Evidence of defined temporal expression patterns that lead a gram-negative cell out of dormancy

Summary

Many bacterial species are capable of forming long-lived dormant cells. The best characterized are heat and desiccation resistant spores produced by many Gram-positive species. Less characterized are dormant cysts produced by several Gram-negative species that are somewhat tolerant to increased temperature and very resistant to desiccation. While there is progress in understanding regulatory circuits that control spore germination, there is scarce information on how Gram-negative organisms emerges from dormancy. In this study, we show that R. centenum cysts germinate by emerging a pair of motile vegetative cells from a thick cyst cell wall coat ~ 6 hrs post induction of germination. Time-lapse transcriptomic analysis reveals that there is a defined temporal pattern of gene expression changes during R. centenum cyst germination. The first observable changes are increases in expression of genes for protein synthesis, an increase in expression of genes involved in the generation of a membrane potential and the use of this potential for ATP synthesis via ATPase expression. These early events are followed by expression changes that affect the cell wall and membrane composition, followed by expression changes that promote chromosome replication. Midway through germination, expression changes occur that promote the flow of carbon through the TCA cycle to generate reducing power and parallel synthesis of electron transfer components involved in oxidative phosphorylation. Finally, late expression changes promote the synthesis of a photosystem as well as flagellar and chemotaxis components for motility.

Keywords:

Gene expression – DNA replication – Gene regulation – Cell walls – Flagella – Gram negative bacteria – Protein synthesis – Bacterial spores

Introduction

Many bacteria enter a period of non-replicative dormancy when faced with nutrient deprivation or harsh environmental conditions. The best studied are dormant long-lived spores by Gram-positive Firmicutes. Many other species are also capable of producing long-lived dormant persister cells that are resistant to antibiotics. Less studied are long-lived dormant cysts synthesized by several Gram-negative species such as Rhodospirillum centenum, Azospirillum sp., and Azotobacter sp. Cysts from these species are not as resistant to environmental stresses as are Gram-positive spores, but they do offer considerable resistance to desiccation and some resistance to heat [1, 2]. Information exists about regulatory events that control the development of Gram-positive spores [3, 4] with just a few studies undertaken on the development of Gram-negative cysts. However, several recent studies do provide at least a baseline understanding of regulatory proteins that control cyst development by R. centenum, a photosynthetic member of the Azospirillum clade [5–11].

Cues initiating Gram-positive spore germination have been well defined and encompass both small molecule nutrient germinants such as amino acids, organic acids, nucleosides and cholesterol, and non-nutrient germinants such as calcium dipicolinate, peptidoglycan fragments, dodecyl amine and even high pressure [12, 13]. Nutrient germinants are thought to bind to a cluster of germination receptors located in the cell membrane [12–15]. This interaction initiates several stages of germination leading to cell wall remodeling, restoration of membrane lipid composition, and core protein movement [12, 13]. Once germination becomes an irreversible process, then there is an outgrowth stage during which there is increased gene expression affecting transport, protein synthesis, and DNA replication [12–14].

In contrast to spore germination, an understanding of Gram-negative cyst germination remains largely unexplored. Several decades ago there were a few microscopic and physiological studies on the germination of cysts from the Gram-negative bacterium Azotobacter vinelandii. These microscopic studies observed that A. vinelandii cysts underwent division during the germination process just before rupture of a thick exine exopolysaccharide coat that encloses individual cysts [16, 17]. Additional physiological analyses also showed that A. vinelandii cysts exhibit rapid induction of protein synthesis, nucleic acid synthesis, and then respiration upon induction of germination [18]. These early events were followed by changes in metabolism, initiation of DNA synthesis, and nitrogen fixation just before emergence from the exine coat [18]. Regulatory circuits that control the induction of these ordered processes in germinating A. vinelandii cysts are not yet defined.

In this study, we have analyzed the germination of R. centenum cysts using a combination of time-lapse microscopy coupled with RNA-seq based transcriptomic analysis. Results from time-lapse microscopy show that the germination period is ~ 6 hrs in length terminating with cyst wall hydrolysis and the emergence of more than one motile vegetative cell. Transcriptomics performed at defined periods in the germination developmental pathway show that genes responsible for protein synthesis, energy production and cell defense are among the first to be expressed. This ramp-up in energy production was followed by periods of DNA replication, flagellar synthesis, and cell wall remodeling. The combination of time-course microscopy, coupled with transcriptomic data, provides the first reported molecular chronology of events that take place during the emergence of a Gram-negative cyst from dormancy.

Results and discussion

Microscopic analysis of cyst germination

Fig 1A shows a scanning electron micrograph of an R. centenum colony undergoing a transition from vegetative to cyst phases. The individual viroid shaped replicative cells are quite distinct from cysts that are large clusters of dormant cells encased in a thick exine layer of exopolysaccharides. Transmission electron micrographs of replicative and cyst cells also highlight substantial morphological differences between these two cell types (Fig 1B–1D). Liquid grown R. centenum vegetative cells are viroid in shape with a single sheathed polar flagellum (Fig 1B), while vegetative cells grown on an agar solidified media exhibit both a sheathed polar flagellum and numerous unsheathed lateral flagella (Fig 1C). In contrast, dormant cysts are round non-flagellated cells encased within a thick exine polysaccharide extracellular wall (Fig 1D). Cyst cells also are replete with large pools of highly refractive polyhydroxy butyric acid (PHB) that is thought to be used as an internal energy source for long-term cyst survival and emergence from dormancy [19].

