Which of the following transcriptional control features are found in both prokaryotes and eukaryotes?

Introduction

Prokaryotic cells such as E. coli can regulate their gene expression to adapt to the changes in available nutrients in the surrounding environment. Their regulation of gene expression occurs mainly through regulation of the rate of mRNA synthesis, since their transcription and translation is coupled. For example, in a lactose rich medium, Escherichia coli activates expression of the gene clusters called operons that are required for the processing of lactose, e.g., β-galactosidase, permease, etc. However, eukaryotic multicellular organisms regulate their gene expression through five major levels: modification of chromatin structures, the number of mRNA molecules synthesized (transcriptional), processing of the mRNA (post-transcriptional), the rate of mRNA translation (translational), and modifications of the proteins for proper function (post-translational regulation). Eukaryotic regulation of gene expression is thus much more complex than prokaryotic gene expression and involves a variety of factors and elements. In this chapter, the numerous mechanisms and factors by which regulation of gene expression is achieved in prokaryotes and eukaryotes are described.

I The Diversity of Prokaryotic Organisms

Prokaryotic cells represent the smallest and simplest form of life that can metabolize, grow and reproduce. They are presumed to resemble the earliest forms of life and they reproduce much more quickly than multicellular organisms do. Taken together, these two properties imply that prokaryotic organisms have had more opportunity to evolve (by orders of magnitude) than plants or animals have had; accordingly, this predicts that prokaryotes should have the most functionally efficient, diverse and specialized of cells, despite their structural simplicity. The key to this paradox is the recognition that genetic variation and natural selection should allow a unicellular organism to improve its performance and acquire new functions without becoming structurally complex. Furthermore, as environments, survival strategies and ecological niches change over time, the criteria of optimal cellular function change which, in turn, sets the stage for new rounds of optimization in various directions. The resulting diversification and specialization can also be expected to make certain features superfluous in certain lineages, leading to cells that may be even simpler than their predecessor.

Functional specialization and optimization of a structurally simple cell seems to account for the observed diversity of modern prokaryotes. Molecular measures of divergence, such as small-subunit ribosomal RNA sequence, indicate that two prokaryotic lineages separated very early and that each encompasses more molecular diversity than multicellular organisms (Fig. 50.1). The two groups distinguished by this early split, Bacteria and Archaea, each have phylogenetic status equivalent to that of all eukaryotic organisms and the three resulting taxonomic units have been termed Domains (Woese et al., 1990). This extensive divergence is also evident in terms of cellular function. Certain bacteria and archaea have metabolic properties not represented among eukaryotes, including N2 fixation, anoxic photosynthesis, additional routes of CO2 assimilation and adaptation to extreme environmental conditions. Similarly, archaea (singular: archaeon, or less commonly, archaeum) have cellular features and metabolic pathways not found in bacteria. These uniquely archaeal features include isoprenoid membrane lipids (found in all archaea) and the ability to make methane (found in a number of genera).

One practical consequence of the deep diversity of the bacterial and archaeal lineages is that it precludes any one organism, such as the bacterium Escherichia coli, from modeling all aspects of prokaryotic physiology, even though this and several other species can be analyzed in great detail. It should also be noted that some components of eukaryotic cells have bacterial origins. In particular, at least two eukaryotic organelles, the mitochondrion and the chloroplast, resulted from endosymbiotic acquisitions of bacteria by progenitors of modern eukaryotic cells (Scwartz and Dayhoff, 1978). This relationship provides a context for understanding molecular structure and function of both the eukaryotic organelle and the bacterial cell.

CELL WALLS

Prokaryotic cells are surrounded by a generally rigid cell wall, protecting the cell from osmotic lysis (Figs. 5.5 and 5.6). The cell wall has also been important in identification and classification of bacteria, providing a major division between gram-positive and gram-negative bacteria, defined on the basis of the Gram stain. The cell wall of gram-positive cells consists of a single thick layer of peptidoglycan, surrounding the cytoplasmic membrane. Peptidoglycan is a polymer consisting of a backbone of alternating N-acetylglucosamine and N-acetylmuramic acid residues connected to cross-linked peptide chains of four amino acids. Gram-positive cell walls also usually contain teichoic acids, polymers of glycerol or ribitol, linked by phosphate groups and containing amino acids and sugars. In the more complex gram-negative cell wall, the peptidoglycan layer is much thinner and is surrounded by an outer membrane enclosing a periplasmic space, which contains enzymes involved in nutrient acquisition, electron transport, and protection from toxins. In contrast, archaeal cell walls have variable chemical structure, consisting of proteins, glycoproteins, or polysaccharides, but do not contain peptidoglycan (Fig. 5.6).

