Who’s in control? Regulation of metabolism and pathogenesis in space and time

Abstract

Bacterial pathogens need to sense and respond to their environments during infection to align cell metabolism and virulence factor production to survive and battle host defenses. Complex regulatory networks including ligand-binding transcription factors, two-component systems, RNA-binding proteins, and small non-coding regulatory RNAs adjust gene expression programs in response to changes in metabolic fluxes, environmental cues, and nutrient availability. Recent studies underlined that these different layers of regulation occur along varying spatial and temporal scales, hading to changes in cell behavior and heterogeneity among the bacterial community. This brief review will highlight current research emphasizing that cell metabolism and pathogenesis are inextricably intertwined in both Gram-positive and Gram-negative bacteria.

Graphical Abstract

Introduction

All microorganisms must sense the environment and align their physiology to the current conditions. Pathogenic bacteria must contend with hostile environments of the host, which impose nutritional and environmental stresses and deploy immune cells to attack and eventually destroy invading microbes. Thus, pathogens must reconfigure both metabolism and the production of virulence factors simultaneously. Such a response is essential to optimize the production of needed proteins and enzymes as well as biological polymers including DNA, RNA, cell wall components and extracellular polymeric substances (EPS). Individual bacterial cells align and coordinate production of these molecules in part using regulatory proteins that monitor the availability of key nutrients. These proteins can act individually or in combination with multiple global and operon-specific regulators in a hierarchical fashion to adjust gene expression. For instance, the prokaryotic transcription factor Lrp (responsive to leucine) and the Firmicute-specific transcription factor CodY (responsive to branched-chain amino acids (BCAAs) and GTP) serve to link the physiological state of the cell to pathogenicity [1,2]. Bacteria utilize transcriptional regulators responsive to ligands [3] and two-component systems (TCSs) to sense and respond to environmental cues including secreted defense molecules in a transcriptional manner [4]. In addition, RNA binding proteins (RBPs) and small, non-coding regulatory RNAs (sRNAs) mediate cellular responses to physiological and environmental stresses post-transcriptionally [5]. However, bacteria rarely act as individual cells. Rather, they communicate using a chemical language that allows them to sense one another, essentially behaving socially with complex intra- and interspecies interactions of a cooperative or an antagonistic nature [6,7]. Previously, researchers used reductionist approaches to study the molecular basis of individual components. While we now have a greater appreciation for their role in more complex systems, our knowledge about physiology and gene regulation during infection is still limited. We posit that integrating this information on varying time and spatial scales is key to understanding pathogen behavior and can inform our design of new therapies as our antibiotics become increasingly ineffective. Just as The Doctor in the revered UK television program explores all of space and time, in this short review, we will explore recent developments on spatio-temporal regulation of gene expression, emphasizing the intertwining of metabolism with production of virulence factors. For the sake of brevity, we provide highlights from Salmonella enterica (hereafter referred to as Salmonella) and Escherichia coli, model Gram-negative intestinal pathogens that intercept cues from the microbiota or the host in the gastrointestinal tract [8], and Staphylococcus aureus, a model Gram-positive pathogen that can exploit nearly every niche of the body. However, regulating metabolism and virulence in space and time applies to all bacteria of medical importance and we encourage the reader to reflect on the concepts described here as to how they might apply to other microorganisms.