The process of cyst germination was monitored by undertaking time-lapse live cell imaging of cysts placed onto nutrient rich CENS agar (Fig 2 and S1 Movie). During early phases of the germination process, the cyst cluster undergoes a slight swelling in size reminiscent of swelling of a Bacillus spore that occurs while they take up water from the environment [12, 20]. There is also a distinct movement of phase bright cells within the exine polysaccharide extracellular wall ~30 minutes of the initiation of germination. The first emergence of vegetative cells from the exine layer typically occurs 5.5 hrs post induction of germination with most cysts germinated by 6 hrs. Typically, pairs of cells emerge together from cysts indicating that cells divide before hydrolysis of the exine polysaccharide extracellular wall. Finally, after emergence there is an empty exine polysaccharide extracellular wall left behind that can be visualized as dark empty “husks” as the germinating vegetative cells continue to divide outside of the exine wall.

Overview of cyst germination transcriptomic analysis

We undertook RNA-seq based transcriptomic analysis at various time points during cyst germination to assess temporal changes in expression that occur when cells leave dormancy and enter a period of replication. We first developed a method to remove any remaining vegetative cells from cyst inducing growth medium and then measured the amount of cyst enrichment by flow cytometry. Cysts used in this study were typically >95% enriched (S1 Fig).

Using isolated cysts, a pilot experiment was initially undertaken to address how quickly changes in expression are observed post initiation of germination. After assessing these results, we undertook replicate studies (three biological replicates) where germination was induced by transferring isolated cyst into nutrient-rich CENS medium from which aliquots were removed for RNA extraction. RNA samples were taken every ½ hour for the first 3 hours, followed by every 1 hour until reaching 8 hrs. Microscopic analysis shows that by 8 hours, the cysts were fully germinated so this last time point also provided a reference of RNA levels observed from vegetative cells relative to the RNA levels observed from fully dormant cysts. RNA extracted from every time point (0.5, 1, 1.5, 2, 2.5, 3, 4, 5, 6, 7, and 8 hr) was compared to RNA extracted from dormant cysts (0th hour) with regions considered differentially expressed (DEG) if the log₂ value from this comparison was greater than or equal to ±1.5 with a p-value <0.05. Finally, RNA-seq data sets were independently validated by assaying expression changes of five genes at four time points (0, 0.5, 2.5 and 6 hr) using quantitative PCR (qRT-PCR). For each gene and time point, log₂fold change with respect to the 0^th hour from qRT-PCR was calculated and plotted against the same fold changed obtained from independent RNA-seq data sets (S2 Fig). The obtained R² value between qRT-PCR and RNA-seq data sets for these genes shows excellent correlation with an R² = 0.91.

During the time course of induction, we observed that 40.64% of the R. centenum genome (1,667 genes) underwent significant changes in expression relative to fully dormant cysts (S1 Table). Given this large number of gene expression changes, we analyzed the data by first binning into clusters of orthologous groups (COGS) where genes that control similar processes were individually analyzed (S3 Fig). In the sections below, we discuss individual COGS that best highlight molecular events that lead to the emergence of a fully functional vegetative cell. Finally, we compared transcriptomic data collected during germination in this study to a prior study from our laboratory that analyzed transcriptomic changes in expression that occur during cyst formation [21]. Interestingly, 317 genes exhibit DEG in both data sets with 194 of these showing opposing regulation defined as upregulation during cyst formation and downregulation during cyst germination or vice versa (S4 Fig). Comparison of gene expression changes during germination with that observed during cyst formation contributed to a more in-depth insight into both the germination and encystation processes in this Gram-negative species.

Early events initiating 0.5 to 1.5 hrs

Translation machinery

An important set of genes expressed earliest in germination include those that encoded proteins of the large and small ribosomal subunits (Fig 3 and S5A and S5B Fig). This group of genes uniformly showed significant increases in expression at the earliest 0.5 hr time point and continued to increase in expression throughout germination. Most initiation factors (IF-1 and IF-2), elongation factors (Tu, G and P) and release factors involved in translation also show significant upregulation early in germination with sustained increases in expression throughout (S5C Fig). The one exception is the elongation factor G (EF-G) coded by RC1_1873 that is inexplicably downregulated during germination with the caveat that the R. centenum genome also codes for two additional EF-G's (RC1_0708 and RC1_3537) that both exhibit early increases in expression. Note that early induction of protein synthesis is known to occur soon after the initiation of Bacillus spore germination [22].

In addition to synthesis of translation machinery, genes coding for proteins involved in protein folding, such as HSP40, and HSP90 and GroEL, are also upregulated during this early time point with some showing significant regulation across the time spectrum (early to middle–SurA^RC1_1563, CcmG^RC1_3021, middle–HSP20, HSP10 and middle to late–RC1_0986, RC1_2666) (S5D Fig). There is also upregulation of ftsY^RC1_2195 –a gene involved in the signal recognition particle system for proper insertion of membrane proteins (Fig 3 and S5E Fig) [23, 24].

Charging the membrane for ATP production

Another set of genes expressed at nearly the same early time as those involved in protein synthesis are those involved in generating a membrane potential. Specifically, at 1.5hr after induction of germination there is rapid induction of components of NADH dehydrogenase–complex 1 that moves electron and protons from NADH to the ubiquinone pool (Fig 3, S6A Fig). Complex 1 also shuttles 4H+ across the cytoplasmic membrane so its expression helps germinating cysts develop a membrane potential that can be used for energy production (Fig 3). Concurrently, there is early (0.5 to 1 hr) induction of genes coding for subunits of an F-type H⁺-transporting ATPase (complex V) that utilizes a proton gradient to generate ATP (Fig 3 and S6B Fig). Starting at 0.5 hr, and peaking at ~2.5hrs, there is also induction of genes coding for components of pyruvate dehydrogenase that utilizes pyruvate to generate NADH, acetyl CoA and CO₂ (S6C Fig). NADH from this reaction could also be used by NADH dehydrogenase to sustain energy production via ATPase. We also note a 4 to 5-fold in upregulation of pyruvate dehydrogenase genes during cyst formation, just like that observed during cyst germination [21] (S6C Fig).