The cell walls of many bacteria are encased within extracellular material (Fig. 5.5), ranging from apparently rigid and distinct capsules of specific thickness to more diffuse (chemically and physically) extracellular polymeric substances. Many roles have been assigned to this material, including protection from predation, adhesion to solid surfaces, and biofilm formation. In the free-living N2-fixing bacterium Azotobacter, extracellular material is important in creating anaerobic regions required for N2 fixation. Biofilm formation is particularly important, with suggestions that the majority of the soil microbial community is attached to particulate matter (clay minerals, soil organic matter, plant roots, and animals). Particulate material provides a concentration of nutrients necessary for microbial growth, and surface attachment has been shown to increase survival of bacteria and to protect them from environmental stress, including low pH, starvation, and inhibition by antibiotics and heavy metals. An example is the production by nitrifying bacteria, in model soil systems, of copious amounts of extracellular material that effectively forms a blanket over colonies, such that individual cells are not visible. This occurs despite the fact that these autotrophic organisms gain barely sufficient energy from oxidation of ammonium or nitrite, use much of this energy to generate reducing equivalents, and require more reducing equivalents because of the requirement to fix CO2. However, once formed, biofilms of these organisms are protected from a wide range of factors to which suspended cells are susceptible. Attachment of cells to surfaces is also facilitated by short, hair-like fimbriae, while similar structures, sex pili, are involved in cell–cell contact associated with plasmid transfer.

C Cell Walls

Prokaryotic cells are surrounded by a rigid cell wall, protecting the cell from osmotic lysis (Figs. 3.5 and 3.6). The cell wall is important in diagnosis and phenotypic classification of bacteria, providing a major division between Gr+ and Gr− bacteria. The cell wall of Gr+ cells consists of a single layer of peptidoglycan, surrounding the cytoplasmic membrane. Peptidoglycan is a polymer consisting of a backbone of alternative N-acetylglucosamine and N-acetylmuramic acid residues connected to cross-linked peptide chains of four amino acids. Gram-positive cell walls also usually contain teichoic acids, polymers of glycerol or ribitol, linked by phosphate groups, and containing amino acids and sugars. In the more complex Gr+ cell wall, the peptidoglycan layer is much thinner and is surrounded by an outer membrane enclosing a periplasmic space, which contains enzymes involved in nutrient acquisition, electron transport, and protection from toxins.

Archaeal cell walls differ from those of bacteria and provide a major exception to the original definition of prokaryotes. They lack an outer membrane, and peptidoglycan is replaced by a range of compounds, including pseudomurein (similar to peptidoglycan) or other polysaccharides. Many archaeal cell walls contain paracrystalline protein or glycoprotein, the S-layer, which is also found in some bacteria. In addition, some prokaryotes do not possess a cell wall (e.g., intracellular mycoplasmas and the archaeon Thermoplasma, which have strengthened membranes).

The cell walls of many bacteria are encased within extracellular material (Fig. 3.5), ranging from apparently rigid and distinct capsules of specific thickness to more diffuse (chemically and physically) extracellular polymeric substances. Many roles have been assigned to this material, including protection from predation, adhesion to solid surfaces, and biofilm formation. In the free-living, N2-fixing bacterium Azotobacter (a member of the Pseudomonadales), limited O2 diffusion through extracellular material creates anaerobic regions required for N2 fixation. Biofilm formation is particularly important, with suggestions that the majority of the soil microbial community is attached to particulate matter (clay minerals, soil organic matter [SOM], plant roots, and animals). Adsorption of nutrients to particulate material provides concentrations necessary for microbial growth. Surface attachment can increase survival of bacteria and protect them from environmental stress, including low pH, starvation, and inhibition by antibiotics and heavy metals. An example is the production by nitrifying bacteria in soil model systems of copious amounts of extracellular material that effectively form a blanket over colonies. This occurs despite the fact that these autotrophic organisms barely gain sufficient energy from oxidation of ammonia or nitrite, use much of this energy to generate reducing equivalents, and then require more reducing equivalents because of the requirement to fix CO2. However, once formed, biofilms of these organisms are protected from a wide range of factors to which suspended cells are susceptible (Prosser, 2011). Attachment of cells to surfaces is also facilitated by short, hair-like fimbriae, whereas similar structures, sex pili, are involved in cell-to-cell contact associated with plasmid transfer.