Nutrient availability and host signals influence transcriptional responses

Nutrient gradients and gene expression

The ability to acquire and process available nutrients efficiently is essential for pathogen survival and multiplication in the host as well as eventual dissemination. While the production of secreted virulence factors can act to extract needed nutrients from host cells or tissues, key regulators can adjust the activities of catabolic pathways to exploit carbon source availability in various environments, particularly when host metabolism is disturbed. Classic carbon catabolite repression (CCR) mechanisms, in part, mediate this tissue-specific response and are reviewed elsewhere [9]. Notably, the availability of a rapidly metabolizable carbon source like glucose affects the activity of the global regulator CcpA in Firmicutes like S. aureus (reviewed in [10]). Bischoff et al. demonstrated CcpA-dependent upregulation of α-hemolysin activity in tissues with high glucose levels, such as the liver [11]. Richardson and colleagues showed S. aureus tissue invasion requires a transition to an environment relatively high in glucose but low in oxygen (or one where respiration is impaired), and that transition is dependent on a functional glycolytic oathway [12]. Diabetics have higher serum glucose levels, which affect the metabolism of invading pathogens. In two mouse models of diabetes where glucose levels are elevated, the mice were found to be more susceptible to S. aureus in the context of skin infection. Host immunity was affected, but the investigators also found that genes associated with glycolysis, translation, stress response, virulence, and the Clp proteases were upregulated; these likely promote adaptation to the environment and facilitate improved growth [13]. This conceivably contributes to delays in chronic wound resolution in these patients [14]. In wounds that do resolve, the abscesses that arise deprive S. aureus of glucose and oxygen, forcing the bacteria again to adapt. On the topic of wounds, multiple species of bacteria tend to be present, and it’s worth noting that not all bacteria adopt the CCR strategy; some actually use a so-called reverse CCR (rCCR) that inverts the order of preferred substrates and reduces competition for the same carbon source [15,16]. Here, we clearly see regulation of metabolism and virulence is altered on spatial scales ranging from within a wound to organ systems to exploit various niches.

Recent studies have suggested that similar metabolic signals regulating the activity of CcpA also regulate activity of the ribose-responsive regulator RpiRc [17]. Although the RpiRc ligand remains unknown, there is a strong connection between RpiRc activity and the Pentose Phosphate Pathway (PPP), as S. aureus strains lacking the PPP enzyme transketolase have altered transcriotional activity of RpiRc [18]. RpiRc represses RNAIII transcription, the effector of S. aureus Agr quorum sensing system (discussed in further detail later) (Fig. 1) [17]. Increased activity of RpiRc is correlated with exponential phase growth where nutrients are abundant and biosynthetic intermediates are plentiful. RNAIII activity is low and the abundance of negatively regulated RNAIII targets, such as the secreted immunoglobin G binding protein A (SpA), are highest [17]. During the postexponential and stationary phases of growth, available nutrients and metabolic intermediates decrease along with RpiRc activity. As a result of altered metabolic fluxes, RpiRc adjusts the expression of metabolic pathways as well as the leukocidins in an Agr/RNAIII dependent manner, leading to an overall enhancement of S. aureus virulence in a bloodstream infection model [17,19].

Figure 1. Integration of metabolism and virulence in S. aureus.

Cells sense nitric oxide (NO*) and hypoxia imposed by the host innate immunity through several TCS, including SrrAB and SaeRS, most likely due to increases in levels of reduced menaquinones (rMK) in the electron transport chain (NDH: nicotinamide adenine dinucleotide dehydrogenase, Cyt: cytochrome aa3/bd). Activation of SrrA and SaeR modulate the expression of genes involved in NO* detoxification. Additionally, SrrA activates the transcription of two sRNAs, RsaD and RsaE. RsaE reduces amino acid catabolism and represses most of the TCA enzyme genes, reducing NADH production and consequently would decrease the rMK pool. CodY-regulated RsaD represses the expression of acetolactate synthase AIsS to fine-tune cell death regulation. Low TCA cycle activity and altered metabolic fluxes alleviate RpiRc repression of RNAIII transcription; adjusting expression of metabolic and virulence factor genes. Once a certain cell population density is reached, the Agr system is induced through the accumulation of auto-inducing peptides (AIP) (quorum sensing) that activates AgrA, the transcriptional factor of agr operon. Disruption of Mn²⁺ homeostasis also contributes to superoxide (O₂*-) detoxification by activating RsaC sRNA expression. RsaC represses Mn²⁺-dependent superoxide dismutase SodA expression to favor the cambialistic SodM enzyme.

Two-Component Systems

Bacterial pathogens use two component systems (TCSs) or more complex phosphorelays to sense and respond to their environments [20–22]. TCSs consist of a sensor kinase and a response regulator; upon activation by a signal, the sensor kinase undergoes autophosphorylation and transfers the phosphoryl group to the response regulator, ultimately leading to differential expression of downstream genes. The ability to sense a variety of environments and stresses is linked to both niche and metabolic versatility. S. aureus encodes 16 TCSs, which contribute to its ability to grow as a commensal or as an opportunistic pathogen causing devastating infections ranging from skin and soft tissue infections to bacteremia and sepsis [23,24]. In a pioneering study, researchers knocked out the 15 non-essential TCSs, which rendered S. aureus incapable of sensing stresses or the surrounding environment, paving the way for biologists to ask deeper questions on signal integration, cross-talk, and niche accommodation [25].