Drug resistance

Another set of genes that undergo early expression increases are three genes coding for components of two multidrug exporters (emrA^RC1_4050, emrB^RC1_4051, and acrB^RC1_2565) (Fig 3, S7 Fig). The macrolide transporter MacBRC1_2439 also gets upregulated, though slightly later at 2 hrs. This expression pattern presumably confers resistance to extracellular toxins, thereby providing germinating cysts a defense mechanism during this transient and vulnerable stage. Expression of multidrug transporters during germination is also seen in Bacillus subtilis [25]. Note that the expression of these three resistance genes are markedly reduced when cells undergo the development of cysts. Since cysts are metabolically dormant, there is presumably little need for expression of these drug exporters while in the cyst phase.

Overall, these results collectively show that germinating cysts initially induce expression of genes involved in protein synthesis and NADH dehydrogenase that utilizes NADH for generating a membrane potential. Concurrently, there is the expression of ATPase that uses a membrane potential for the production of ATP. These results indicate that there is an initial preference for both the production of protein synthesis machinery and the initiation of energy production. Finally, there is also an early expression of proteins involved in defense against antimicrobial compounds.

Early to middle events (1–3 hrs.)

Cell wall and membrane remodeling

The cell membrane and outer cell wall are quite different between desiccation resistant cyst cells that do not undergo replication versus replicating vegetative cells (Fig 4). In A. vinelandii, where the cysts cell membrane has been best studied, phospholipids present in vegetative cells are converted to 5-n-alkylresorcinols and 6-n-alkylpyrones during cyst development [26, 27]. These alkyl cyst lipids are thought to create a more impermeable membrane matrix, thereby conferring desiccation resistance [26, 27]. Not surprisingly, the fabZ, fabF, fabH, fabG, and fabK genes involved in fatty acid synthesis [28] all exhibit upregulation in the early to mid-timepoint (S8A Fig).

Cysts are also coated with a very thick layer of exopolysaccharides, so it is also not surprising that there are many changes in expression of genes involved cell wall metabolism early in germination (starting at ~1 hr) and throughout the germination period (Fig 4, S8 Fig, S9 Fig). For example, genes involved in peptidoglycan biosynthesis (murb, murG, murD, mraY, murF, ddiB and uppS) (S8B Fig) [29] and genes involved in building components of the outer membrane, such as those involved in synthesis of the lipid A-core (lpxA, hrtB, RC1_0231, RC1_1918) (S9A Fig) show upregulation in early to mid-timepoints. It is to be noted that the same mur peptidoglycan biosynthesis genes decrease expression during development of cysts (S8B Fig). However, genes involved in synthesis of the core oligosaccharide layer and/or O antigen layer (algC, galU, rkpK, rfbA, rfbC, rfbD) all exhibited upregulation later at ~4–6 hours after induction of germination (S9B Fig) indicating later completion of cell wall synthesis around just before the cyst husk breaks open. Interestingly, genes involved in anhydromuropeptide recycling (nagA and RC1_0623), a process that happens during cell growth [30], shows downregulation (S9C Fig). This result suggests an absence of cell wall recycling during the transition of cysts from dormancy into vegetative cells.

A set cell wall of genes coding for enzymes involved in the production of UDP-α-D-galactose (galE and exoB) and GDP-L-fucose (RC1_3992) undergo opposing expression patterns (S9D Fig). Specifically, these genes exhibit increased expression during cyst development and decreased expression during cyst germination. This expression pattern suggests that these genes may code for enzymes involved in modulating cell wall changes specific to the cyst cell wall.

DNA replication and Nucleotide metabolism

Time-lapse videos of germination (see Fig 2 and S1 Movie) indicate that motile cells emerge from the exine coat in pairs suggesting that cellular replication occurs as part of the germination process. This hypothesis is supported by the expression pattern of genes associated with DNA replication. For example, 2.5 hours after induction of germination there is significant upregulation of two genes coding for subunits of DNA polymerase III (RC1_0787, RC1_2553) as well as ruvB coding for a DNA replication helicase (Fig 4, S10A Fig). At 2–2.5 hrs. post initiation of germination, there is maximal expression of the chromosome partitioning protein (parB^RC1_2813) [31]. We suspect that dormant cyst cells do not contain multiple replication forks and note that it takes 60 to 90 min to replicate a 4 Mbp prokaryotic genome. Consequently, the initiation of DNA replication would likely occurs well before the completion of chromosome replication and subsequent segregation into daughter cells that occurs late in germination [32, 33]. Along these lines there is also ~4-fold increased expression of hupB that codes for HU that is involved in chromosome packing at 2.5 hrs post initiation of germination (S10A Fig). Finally, most of these DNA replication genes are also down-regulated during cyst development which is not surprising as dormant cysts are not undergoing chromosome replication.