Normal Physiological Processes

Prokaryotic cells derive their energy from glycolysis, a process that oxidizes and splits the six-carbon glucose molecule into two three-carbon pyruvate molecules (Figure 1). The energy released during glycolysis is conserved by the phosphorylation of adenosine diphosphate (ADP) to adenosine triphosphate (ATP) and by the reduction of nicotinamide adenine dinucleotide (NAD+) to NADH. Glycolysis would stop if it lacks supply of NAD+. Higher organisms regenerate NAD+ by reducing pyruvate to lactate, allowing glycolysis to continue indefinitely:

Glucose+2ADP+2Pi→2lactate+ 2ATP

where Pi denotes inorganic phosphate.

The hydrolysis of ATP provides the energy required to power vital cellular processes, such as protein synthesis, locomotion, and membrane-associated ionic pumps:

2ATP→2 ADP+2Pi+2H++energy

Combining the previous equations yields the general expression for glycolysis:

Glucose→2lactate+2H++energy

The emergence of mitochondria, remarkable organelles capable of handling oxygen safely, provided the energy required for the evolution of eukaryotic cells. Mitochondria are thought to be descendants from symbiotic bacteria since they contain their own DNA and replicate independently.

Mitochondria generate vast quantities of energy from O2, generating CO2 and water as by-products. In the mitochondria, pyruvate is oxidized to acetyl-CoA, which enters the tricarboxylic acid cycle, producing CO2 and electrons. These electrons move down a cascade of progressively lower energy states until transferred to oxygen, the final electron acceptor. The energy liberated by the mitochondrial electron cascade is coupled to the partly understood mechanism of oxidative phosphorylation, producing 36 ATP molecules per glucose molecule.

The abundant cellular energy provided by aerobic metabolism allowed eukaryotic cells to form complex multicellular organisms. Growth, however, was limited by the diffusion of O2 from the environment. Species adapted to prevailing ecological conditions by developing organs to carry O2 from the surrounding media to all tissue cells. Animal O2 transport systems consist of heme pigments to reversibly bind O2 (hemoglobin), specialized cells to enfold but not consume the potentially toxic O2 (erythrocytes), organs to promote O2 diffusion from the environment into blood (lungs or gills), and the hydraulic pumps and circulatory conduits needed to convey O2-containing cells to and from the tissues (the heart and blood vessels).

Air reaches the alveolar spaces following a negative pressure gradient during inspiration. Alveolar O2 partial pressure (PAO2) may be estimated from the alveolar air equation

PAO2=FIO2(Patm−Pwater)−PaCO2[FIO2+(1−FIO2)/RQ] mmHg

where FIO2 is the inspired O2 fraction (0.21 for air), Patm is the atmospheric pressure, Pwater is the water vapor pressure, PaCO2 is the arterial blood CO2 partial pressure, and RQ is the respiratory quotient.

When breathing air at sea level, this equation approximates to

PAO2=150−PaCO2/0.8mmHg

Oxygen diffuses from the alveoli into pulmonary capillary blood, where it dissolves in plasma or binds to hemoglobin. The alveolar–arterial gradient is the difference between alveolar and arterial PO2 (A−a=PAO2−PaO2). Ventilation perfusion inequality, shunt, or (rarely) diffusion limitation of oxygen can widen the A–a gradient across the blood–gas barrier.

The sigmoid-shaped oxyhemoglobin dissociation curve (ODC) defines the relationship between plasma PO2 and fractional hemoglobin O2 saturation (SO2). The concentration or O2 content of blood is the sum of O2 bound to hemoglobin and O2 dissolved in plasma:

O2content=13.9×SO2×[Hb]+0.031 ×PO2mlO2perliterof blood

where the term 13.9 is the O2 carrying capacity of hemoglobin, [Hb] is the concentration of hemoglobin in blood (g l−1), and the term 0.031 is the solubility of O2 in plasma for PO2 in mmHg.

Arterial blood delivers oxygen and metabolic substrate to the tissues in accordance to local energy requirements. Higher level control of the circulation is aimed primarily at counteracting gravitational effects by stimulation of arteriolar α and β receptors. The microvasculature – the vast array of terminal arterioles, capillaries, and venules – regulates intraorgan blood flow through vasomotion, rhythmic oscillations in vascular tone.