An emerging area ot study is the role of TCSs in sensing respiratory flux, and how this contributes to fitness in vitro and during infection. Nitric oxide (NO*) is a membranepermeable radical and a major innate immune defense against invading pathogens. While bacteria like Escherichia coli and Bacillus subtilis use general and oxidative stress responses to mitigate the effects of NO*, S. aureus SrrAB TCS coordinates the response to nitrosative stress in vivo and in vitro [26–28]. Exposure to NO* prompts SrrAB activity to detoxify the NO* and maximize cytochrome production [27]. Reminiscent of the ArcAB system in E. coli, SrrAB activity is thought to increase when reduced menaquinones accumulate in the respiratory chain as a result of damage to cofactors in cytochromes (Fig. 1) [29]. The SaeRS TCS is thought to operate via a similar mechanism under low oxygen conditions. SaeRS contributes to the regulation of over 20 virulence genes, and along with SrrAB regulate fermentative biofilm formation [29,30]. The coordination of the SaeRS and SrrAB response to S. aureus respiratory status exemplifies the layers of regulation that integrate physiological cues to fine-tune virulence gene expression in response to host signals.

Another class of important innate defense molecule is reactive oxygen species (ROS). During the course of infection, Salmonella colonizes the small intestine, invades the epithelial later, and is engulfed by host cells where it encounters toxic levels of ROS [31]. Salmonella’s ArcAB TCS is critical for survival not only in response to changing oxygen levels, but also fo’lowing exposure to ROS, in particular H₂O₂ in the phagosome [32]. ArcAB represses the expression of several porins, and lack of porin proteins on Salmonella’s cell surface leads to increased ROS resistance [33,34]. When inside neutrophils, specific porins are up-regulated by ArcAB; however, ArcAB down-regulates many of those same porins inside macrophages (Fig. 2) [32]. Further, activation of ArcAB in neutrophils, which display harsher environments that macrophages, results in upregulation of the type III secretion system (T3SS) of the Salmonella Pathogenicity Island 1 (SPI-1), which is necessary for intracellular invasion and maintenance of the salmonella-containing vacuole (SCV) [32]. This ArcAB regulation inside neutrophils suggests that Salmonella is spatially aware of its surroundings and uses ArcAB to respond to various cell types strategically.

Figure 2. Integrating physiology and pathogenesis in S. enterica while inside the phagosomes of macrophages.

PhoQ is activated by changes in pH, the presence of certain antimicrobial peptides and/or Mg²⁺ ions - signals indicating that the bacterium has been phagocytized by the host. PhoP in turn directly represses hilA transcription, which encodes the main transcriptional activator of SPI-1 genes. SPI-1 encodes a T3SS. PhoP also represses hilA transcription indirectly by transcription of an Hfq-dependent sRNA, PinT, which posttranscriptionally downregulates hilA mRNA. Additionally, by repressing CRP, PinT prevents hilA transcriptional activation by several factors (RtsA, HilC, HilD). HilA repression prevents SPI-1 genes expression, only required for the invasion stage of the intestinal epithelial cell. CsrA is predicted to enhance SPI-1 repression by binding to multiple virulence genes. Transcriptional activation of Zn transporter ZnuABC and PhoP-dependent activation of Mg²⁺ transporter MgtA reinforce resistance to nitro-oxidative stress by maintaining homeostasis and bolstering glycolytic fluxes. AC, adenylate cylase

Phenotypic heterogeneity and spatial regulation of gene expression

Host signals or resources are often diffusible or otherwise spatially distributed and can affect gene expression and physiology in individual cells, leading to distribution of labor and cellular “differentiation”. A recent study observed competition for iron (Fe) oetween the host and pathogen and how this competition affects S. aureus infection outcomes [35]. Fe availability varies within host tissue; this is reflected in abscesses that indicated Fe starvation not only in different organs, but also within distinct regions of the same infection focus. Interestingly, this study also identified “microdomains” of other metals within abscesses, including calcium, phosphorous, and manganese [35]. Our laboratory made a similar observation regarding phenotypic heterogeneity and discovered that the expression of virulence gene nuc (secreted nuclease) is high in staphylococci at the core of abscesses but is muted in cells at the periphery of the abscess [36]. The mechanistic underpinnings of this phenotype are currently under investigation, but given the proximity of peripheral staphylococci to neutrophils, a diffusible host signal is likely to be part of the story.