In regard to nucleotide metabolism, there is upregulation of a transcriptional repressor of ribonucleotide reductases (nrdR^RC1_1341) concurrent with downregulation of ribonucleotide reductases (nrdR^RC1_1341) that are involved in the synthesis of deoxyribonucleotides (S10B Fig). While down-regulation of ribonucleotide reductase may seem contradictory to the hypothesis that DNA replication occurs during early/mid germination, Gram-positive spores are known to contain a significant reserve pool of deoxyribonucleotides [34]. Consequently, it is also possible that Gram-negative cysts may likewise retain a deoxyribonucleotide reserve pool. Along this vein, we note that at ~1hr, and proceeding throughout germination, there is a significant increase in expression of a gene coding for a nucleoside-diphosphate kinase (RC1_1554) (S10C Fig). This enzyme promotes an exchange of phosphate between different nucleoside triphosphates that would presumably help promote a balanced equilibrium between the various nucleoside triphosphate pools [35].

Finally, we also observed early/mid upregulation of RC1_1499 involved in the synthesis of (p)ppGpp (S10C Fig). This expression pattern during germination is opposite to that observed during cyst development. Furthermore, spores of Bacillus megaterium are known to synthesize (p)ppGpp minutes after the induction of germination [36]. In addition, there is mid to mid/late expression of an inorganic pyrophosphatase gene ppa (S10D Fig). Pyrophosphate is generated as a byproduct of DNA and RNA synthesis and needs to be broken down to inorganic phosphate to maintain a balanced phosphate pool, which is essential for cellular viability [37, 38].

Middle to late events (3–6 hrs.) and later (6–8 hrs.)

Central metabolism

There is a clear increase in expression of genes coding for enzymes involved in central metabolism in the mid to late phase of the germination cycle. Specifically, there are significant increases in the expression of genes coding for enzymes that flow carbon from pyruvate into the TCA cycle which drives NADH production (Fig 5, S6C Fig and S11A Fig). There is also increased expression of respiratory components such as succinate dehydrogenase (Complex II), cytochrome bc₁ (Complex III), as well as components of a cbb₃ type cytochrome c oxidase (Complex IV) (S11B–S11D Fig) that utilizes electrons from NADH and succinate dehydrogenase to drive oxidative phosphorylation. There are also mid/late increases in expression of ubiB that codes for a protein catalyzing the first monooxygenase step in ubiquinone biosynthesis. This is a step in ubiquinone biosynthesis that is proposed to have a regulatory role in the synthesis of ubiquinone (S11E Fig) [39].

At the same time that germinating cysts are ramping up expression of genes involved in aerobic respiration, they are also significantly reducing expression of genes involved in anaerobic fermentation that lead to the production of acetate and lactate respectively (Fig 5 and S12A and S12B Fig). The same reduced expression also occurs for genes that code for enzymes involved in fatty acid β-oxidation that promotes the brake-down of fatty acids into acetyl CoA, genes for the conversion of acetyl CoA to acetate, and genes that code for enzymes involved in PHB synthesis (S12C and S12D Fig). Interestingly, all observed changes in gene expression that affect central metabolism during germination are inverse to what has been observed to occur during cyst development [20] (Fig 6). This opposing regulation fits the logical narrative that cyst development and cyst germination are bifurcating processes and also validates both data sets. Furthermore, it is rather clear from these observed changes that cysts derive energy via anaerobic fermentation while germinating cells reprogram metabolism to derive energy via oxidative phosphorylation.

Amino acid biosynthesis

While the protein synthesis machinery showed upregulation at the earliest time point, amino acid biosynthetic pathways and many aminoacyl tRNA synthetases exhibit upregulation only at the middle to late time points (S13 Fig). We suspect that endogenous amino acids pools are present in dormant cysts for use during initial protein synthesis, as seen in germinating spores of Bacillus megaterium [40]. We speculate that protein synthesis with newly synthesized amino acids happens only after the middle time point.

While most genes involved in amino acid biosynthesis are upregulated in this data set, there are a few interesting examples of decreased expression of important genes in this category. This includes reduced expression of two NifS isomers (RC1_0478 and RC1_1616) coding for putative cysteine desulfurases that generates Ala and elemental sulfur from Cys (S13 Fig). This reaction has been implicated in the generation of sulfur for use in the assembly of iron-sulfur centers in nitrogenase as well as for the sulfur in the ring of biotin [41, 42]. Germination was initiated in a nitrogen-rich complex medium; so, these germinating cells may be nitrogen replete and therefore not assembling an active nitrogenase. This hypothesis is supported by reduced expression of the small chain of glutamate synthase that is also needed for nitrogen assimilation (S13 Fig).

Photosynthesis

At ~3 hours post induction, when cells are ramping up expression of genes involved in TCA cycle/oxidative phosphorylation, these cells are also ramping up expression of genes involved in the synthesis of a photosystem (Fig 5 and S14 Fig). Specifically, several bch and hem genes coding for enzymes for bacteriochlorophyll and heme biosynthesis exhibit elevated expression 3 hrs. post induction of germination. At ~6 hrs. post induction of germination, the puf and puh genes coding for light harvesting I and reaction center structural components also ramp up expression (S14 Fig). This expression pattern presumably ensures that mature vegetative cells exit dormancy fully capable of producing energy via photosynthesis.

Motility

R. centenum is known to synthesize two different flagella types, a sheathed polar flagellum and unsheathed lateral flagellum [43, 44]. Two sets of genes are present in the chromosome that are dedicated either to lateral or polar flagella components with phylogenetic analysis indicating that lateral flagellar genes were obtained via lateral gene transfer [45, 46]. In liquid medium, these cells only synthesize a polar flagellum, whereas when grown on an agar surface they synthesize both flagellum types [44, 47]. Cysts in this study were obtained from agar grown cultures that would synthesize both flagellar types while cyst germination occurred in liquid medium where only the polar flagellum is synthesized. It is therefore not surprising that we observed a ramp-up of polar flagellar gene expression and down-regulation of lateral gene expression during germination (S15A and S15C Fig, respectively).