Accurate measurements of regional O2 delivery are difficult to obtain. A first approximation to the adequacy of cellular O2 transport may be obtained by calculating the systemic rates of O2 delivery (DO2) and O2 uptake (VO2),

DO2=cardiacoutput×O2contentmlO2perminute

Systemic O2 uptake (VO2) may be measured directly from the expired gases as

VO2=[FIO2(1−FEO2−FECO2)/(1−FIO2)−FEO2]×VEmlO2perminute

where FIO2 and FEO2 denote the inspired and expired fractional O2 concentration, respectively; FECO2 is the fractional concentration of expired CO2; and VE is the expired minute volume. A practical problem with this expression is that small inaccuracies in measuring FIO2 at values >0.60 may produce large errors in VO2.

Systemic VO2 may also be calculated with the aid of Fick's principle as

VO2=cardiacoutput×(arterial O2content−mixedvenousO2content )mlO2perminute

This calculation requires the insertion of a pulmonary artery catheter to gain access to pulmonary artery blood. Fick's principle underestimates total VO2 since it does not account for pulmonary VO2. The latter can be substantial in severe pneumonia or in adult respiratory distress syndrome.

The ratio of tissue CO2 production (VCO2) to VO2 is the respiratory quotient,

RQ=V CO2/VO2

The type of substrate consumed determines the value of RQ, being 0.7 for free fatty acids and 1.0 for glucose.

The O2 extraction ratio (ERO2), the fraction of DO2 taken up by the tissues, is another useful parameter:

ERO2=VO2/DO2

ERO2 is approximately 20–25% at rest, increasing to 50% or greater with exercise or during cardiac failure at rest. ERO2 values <15% are suggestive of functional peripheral shunting, as might occur during sepsis.

The DO2–VO2 relationship (Figure 2) is the subject of great clinical interest. Experimental preparations subjected to decreases in DO2 initially show little variation in VO2 resulting from microvascular responses that increase ERO2. Decreases in VO2 occur below a critical DO2 because cellular O2 supply cannot keep up with the rate of aerobic metabolism. Anaerobic processes now must be recruited to complement the faltering mitochondrial supply of ATP. This condition, if protracted, may lead to the death of the animal.

Measurements of DO2 and VO2 from critically ill individuals uniformly show a straight line, one lacking a critical DO2 defining the threshold for anaerobic metabolism. Failure to account for the passage of time in the DO2–VO2 relationship may explain the difference in findings between clinical and experimental data. Animal experiments usually take a few hours, during which time DO2 is relentlessly lowered. Clinical measures are gathered over a longer period, typically several days, and no efforts are made to decrease DO2. During this time, the patient's metabolic rate is bound to change, leading to variations in DO2 in response to changes in VO2, not the reverse. By considering time as an additional dimension, the two-dimensional DO2–VO2 curve becomes a surface whose contour varies in concert with the patient's clinical condition. Figure 3 illustrates the temporal evolution of a hypothetical DO2–VO2 relationship. In this example, the person's condition worsens with time as increases in DO2 become less effective in their ability to increase VO2. The figure also shows that samples of DO2 and VO2 taken at different times (red circles) may project as a straight line if viewed from a two-dimensional perspective.

Polycistronic mRNA

In prokaryotic cells, a single mRNA may code for several proteins. Each message on the mRNA is contained in a single ‘open reading frame,’ a sequence of codons bound by start and stop codons. There are no start or stop codons within the reading frame itself. The arrangement of messages in tandem along a single strand of mRNA allows the proteins (often called gene products) to be translated simultaneously; these gene products are often related in function. Because mRNAs are single stranded, some mRNA molecules are able to base-pair within themselves and can form secondary and tertiary three-dimensional structures. These structures can regulate the synthesis of polypeptides in the polycistronic mRNA. One example of this mechanism is MS2 bacteriophage (Kozak, 1983). The A protein is coded at the 5′ end of the polycistronic message, but is needed in only small quantities. The 5′ end of the mRNA is often blocked by tertiary folding of the mRNA allowing only limited translation of the A protein while allowing translation to occur at more accessible sites downstream from the A gene.

