzed host-free metabolic activities of mixtures of EBs and RBs of Chlamydiaceae and reported that no significant loss of activity was observed, when purified bacteria were stored for two days prior to activity assessment (Weiss, 1965, Weiss, 1967). These findings are inconsistent with the fragility of chlamydial RBs described some years later, when methods for the separation and individual characterization of developmental stages became available (Tamura & Higashi, 1963, Matsumoto, 1988, Omsland, et al., 2012). Thus, although it was not specifically addressed by Weiss and coworkers some of the activities they observed might have in fact been carried out by the more stable EB stage. In addition, it has been shown that nucleoid decondensation during the redifferentiation of EBs to RBs, shortly after their uptake into a host cell depends on chlamydial de novo transcription and translation to synthesize an intermediate of the non-mevalonate isoprenoid synthesis pathway, which causes release of the histone-like proteins from bacterial DNA (Grieshaber, et al., 2004). Therefore, a certain level of metabolic capacity must be maintained in EBs for initiation of activity once in the appropriate environment.

The availability of chlamydial genome sequences facilitated proteomic analyses and thus enabled a better characterization of the composition and metabolic repertoire of chlamydial EBs. The analysis of the C. pneumoniae EB proteome, representing the first comprehensive proteome study of a chlamydial species in the post-genomic era, revealed the presence of a surprisingly large number of proteins involved in protein synthesis and energy generation in the infective stage, including enzymes involved in glycolysis and the TCA cycle (Vandahl, et al., 2001). The authors thus suggested that an active energy metabolism may take place in EBs and that this stage may possess functional transcriptional and translational machineries. Based on these findings they questioned metabolic dormancy in EBs and thus proposed to term them a “non-dividing” stage instead (Vandahl, et al., 2001). A similar high representation of proteins involved in energy metabolism, transcription, and translation was also observed in EBs of C. trachomatis and P. amoebophila and was proposed to demonstrate either that the infective stage is primed for germination once in the appropriate environment or that the EB stage itself has the capacity for metabolic activity (Shaw, et al., 2002, Skipp, et al., 2005, Sixt, et al., 2011). Most notably, a recent quantitative proteomic comparison of C. trachomatis EBs and RBs confirmed the presence of proteins involved in translation, glucose utilization, and intermediary metabolism in EBs and showed that while RBs were enriched in proteins associated with translation, nutrient transport (including import of ATP), and the electron transport chain, proteins involved in central carbon metabolism and glucose catabolism were more abundant in EBs (Saka, et al., 2011).

Haider and colleagues recently used confocal Raman microspectroscopy to differentiate between P. amoebophila EBs and RBs, which enabled the analysis of developmental form-specific activity at a single-cell level (Haider, et al., 2010). Raman microspectroscopy represents a nondestructive vibrational spectroscopic method that provides information about the molecular composition of a sample (Wagner, 2009). In microbiology, Raman spectra of single bacterial cells provide fingerprints that can be used for the identification of and discrimination between bacterial species (Jarvis, et al., 2006). In density gradient purified fractions of P. amoebophila, Haider et al. observed two distinct classes of spectra that could be assigned to either EBs or RBs based on an accompanying electron microscopy-based quantification of developmental stages in the respective fractions (Haider, et al., 2010). Interestingly, uptake of externally added ¹³C-labeled phenylalanine into single bacterial cells, observed as a pronounced shift in the phenylalanine peak of their Raman spectra, was observed not only during intracellular growth of P. amoebophila in amoebae, but also during incubation of bacteria in a host-free defined medium (Haider, et al., 2010).

Host-free activity of P. amoebophila was detectable for up to 21 days, predominately in bacteria exhibiting typical EB Raman spectra (Haider, et al., 2010). Addition of carbonylcyanide m-chlorophenylhydrazone (CCCP), an ionophore that dissipates the membrane potential, blocked uptake of phenylalanine, consistent with the notion that this amino acid is imported via a proton/sodium neutral amino acid symporter (Haider, et al., 2010). In the same study, it was also shown that host-free EBs were able to reenergize their membrane after removal of CCCP, demonstrating active electron transport processes in the infective stage (Haider, et al., 2010). Consistent with this observation, it was reported more recently that host-free P. amoebophila also reduce the tetrazolium dye CTC (Sixt, et al., 2013), indicating respiratory activity (Rodriguez, et al., 1992, Creach, et al., 2003); this activity could be attributed to the EB stage, because the proportion of active bacteria in density gradient fractions enriched for this developmental stage clearly exceeded the expected proportion of co-purified RBs and transition stages (Sixt, et al., 2013).

Haider and co-workers furthermore observed incorporation of ¹⁴C-labeled phenylalanine into proteins by host-free P. amoebophila and C. trachomatis even 24 h after release from their host cells, and suggested de novo protein synthesis in EBs (Haider, et al., 2010). Finally, in a recent comparative characterization of metabolic features of C. trachomatis EBs and RBs, de novo protein synthesis and ATP generation in the infective stage of this chlamydial species was demonstrated (Omsland, et al., 2012).