In addition to flagellar structural genes, the genome of R. centenum also codes for three independent Che-like gene clusters [45]. The Che1 gene cluster involved in regulating chemotaxis and phototaxis [48, 49], the Che2 gene cluster involved in regulating lateral flagellum synthesis [44], and the Che3 gene cluster involved in regulating cyst cell development (discussed in the Signal Transduction section below) [8, 10, 50]. Interestingly the Che1 and Che2 gene clusters both mimic their roles relative to polar and lateral flagella gene expression. Specifically, the che1 operon responsible for phototactic and chemotactic behavior, shows a significant increase in expression ~ 3rd hours onwards as do the expression of most polar flagellar structural genes (S15A and S15B Fig). Likewise, the che2 operon that is involved in the control of lateral flagellar synthesis [44] is downregulated during germination as are that of the lateral flagella structural genes (S15C and S15D Fig).

R body genes

During middle/late germination, there is also a notable ramp-down in the expression of genes that code for R bodies (RC1_ 1995, RC1_ 1997, RC1_ 1999) (S16 Fig). R bodies are a proteinaceous coiled ribbon that is thought to confer resistance to predation [51]. During cyst development, R body gene expression is significantly ramped-up presumably in response to the fact that cysts are non-motile, so the production of a predation defense is likely advantageous.

Gene expression events that span early through late germination

Transport

While it is likely that transport of organic and inorganic materials across the cell wall/membrane is limiting during dormancy, transport systems most likely become important during cyst germination. For example, spores of Bacillus exhibit high specificity of timing on what can enter and exit the cell at different periods of germination and outgrowth [12, 52]. A varied expression pattern for genes involved in transport during R. centenum cyst germination likewise shows the existence of similar selective permeability. For example, genes coding for zinc/manganese/iron transport system (troC^RC1_2461 and troD^RC1_2462) show upregulation at early time points (0.5 hrs.– 1.5 hrs) similar to that observed in Bacillus spore germination (Fig 3 and S17A Fig) [25]. However, unlike Bacillus, genes coding for ABC transporters of phosphate (pstS^RC1_3262, pstC^RC1_3263, and pstB^RC1_3265) also show early upregulation (Fig 3 and S17B Fig). Though single genes involved in the efflux of monovalent cations show both early (RC1_2885) and late (RC1_3663) downregulation (S17C Fig), genes coding for a multicomponent Na+/H+ antiporter (RC1_3665, RC1_3666, RC1_3667, RC1_3668 and RC1_3669) [53, 54] show early upregulation (Fig 3 and S17C Fig) indicating an efflux of monovalent cations early in germination, like that observed during Bacillus germination [12, 25]. Also, it is of relevance here to note that a gene for K+ stimulated pyrophosphate energized sodium pump goes up late (between 6–8 hours) in germination (S17D Fig). This gene is known to code for a pump that pushes sodium ions across the cytoplasmic membrane by using energy from the hydrolysis of pyrophosphate [55]

Later during germination at ~2 hrs, there is increased expression of lipopolysaccharide transporters (RC1_0820, YjgP^RC1_0756, YjgP^RC1_0757, and YjgP^RC1_1564) which corresponds to the time of increased expression of enzymes involved in LPS synthesis (Fig 4, S18A Fig and S9A Fig). Aptly, RC1_0820, and RC1_1564 were previously shown to be downregulated during the formation of non-dividing dormant cysts (S18A Fig) [21]. Also, genes annotated to be involved in molybdenum transport (RC1_2216, RC1_2218, RC1_2219) show downregulation during this time point (S18B Fig).

ABC transporters specific for sugar transport (RC1_0249, RC1_0250, and RC1_0251) show upregulation at the middle to late timepoint ~4 hrs into germination at the same time that enzymes for central metabolism (notably component of the TCA cycle) are ramped up in expression. We speculate the necessity to rapidly supply a germinating cyst with sugar for efficient energy generation. (Fig 5 and S19A Fig) [21].

Genes involved in branched-chain amino acid transport (livG^RC1_1833), putrescine transport (potI^RC1_1457, potH^RC1_1458, and potG^RC1_1459), and oligopeptide transport (oppB^RC1_3364 and oppC^RC1_3365) are upregulated during the early, middle and late time points, respectively (Fig 5 and S19B, S19C and S19D Fig). A gene for potassium transport (RC1_0384) is upregulated late in germination (6 hours) (Fig 5 and S19E Fig), unlike Bacillus where potassium transport is known to be upregulated early in spore germination and outgrowth [25]. Conversely, genes in the operon involved in the active transport of potassium (kdpC^RC1_0434 and kdpA^RC1_0436) are all downregulated (S19F Fig). Also, unlike Bacillus, where copper transport is upregulated [25], this data set shows down-regulation of the copper transport gene (nosF^RC1_3910) late in germination (S19G Fig).

It is at the middle to late timepoints where genes of the tonB transport system (exbB^RC1_0681, exbD^RC1_0682, and exbD1^RC1_0683) that are involved in the transport of iron, nickel, small molecules, vitamin B₁₂ and carbohydrates show upregulation (Fig 5 and S20A Fig). Eleven TonB dependent receptors also show differential regulation during various times. Specifically, 6 TonB receptors are downregulated (RC1_0278, RC1_2392, RC1_2946, RC1_3723, RC1_3920 and RC1_3980) while 5 are upregulated (RC1_0402, RC1_0463, RC1_0580, RC1_0945 and RC1_3711) including RC1_0945 which is a siderophore specific TonB receptor (Fig 5 and S20B Fig). Interestingly, many of these TonB system genes also show an opposing expression pattern during cyst formation (S20A and S20B Fig) [21].