Expression of nuc is also heterogeneous in biofilms; this mechanism has been clarified recently [37]. Early in the formation of S. aureus biofilms, a subpopulation of cells detaches from its substrate in an event termed “exodus” through Sae-dependent stochastic expression of nuc, leading to degradation of extracellular DNA (eDNA), a major component of the matrix that maintains biofilm structural integrity. This allows for the formation of microcolonies that are metabolically heterogeneous [37,38]. While the stochastic expression of nuc is Sae-dependent, Agr-dependent genes have been shown to also exhibit heterogeneity in expression [39]. However, they exhibit a different pattern of expression during biofilm development. The Agr system, described further below, is activated relatively later in growth when levels of autoinducing peptide (AIP) accumulate. Consistent with this regulation, the Agr-dependent genes psmβ, plc, and sek are not expressed stochastically during early biofilm development. Rather, expression is confined to late stages of biofilm development when towers form [37]. Clearly, the activities of multiple regulators are integrateo during biofilm development to form these heterogenous subpopulations.

The strategy of spatially regulating gene expression in tissues is not limited to Gram-positive pathogens like S. aureus. Salmonella differentiates into at least six subpopulations based on temporal and spatial fluctuations in ROS and reactive nitrogen species with distinct phenotypes and cell fates [40]. More recently, heterogeneity and spatial regulation of pathogen gene expression in host tissue was reported by Davis and colleagues during murine infection with the Gram-negative bacterium Yersinia pseudotuberculosis. Y. pseudotuberculosis cells on the periphery of lesions upregulate the expression of the gene coding for the NO*-detoxifying enzyme Hmp, protecting cells in the center of the lesion from diffusing NO* and sparing them the need to upregulate hmp. Another subpopulation of cells on the periphery of lesions in contact with host cells express the T3SS to inhibit phagocytosis by neutrophils [41]. NO* exposure damages iron-sulfur clusters, leading to inhibition of respiration stated above. Thus, it makes physiological sense that expression of the iron-sulfur cluster repair protein gene ytfE is increased on the periphery of the Yersinia microcolonies, promoting the survival of these bacteria on the “front lines” of the host pathogen battle. [42]. These examples demonstrate that individual bacterial cells in a population exhibit social behaviors; there are others, and the concept is reviewed in more detail elsewhere [7]. Stochasticity, spatial location, and varying time scales contribute to this apparent distribution of labor during infection to alter metabolism and virulence gene expression in individual cells of a species and it will be interesting to see how this pans out for mixed infections in the future.

Nutritional Immunity

Essential metals are sequestered by the host using a strategy known as nutritional immunity [23]. For instance, local restriction of metals (iron (Fe), manganese (Mn), zinc (Zn)) by neutrophil-produced chelator calprotectin (CP) or divalent-cation transporter NRAMP1 in acidic phagosomes halts bacterial growth and provide a natural resistance to infections by the host [31,43]. Thus, a deep interest in nutritional immunity has emerged and we hignlight a few examples here.

Zn is essential for bacterial DNA replication/repair as well as serving as a co-factor for glycolytic enzymes. After engulfment by macrophages, Salmonella immediately senses the harsh environment and combats host nutritional immunity by adjusting the cellular physiology [44,45]. Nitrosative stress induces a derepression of the high-affinity Zn transporter, znuABC, which is negatively regulated by the global regulator Zur. By activating the import of Zn (necessary for activity of fructose bisphosphate aldolase), ZnuABC fuels glycolytic flux under NO* stress (Fig. 2). Additionally, low pH and deprivation of divalent cations within acidifying phagosome vacuoles induces the activation of the Salmonella PhoPQ TCS to modulate outer membrane content for immune evasion and shut down the no longer needed T3SS [46,47] (Fig. 2). PhoPQ TCS also activates the transcription of magnesium transporter MgtA to maintain Mg²⁺ homeostasis, protecting cells against NO* stress possibly by promoting protein synthesis and RNS-induced damage [48]. Activation of the PhoPQ system illustrates how Salmonella can detect where and when to turn on virulence genes to resist innate immunity.