Thus, while it is clear that chlamydial EBs represent a non-dividing stage, which per se indicates a reduced activity compared to the replicative form, there is accumulating evidence for the occurrence of some degree of metabolic activity in this infective stage. The concept of chlamydial EBs as metabolically inert or spore-like may therefore be in need of refinement. The discrepancy between recent findings and the failure to detect EB metabolic activity in earlier studies (Hatch, et al., 1982, Hatch, et al., 1985, Hatch, 1988, Plaunt & Hatch, 1988, Crenshaw, et al., 1990) may be due to improvements in media composition and differences in purification and pre-treatment of EBs before assessment of activity.

Refining the energy parasite hypothesis

Recently, highly enriched C. trachomatis EB and RB populations were used to assess the ability of each cell form to respond to specific nutritional and physiochemical conditions (Omsland, et al., 2012). As presaged by the pioneering studies of Weiss (Weiss, 1965) and Hatch (Hatch, et al., 1985) showing utilization of D-glucose 6-phosphate and ATP by Chlamydia spp., respectively, these recent experiments revealed developmental stage-specific energy source requirements. Axenic protein synthesis by EBs and RBs was restricted to specific physiochemical environments and nutritional conditions (Omsland, et al., 2012). Notably, metabolic activity of both cell types was highly favored by ionic conditions characteristic of the inclusion lumen. In addition to D-glucose 6-phosphate and ATP, the medium designed in this study for axenic activity of C. trachomatis also contained nucleotides, amino acids, and the reducing agent DTT, and maintenance of activity was improved under microaerophilic conditions (Omsland, et al., 2012). These measures facilitated a previously unseen extent of host free-activity of Chlamydiae. The comparison of EBs and RBs revealed that EBs were dependent upon D-glucose-6-phosphate for ATP synthesis, transcription, and translation whereas RBs were entirely dependent upon ATP as an energy source (Omsland, et al., 2012), the latter of which is consistent with earlier findings by Hatch (Hatch, et al., 1985).

Developmental stage-specific nutrient requirements have not been investigated for a member of the environmental chlamydiae, but it could be shown that host-free maintenance of metabolic activity of P. amoebophila EBs was dependent on the presence of D-glucose (Sixt, et al., 2013). This was inferred from the observation that replacement of this substrate with the non-metabolizable stereoisomer L-glucose caused a more rapid decline in their potential to reduce CTC, even when other potential energy sources such as amino acids were present in the medium (Sixt, et al., 2013). Keeping in mind that P. amoebophila, in contrast to C. trachomatis, is capable of phosphorylating D-glucose by its own (Horn, et al., 2004, Horn, et al., 2006), EBs of both species thus seem to have a similar requirement for D-glucose-(6-phosphate).

While it can currently not be excluded that D-glucose could be replaced by other substrates, such as intermediates of glycolysis or the TCA cycle, D-glucose uptake by host-free P. amoebophila EBs could also be demonstrated experimentally on the single-cell level by the use of a fluorescently labeled D-glucose analog (Sixt, et al., 2013). In the same study the utilization of ¹³C-labeled D-glucose by purified P. amoebophila EBs was further analyzed by a variety of mass spectrometry-based methods, including isotope-ratio mass spectrometry (IRMS), ion cyclotron Fourier transform mass spectrometry (ICR/FT-MS), and ultraperformance liquid chromatography mass spectrometry (UPLC-MS). These analyses revealed several differences compared to extracellular activities reported previously for host-free Chlamydiaceae and thus provided experimental evidence for the increased metabolic potential of environmental chlamydiae as inferred from genomic analysis. More precisely, imported D-glucose was subsequently phosphorylated and primarily metabolized via the pentose phosphate pathway (Sixt, et al., 2013). In addition, complete TCA cycle activity could be observed, as inferred from the release of carbon 6 from ¹³C-labeled D-glucose as CO₂ and the formation of a labeled metabolite whose mass-to-charge ratio corresponded to the TCA cycle intermediate citrate or isocitrate. Thus, EBs of environmental chlamydiae exploit their increased metabolic potential compared to the Chlamydiaceae, at least under host-free conditions.

In summary, the apparent dependence of Chlamydia RBs on ATP under axenic conditions is reminiscent of the energy parasite hypothesis argued by Moulder in 1962 (Moulder, 1962), but in the EB stage there is also evidence for metabolic activity and energy generation. While these findings may indicate that the two cell forms rely on distinct mechanisms of energy metabolism (i.e. ATP scavenging in RBs versus ATP synthesis by substrate level-phosphorylation and/or oxidative phosphorylation in EBs), further studies will be required to dissect EB and RB energy metabolism in conditions faced during infection.