Transcription and signal transduction

To access whether it is advantageous to initiate germination, dormant cysts presumably monitor both their extracellular environment and their internal metabolic state. Once it is perceived that its advantageous to germinate then these cells initiate an ordered cascade of gene expression changes that lead to germination. Not surprisingly, a large number of genes coding for sigma factors, other transcription factors, and signaling proteins change expression throughout the germination cycle. Presumably some function as master regulators and others as downstream players that orchestrate the germination event.

Early on at ~1.5 hrs, there is ramped up expression of two genes coding for subunits of RNA polymerases (rpoC^RC1_0702, rpoA^RC1_0735) (S21A Fig). Increased RNA polymerase synthesis is not unexpected, given that there are numerous increases in gene expression early in germination. There are also nine sigma factors that exhibit changes in expression at various times during germination (S21B Fig). Intriguingly, RC1_1842, RC1_2001, RC1_2169, and RC1_3812 exhibit reduced expression during germination while during cyst formation, these same sigma factors undergo increased expression [21]. This bifurcation in expression indicates that they may have a role in cyst development or perhaps cyst maintenance. There is also, a gene coding for sigma factor regulatory protein (fecR^RC1_3844) involved in ferric iron transport that is upregulated early and throughout the germination process (S21B Fig).

Looking at annotated transcription factors, there are 39 that show significant downregulation and 16 that undergo increased expression throughout the germination time spectrum (S22A, S22B and S22C Fig). Notable on the list are the transcription factors RC1_1377 and RC1_3187 that show very early increases in expression at 0.5 hrs and 1.5 hrs after induction of germination. Presumably, these would be good candidates as regulators involved in early germination events. It is of note that some of these transcription factors show an opposite direction of regulation during germination versus cyst development. For example, RC1_1377 shows early ramped-up expression during germination but also a significantly ramped-down expression during cyst development (S22C Fig). This is an expression pattern expected from a master controller of either the germination or cyst development process. Later, at 4–5 hrs into germination, there is increased expression of a MucR type regulator (RC1_2608) and an ExpG regulator that in other species are known to regulate expression of genes involved in exopolysaccharide synthesis. These regulators are likely involved in controlling genes involved in late cell wall remodeling before emergence from the exine husk.

Regarding signal transduction, there are 12 histidine kinases (HKs) and 14 response regulators that are differentially expressed during germination (S23 Fig). The first histidine kinase to show upregulation is RC1_0476 at 2.5 hours post induction of germination followed by HK RegB^RC1_2541 that increases expression ~3 hrs post induction of germination (S23A Fig). RegB^RC1_2541 is a redox monitoring HK that is known to coordinate synthesis of the energy generating pigment and protein components of the photosystem as well as respiratory electron transfer components (cytochrome oxidase, ubiquinones etc.) [56]. These energy generating processes are also synthesized at ~ 3hrs post induction of germination (Fig 5). RC1_0688 and RC1_3999 also show opposing regulation during cyst formation and cyst germination (S23A Fig) suggesting that they may be directly involved in regulating one or more steps of these developmental processes.

Among the seven response regulators that show upregulation, the earliest is narL^RC1_1407, involved in nitrogen metabolism, showing significant expression at 3 hours post induction. ctrA^RC1_1752 and divK^RC1_1533, are also two cell cycle response regulators that also show upregulation late in germination at ~6 hours post induction. RC1_1592, RC1_1779, RC1_2146, and RC1_2399 also show opposing regulations during germination and encystation and are thus good candidates for regulatory genes involved in germination (S23B Fig).

Finally, expression of the che3 gene cluster is also ramped down as cells undergo germination (S23C Fig). The che3 signal transduction cascade is known to be involved in the formation of cyst formation, so it is not surprising that its expression is ramped down during germination [8, 10, 50].

Conclusions

Microscopic morphological analysis, coupled with transcriptomics during R. centenum cyst germination, has allowed us to obtain chronological parsing of transcription events that occur during germination. Pathways and processes can be readily binned into ‘early’ (within 1.5 hours), ‘early-mid’ (1–3 hours), ‘mid-late’ (3–6 hours) and ‘late’ (6–8 hours) time frames which show that there is a distinct temporal progression of gene expression during germination (Fig 7). Distinct temporal expression changes sequentially promote the following cellular processes; (i) a significant ramp up in protein synthesis capability, (ii) concurrent early synthesis of NADH dehydrogenase that promotes production of a proton gradient, concurrent with the synthesis of ATPase that utilizes a proton gradient to generate ATP, (iii) ramped up synthesis of multi-drug exporters providing resistance to microbial toxins (antibiotics). These early events are followed by early/mid germination expression changes that promote (IV) lipid and LPS remodeling and (iv) DNA replication. Later mid/late germination events then promote (v) changes in central metabolism and the synthesis of electron transfer components that favor energy production via oxidative phosphorylation, (vi) synthesis of a photosystem followed by, (vii) synthesis of flagellar and chemotaxis components promoting motility. Throughout the germination process, there are numerous expression changes in transcription and signal transduction factors that presumably function as conductors that orchestrate observed temporal changes in gene expression.