Mn homeostasis is critical for survival in the host by acting as a cofactor in radicaldetoxifying enzymes and glycolytic enzymes [49]. S. aureus circumvents the restriction of Mn by sensing decreased Mn availability through the ArIRS TCS. ArIRS does not sense the depletion of Mn itself; rather, the signal appears to be Mn-induced perturbations of glycolytic flux [50]. ArIRS can be considered a global regulator of Mn starvation and is a prime example of hierarchical regulation, acting either directly on transcription or by altering the activities of other staphylococcal regulators including Agr, MrgA, Rot, and LytSR [51,52]. As a result, carbon flux is shifted to pathways with lower Mn demand [51], reminiscent of the iron sparing response mediated by Fur and the RyhB small regulatory RNA that prioritizes Fe-dependent pathways during iron scarcity [53]. A second strategy to bypass nutritional immunity is to express enzymes that are metal-independent or cambialistic, which is the ability to bind multiple metals for enzymatic activity. For instance, S. aureus inhibits the synthesis of Mn-dependent superoxide dismutase SodA in favor of expressing the cambialistic SodM (Fig. 1), assuring resistance to oxidative stress in Mn-depleted conditions like those generated in the presence of CP [54]. Similarly, GpmA is the secondary, metal-independent phosphoglycerate mutase in S. aureus that specifically functions when S. aureus is starved for Mn [49]. GpmA is essential to establish invasive infection, providing resistance to CP-dependent Mn starvation and retaining glycolytic functions.

Posttranscriptional regulation enhances cell adaptation to the host environment

Recent studies have underscored the importance of posttranscriptional regulation of metabolism for cell fitness in diverse environments, including host niches. RNA binding proteins and sRNAs are critical for reguiating a multitude of bacterial stress responses, by adjusting the flow of metabolites through central metabolic pathways in both Gram-negative and Gram-positive bacteria, therefore being critical for adaptation to the host environment. As we also highlight, they are essential for regulating and timing functions for pathogenicity.

RNA-binding proteins (RBPs) regulate gene expression in response to nutritional status or general stress

The small RBP CsrA, is a core component of the carbon storage regulatory (Csr) system and exemplifies one of the widest phylogenetically distributed factors that coordinates complex physiological processes including carbon metabolism and virulence in response to nutritional conditions [55]. Acting as an RNA chaperone, global regulator CsrA binds to specific single-stranded GGA motifs in RNA molecules and typically results in translation repression and/or stability of downstream mRNA targets. Across multiple proteobacteria, CsrA tends to activate glycolysis and repress stationary phase genes or those genes upregulated during exposure to environmental stresses [55]. Recently, CsrA was found to repress translation of iron storage proteins (entCEAB, entD, feoABC) in exponential growth under minimal stress but not in the iron-limited environment of the large intestine and within host cells [56,57]. The accumulation of short-chain carboxylate compounds, abundant in the intestinal lumen, and the amino acid starvation alarmone (p)ppGpp, stimulates the production of two inhibitory sRNAs, CsrB and CsrC. These two major sRNAs prevent E. coli CsrA activity by mimicking CsrA-binding motifs [55], alleviating the repression of iron storage proteins under oxidative stress and ensuring that cellular iron is available for growth [56]. The same cues result in CsrBC-mediated sequestration of CsrA, preventing it from binding to several transcripts of operons in the SPI-1 T3SS and promoting host cell invasion (Fig. 2) [58]. A similar phenomenon is expected to occur for Salmonella during intracellular growth in acidic vacuoles inside the host. CsrBC-mediated sequestration of CsrA would alleviate repression of sopD2 translation, activating the disruption of endocytic trafficking in the host and promoting stability of the Salmonella containing vacuole [58,59].

The S. aureus genome also encodes small RBPs that modulate gene expression. Notably, CspA post-transcriptionally modulates a broad repertoire of mRNA targets encoding proteins involved in carbohydrate and ribonucleotide metabolism or in virulence [60]. For instance, by controlling the translation of the alternative sigma factor σ^B and the production of S. aureus golden pigment and antioxidant staphyloxanthin, CspA potentially enhances bacterial survival during infection [61]. Together, recent studies discovered new RBP functions that fine-tune complex bacterial decision-making processes.