Biological significance of EB activity

The metabolic capacity is clearly there, but whether EBs utilize this capacity primarily when they are still within the inclusion or also under extracellular conditions faced after they have been released from their host cell, requires further investigation. Recent studies provide some clues in this regard.

The exchange of D-glucose with the non-metabolizable stereoisomer L-glucose, in case of P. amoebophila, or the exchange of D-glucose-6-phosphate by D-glucose, in case of C. trachomatis, leads to a more rapid decrease in infectivity during host-free incubation of chlamydiae (Sixt, et al., 2013). Additional evidence for a biological role of host-free EB activity was provided by the experiments with the ionophore CCCP described above. Although the inhibition of L-phenylalanine uptake by CCCP was in general reversible, no recovery could be detected after pro-longed incubation of host-free bacteria in the presence of this drug (Haider, et al., 2010), indicating that maintenance of a membrane potential may be essential for chlamydial survival, at least in case of P. amoebophila. Together, these findings indicate that metabolic activity in EBs can contribute to their survival in the extracellular environment.

As the host-free environment and nutrient availabilities encountered by chlamydiae after release from their host cells may thus have a significant effect on the outcome of infection, further studies will be required to assess whether EBs have access to sufficient amounts of glucose, glucose-6-phosphate, or other substrates to regenerate ATP pools in their natural extracellular environment (e.g., mucosal surfaces or biofilms), or whether conditions that were used experimentally to prove axenic activity were successful because they mimicked intracellular conditions. Indeed, the robust requirement of C. trachomatis EBs for ionic conditions that reflect cytosolic and intra-inclusion sodium and potassium concentrations (Omsland, et al., 2012) suggests that axenic culture media replicates the intracellular environment and the host-free activity observed experimentally for this species may be dependent upon this simulation.

The importance of host-free metabolic EB activity may differ between Chlamydiaceae and the environmental chlamydiae. The increased metabolic potential of species infecting amoebae (Horn, et al., 2004, Greub, et al., 2009, Bertelli, et al., 2010, Collingro, et al., 2011) together with the observation of prolonged extracellular maintenance of infectivity (Sixt, et al., 2013) may suggest that these bacteria may be better adapted to host-cell free survival than the pathogenic Chlamydiaceae. Amoebae naturally occur in a great variety of different habitats, including soil, freshwater and marine habitats, as well as man-made environments (Rodriguez-Zaragoza, 1994). It is thus conceivable that bacteria released from these protists may be exposed to highly variable and fluctuating conditions and may have to persist in the host-free environment for prolonged periods of time as compared to Chlamydiaceae EBs discharged from infected cells and released into animal tissue, where new host cells are readily available.

Concluding remarks

A fundamental tenet of bacteriology is that by closely enough replicating the natural environment of a bacterium, axenic culture is possible. While this has been achieved for some spore-forming species and other bacteria that - like chlamydiae - exhibit structurally and metabolically distinct developmental forms (such as Coxiella, Legionella, Caulobacter), it may be particularly challenging to decode and mimic the role of the chlamydial inclusion membrane in creating an appropriate environment. The semi-porous nature of the inclusion membrane (Heinzen & Hackstadt, 1997, Grieshaber, et al., 2002) could conceivably regulate concentrations of required nutrients to maintain a homeostasis difficult to predict and likely difficult to maintain in a static culture system. In addition, the porosity of the inclusion could permit diffusion of metabolic byproducts to prevent a toxic accumulation. The recent identification of membrane contact sites (Derre, et al., 2011) or pathogen synapses (Dumoux, et al., 2012) that appear to be discrete sites for contact of the inclusion membrane with the endoplasmic reticulum suggests that the influx of lipid into the inclusion may be regulated. Even if it became possible to design axenic conditions supporting host-free cultivation of chlamydial RBs, studying the chlamydial developmental cycle axenically would require that, in addition to cell division, differentiation also take place in the host-free environment. The signals leading to developmental events are, however, not known. The contact-dependent model for the development of chlamydiae postulates that chlamydial differentiation is regulated by physical contact with the inclusion membrane (Hackstadt, et al., 1997), possibly type three secretion system-mediated (Wilson, et al., 2006). If this is the case, axenic emulation of the chlamydial developmental cycle may be an even greater challenge.

A system for axenic cultivation of chlamydiae would be expected to significantly accelerate current efforts in the development of means for their genetic manipulation. On the other hand, recently developed genetic tools (Tam, et al., 1994, Binet & Maurelli, 2009, Kari, et al., 2011, Wang, et al., 2011, Mishra, et al., 2012, Nguyen & Valdivia, 2012, Wang, et al., 2013), could be used to modify chlamydial metabolic activities. Certain deficiencies may be complemented by introducing metabolic genes, which could potentially minimize the number of supplements needed for any medium. Even growth of genetically modified chlamydiae in cell-free media would be a tremendous advance that would contribute to basic research, diagnostics, and vaccine development.