Finally, comparative analysis of germination gene expression changes in this study with prior analysis of expression changes during cyst development reveals that there are numerous examples of refractile changes in gene expression between the germination and development of cysts. As shown in Fig 7, there is ramped-down in expression of protein synthesis machinery, chromosome replication, and lipid biosynthesis late in cyst development (Fig 7, dotted lines). In contrast, there is ramped-up expression of these same processes during cyst germination (Fig 7, solid lines). There are also numerous regulatory proteins (transcription factors, sigma subunits, signal transduction components, etc) that undergo an inverse expression pattern between cyst cells that are undergoing germination versus vegetative cells that are undergoing the development of cysts. These are potential candidate genes that may be directly involved in regulating either germination or cyst development. Future analysis of these regulatory genes should provide valuable molecular insight on how R. centenum and other Gram-negative species develop dormant cyst and how cysts undertake the process of germination.

Finally, beyond our study on the germination of a Gram-negative cyst, and studies of Gram-positive spore germination [12–14], there are very few reports on what occurs as cells emerge from dormancy. Like R. centenum cyst germination, Gram-positive spore germination also appears to involve programmed expression of different cellular processes such as cell wall remodeling, alteration of lipid composition, followed by increased gene expression involving transport, protein synthesis, and DNA replication [12–14]. The order by which individual processes undergo altered expression is slightly different between cysts and spores but it’s clear that emergence from dormancy in both of these cell types involved ordered programed events. Beyond spores and cysts, it is well known that cells in biofilms and late stationary cultures form non-replicating persister cells that exhibit significantly reduced metabolism [57–59]. Due to their low metabolism, they are often resistant to antibiotics that are designed to target cells with more active metabolism and replication [60]. Like classic dormant cysts and spores, the addition of fresh nutrients has been observed to immediately resuscitate persister cells [58]. Not unlike R. centenum cysts that prioritize ribosome synthesis early in germination, the ribosome content of persister cells are known to play a role in persister cell awakening with cells that have higher levels of ribosomes waking up first [61, 62]. This phenomenon has been attributed to the mandatory need for protein production for successful persister cell awakening [61, 62]. One long-term goal of this study, and others, is to determine if there are commonalities in processes by which different species both become dormant and also awaken from dormancy. If commonalities exist, then it may allow the development of tools to control these cellular events.

Methods

Bacterial strain, media, and cyst isolation

Wild type (WT) Rhodospirillum centenum (ATCC51521) was used throughout this study [47]. CENS medium [43] was used for routine growth, while CENS containing 8 times the nitrogen content (CENS-8XN) was used for inducing cyst formation [5].

For making cysts, a single colony of R. centenum was transferred from agar solidified CENS medium into 5mls of CENS liquid medium and then aerobically grown with shaking at 37°C. 50μL of this overnight culture was then transferred to 25mls of CENS liquid medium and grown aerobically at 37°C for 15–16 hours. OD₆₆₀ of this culture was adjusted to 2.0 with 1ml of the OD₆₆₀ adjusted culture spun down at 8,000 × g for 2 min, washed thrice with 40mM pH 7.5 KHPO₄ buffer and resuspended in 50μL of 40mM pH 7.5 KHPO₄ buffer. This resuspension was used to make four 5μL spots on dry CENS-8XN plates that were incubated at 37°C for 60±5 hours. Presence of a large fraction of cysts was subsequently confirmed using 100x oil immersion microscopy. For cyst isolation, residual vegetative cells were removed from cyst containing wrinkled colonies by addition of lysozyme, which selectively lyses replicative vegetative cells but not dormant cysts. For lysozyme treatment, wrinkled colonies from CENS-8XN plates were harvested after 60±5 hours of growth by resuspending approximately 3.53 × 10⁶ cells/ml in 40mM pH-7.0 KHPO₄ buffer and subsequently pelleting cells by centrifugation at 8,000 × g for 2 min. The pellet was resuspended in TE buffer containing 200μg/ml of lysozyme, incubated at 37°C for 30 minutes and washed multiple times with 40mM, pH 7.0 KHPO₄ buffer. Cysts present in the resulting pellet, were >90% enriched over vegetative cells as based on flow cytometric analyses (S1 Fig). Enriched cysts were prepared fresh before each germination assay/study.

Microscopy

Cysts harvested from 60±5 hour colony grown on CENS-8XN plates, were inoculated on top of a 1% CENS agarose pad on a concavity slide to which a coverslip was added and sealed using VALAP [63]. This slide was incubated on a 60X oil immersion objective of a Nikon TiE inverted microscope with an objective warmer set at 37°C. The cysts on the slide were recorded based on their X, Y coordinates using NIE elements viewer software© and imaged every 5 minutes (phase contrast) for the next 15 hours. ImageJ (FIJI) was used for adding the time stamp and for stitching the images into a movie. Negative staining electron microscopy and scanning electron microscopy was undertaken as described in ref [1].

RNA-Seq transcriptomic analysis of germination

Approximately 1.13 x 10⁵ isolated cysts/ml were inoculated into nutrient-rich CENS medium in 2L Erlenmeyer flasks and incubated at 37°C with continuous stirring on a magnetic plate. At 0 hr, a 10ml control sample was taken immediately after addition of CENS medium to cysts. 10ml samples were then taken every half an hour for the first 3rd hours and every 1 hour after that until a final sample at the 8th hour. Upon collection, samples were immediately pelleted in a cold centrifuge and resuspended in 1ml of Trizol and transferred to a lysing matrix B tube® (MP Biomedicals), lysed in a Fastprep® instrument (MP Biomedicals), and spun down at 8,000 × g. The resulting supernatant was stored at -80°C until RNA extraction.