Temporal control of genes by sRNA during infection or cell-cell communication

Bacterial sRNAs act as major posttranscriptional regulators of metabolic processes and virulence by regulating translation and/or mRNA degradation with an imperfect base-pairing to mRNA targets. The RNA chaperone Hfq promotes the base-pairing between these trans-acting sRNAs and mRNA targets in most Proteobacteria (see [62] for a detailed review) and can act as a spatial and temporal regulatory mechanism during infection. The first temporal stage of Salmonella infection requires bacterial invasion into the host intestinal epithelial cells using the T3SS (Fig. 2). HilA acts as the main transcriptional activator of SPI-1, itself activated by three AraC-like regulatory proteins, HilC, RtsA ana HilD, creating a complex feed-forward loop [63]. Once cells invaded phagosomes of the host cell, SPI-1 T3SS expression is rapidly downregulated to optimize fitness of the population as it is no longer needed after internalization or during systemic replication in macrophages [64]. Kim et al demonstrated that, in response environmental cues inside phagosomes, the PhoPQ TCS activates the expression of the Hfq-dependent PinT sRNA, which, as a post-invasion-activated sRNA, acts as a timer to silence SPI-1 by blocking hilA and rtsA translation (Fig. 2) [63]. The sRNA also represses the expression of several specific SPI-2 genes during early stages after invasion through direct repression of cyclic AMP receptor protein (CRP), the central regulator of CCR, likely adapting carbon metabolism flux inside macrophages simultaneously [65]. In summary, this example illustrates that a sRNA can shape the cellular switch from invasion to intracellular replication by controlling the expression of different sets of genes (SPI-1 and SPI-2) at the posttranscriptional level.

Multifunctional S. aureus RNAIII illustrates the best characterized regulatory RNA in Gram-positive pathogens. RNAIII expression is controlled by the Agr quorum-sensing system in a population density-dependent manner (reviewed in depth elsewhere [21,66]). In brief, autoinducing peptides bind to the AgrC sensor kinase and activate the AgrA response regulator with a necessary but unclear role for the metalloprotease MroQ (Fig. 1) [67,68]. As a result, RNAIII is transcribed. RNAIII carries a short open reading frame encoding cytolytic peptide δ-hemolysin Hla while its noncoding sequence acts as the effector of the Agr quorum-sensing system to post-transcriptionally regulate a set of genes including virulence factors [67,68]. Only few other sRNAs have been functionally characterized in S. aureus but the vast majority are affected by a large network of regulators (CodY, AgrA, SrrA, SaeS and SigB), thereby controlling the expression of virulence factors in response to a wide variety of environmental and metabolic signals [69–72]. For instance, RsaD was shown to be a novel effector of S. aureus SrrA when exposed to NO* and influences overflow metabolism when CodY activity is reduced (Fig. 1) [73,74]. Interestingly, Agr was recently shown to repress RsaD via RNA-Seq[67]. Together with SrrA-dependent RsaE sRNA, these sRNAs may integrate host-induced stresses and other signals, such a quorum sensing, to mount a metabolic response to physiological stresses encountered by S. aureus during infection Given their ability to influence metabolism and pathogenicity, elucidating the function of additional uncharacterized sRNAs in the ESKAPE pathogens is a high priority because of the potential for development of drugs that block their synthesis and activity Recent developments in computational and experimental approaches [75] will pave the way.

Conclusion

The dramatic surge of multidrug resistance to our curreni antibiotics emphasizes the need for new therapies. Pathogens use complex intertwined layers of regulation to adjust metabolism and virulence factoi production during infection; thus, targeting these circuits is one exciting alternative approach in addition to other anti-virulence strategies like quorum quenching [76] that would allow the host immune response to naturally clear the infection. Importantly, emerging views of behavior in vivo reveal metabolic and phenotypic heterogeneity across spatial and temporal scales. Candidate anti-virulence targets must be evaluated to ensure the invading pathogen will be susceptible to the strategy. Bright reporters have already demonstrated their utility [36,77]; emerging methods, including intravital microscopy, multimodal imaging, Dual-RNA-seq, 3D cell culture and others [78–81] will surely contribute to the redesign of our strategies for new therapeutic interventions.

Highlights.

• Regulators act as genetic “time circuits” to adjust metabolism and virulence

• Altered metabolic flux has profound effects on survival and pathogenic potential

• Host signals or nutrients can affect gene expression on various spatial scales

• Spatial and temporal cues result in metabolic and phenotypic heterogeneity