RNA was extracted using the DirectZol^™ RNA miniprep plus (Zymo Research) using manufacturer's instructions. Briefly, after adding an equal volume of ethanol to the Trizol^™ lysed sample, the mix was loaded on to Zymo-Spin^™ IIICG column, centrifuged, flow through discarded and the in-column DNase I treatment was done. After washes, total RNA eluted in nuclease-free water.

Quality control, library preparation, and RNA sequencing

Analysis of RNA quality, cDNA library preparation, and deep sequence analysis was undertaken by the Gene Expression Center and DNA Sequencing Facility at the University of Wisconsin-Madison Biotechnology Center. Briefly, RNA concentration and quality were determined using nanodrop and Agilent 2100 bioanalyzers. Samples were checked for sharp 16S and 23S rRNA peaks before downstream processing. Ribosomal RNA from the total RNA was removed using EpiCentre® Ribo-Zero^™ Magnetic (Bacteria) kit with an input quantity of 1.5μg total RNA. TruSeq^™ RNA Sample Prep kit (Illumina, San Diego, CA) was used to prepare cDNA mRNA-seq libraries. Standard Illumina protocol was then used to perform single end sequencing (1X100bp) on Illumina HiSeq 2000 to a level that provided over 200-fold coverage per nucleotide of the genome.

Raw data processing and data analysis

In-house data analysis pipeline was used for raw data processing. Briefly, Trimmomatic [64] was used to trim the reads off adaptors with low-quality reads from the ends (sliding window– 5:25; minimum read length– 40bp). Bowtie2 [65] then used to align trimmed reads to R. centenum reference genome followed by HT-seq-count [66] to detect the number of reads per gene. Overall there were over 130 million (M) (strand specific) RNA-Seq reads with >99% of these reads mapped to the R. centenum genome yielding over 200X coverage per nucleotide. This sequence data has been deposited in NCBI’s Gene Expression Omnibus [67] and are accessible through GEO Series accession number GSE132286 (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE132286).

Differential expression between the 0th-hour sample, and every other time point, was then calculated using the DE-Seq2 package in R [68]. A gene was considered as a Differentially Expressed Gene (DEG) when the log₂ fold change was ≥ ± 1.5 with a false discovery rate adjusted p-value of <0.05. The differentially expressed genes were sorted into COGs (Cluster of Orthologous Groups) using Eggnog 4.5.1 [69] and Kegg [45, 70–72] with COG X assigned to photosynthesis genes [73]. Final manipulation with the output file (Tab-delimited text file) was done using Microsoft Excel^™.

qRT-PCR validation of RNA-seq results

Five genes that were differentially expressed in the RNA-seq data sets were chosen (RC1_3511, RC1_1228, RC1_0470, RC1_0099, RC1_0097) and independently analyzed for expression using qRT-PCR. Briefly, RNA extracted from samples collected during specific timepoints of germination (0hr, 0.5hr, 2.5hr and 6 hr.) for 3 biological replicates were reverse transcribed using Takara PrimeScript^TM RT Master Mix. TB Green Premix Ex Taq^TM II with ROX was used to perform qRT-PCR in an Applied Biosystems StepOne Plus instrument. The thermocycling conditions used were as follows: 95°C for 30 seconds (1 cycle), 95°C for 5 seconds and 60°C for 30 seconds (40 cycles) followed by a melt curve stage. Bestkeeper analysis and previous qPCR validations have identified RC1_1500 to be an ideal housekeeping gene in R. centenum [74, 75] and the same has been used here for normalization. Log₂fold change with respect to time 0 for each gene and timepoint were calculated and plotted against the log² fold change of the RNA-seq data (S2 Fig).

Supporting information

S1 Movie [avi]
A video showing cyst germination on a 1% CENS agarose pad incubated at 37°C over an 8 hour time period.

S1 Table [deg]
A spread sheet listing 1667 genes that undergo differential expression between the 0th-hour sample, and every other time point, was then calculated using the DE-Seq2 package in R [].

S1 Fig [tiff]
Representative flow cytometry data showing the enrichment of cysts for germination studies.

S2 Fig [tiff]
Correlation validation plot between RNA-seq data and qRT-PCR data sets.

S3 Fig [green]
Differentially expressed genes that are binned into COGS.

S4 Fig [pink]
Comparison of transcriptomics during encystation (cyst formation) and cyst germination.

S5 Fig [tiff]
Expression heat maps of genes involved in protein synthesis.

S6 Fig [c]
Early expression of protein complexes involved in formation of a membrane potential and use of this potential for energy formation.

S7 Fig [tiff]
Heat maps of multidrug transporters show early expression during cyst germination.

S8 Fig [a]

S9 Fig [a]

S10 Fig [a]

S11 Fig [tiff]
Heat map of genes coding for electron transfer components involved in respiratory oxidative phosphorylation.

S12 Fig [c]
Reduction of anaerobic fermentation processes midway/late in germination.

S13 Fig [tiff]
Heat map of gene expression changes for genes involved in FeS and amino acid biosynthesis.

S14 Fig [a]

S15 Fig [a]

S16 Fig [left]

S17 Fig [a]
Heatmaps show the expression profiles of various transport systems expressed at the early time point of germination.

S18 Fig [a]

S19 Fig [a]
Heatmaps show the expression profiles of transport systems expressed at the middle to late time points of germination.

S20 Fig [a]

S21 Fig [a]

S22 Fig [a]

S23 Fig [a]