Regulation of De Novo Pyrimidine Biosynthesis

In prokaryotic cells aspartate transcarbamoylase, an allosteric protein, is inhibited by the end product CTP and activated by ATP. In mammals, carbamoyl phosphate synthetase II (CPS II) activity of the trimeric multifunctional protein (Pyr 1–3) is the primary regulatory site. CPS II allosteric ligand UTP inhibits, whereas PRPP and ATP activate, the enzyme activity. Activation by ATP may be important in achieving a balanced synthesis of purine and pyrimidine nucleotides. During rapid proliferation of cells, either as a normal physiological process or in pathological processes (e.g., tumor growth), there is an increased demand for nucleotides required for nucleic acid synthesis. In those periods of cellular growth, CPS II activity is altered by phosphorylation at its regulatory site, leading to relief of UTP inhibition and enhancement in the stimulation of PRPP activation. The phosphorylation is mediated by mitogen-activated protein (MAP) kinase. The MAP kinase cascade is activated in response to growth factors (e.g., epidermal growth factor). PRPP is also an essential (and probably a rate-limiting) substrate for the orotate phosphoribosyltransferase reaction (reaction 5 in Figure 25.16) promoting de novo pyrimidine nucleotide synthesis by induction of Pyr 1–3. Another potential site of regulation is orotidine-5-phosphate decarboxylase, which is inhibited by UMP, CMP, allopurinol nucleotide, and oxypurinol.

C Transcription Initiation

In prokaryotic cells transcription initiation from many promoters is affected by negative superhelicity of the DNA template (Pruss and Drlica, 1989). A recent report describes the stimulation of the leu-500 promoter, a negative supercoiling-dependent promoter, by divergent transcription from the tet promoter of pBR322 plasmid (Chen et al., 1992). Interestingly, transcription from the tet promoter was shown to be responsible for the generation of hypernegatively supercoiled pBR322 plasmids in E. coli topA mutants (Pruss and Drlica, 1986). It was concluded that the negative supercoils generated behind the RNA polymerase transcribing from Ptet were activating the leu-500 promoter. In eukaryotic cells recent in vitro studies have shown that binding of the transcription factor TFIID to the fibroin gene promoter was stimulated by negative superhelicity of the DNA template (Mizutani et al., 1991). Different eukaryotic promoters respond differently to DNA supercoiling in an in vitro transcription system (Parvin and Sharp, 1993). Using a yeast top1 top2ts cellular extract, which is almost free of DNA topoisomerase activities, it was recently shown that transcription from a yeast rRNA promoter was stimulated by negative superhelicity of the DNA template (Schultz et al., 1992). Most interestingly, transcription from a T7 RNA polymerase promoter is shown to activate transcription from an upstream divergent Xenopus rRNA promoter (Dunaway and Ostrander, 1993).

RNA Polymerase

Both prokaryotic and eukaryotic cells employ the multisubunit RNA polymerase (RNAP) complex as the main effector of transcription. Eukaryotic cells employ three types of RNAP (I, II, and III) for synthesizing three different classes of RNA. RNAPII, that synthesizes mRNA, is composed of 10–12 subunits. The RNAPII holoenzyme contains the polymerase associated with GTFs. Functionally, RNAPII is composed of four multisubunit mobile elements:

The core contains the cleft that is surrounded by the other units; the cleft holds the incoming single DNA strand at the active site.

The clamp is a mobile unit connected to the core that opens and closes access to the cleft, stabilizing the DNA–RNA hybrid with the polymerase.

The shelf

The jaw lobe grips the DNA downstream of the active site.

RNAPII passes the template strand through the core cleft, adding base pairs complementary to the template strand to form an RNA strand which is an exact copy of the coding strand, except that RNA contains uracil in lieu of thymine. RNAPII transcribes 3 to 5′ along the template strand, producing a 5 to 3′ RNA strand. Several RNAPIIs can transcribe a single gene, increasing the amount of RNA produced. The transcription bubble is the unwound DNA structure that allows RNAPII to transcribe along the single template strand. As RNAPII progresses, the DNA downstream is unwound. Subunits of TFIIH have helicase and ATPase properties which facilitate bubble formation. Other functional elements of importance are the C-terminal domain (CTD), which is a repeating heptad sequence that is phosphorylated at specific residues (serine 2 and 5), which govern the initiation and elongation steps. Phosphorylation of this scaffold region regulates transcription by releasing inhibitory factors.

Briefly, GTFs bind to a core promoter sequence (TATA), located upstream of the TSS. This closed, multiunit PIC recruits and binds RNAPII. A small section of the DNA double strand separates, and the template strand lies in the active site of the RNA polymerase. Initial attempts at RNA synthesis may be unsuccessful due to the low stability of the short DNA–RNA strand in the active site; once RNAPII synthesizes an RNA strand which is stable, it leaves behind the PIC, “clearing” the promoter. RNAPII proceeds to synthesize.

This event initiates RNA synthesis where the RNAPII complex is released from the promoter region. After this first step of recruitment and initiation, RNAPII undergoes phosphorylation at serine 2 in the CTD (RNAPIIS2), resulting in productive transcriptional elongation along the coding region. Transcription ends with a termination sequence motif (AAUAAA) that halts RNAPII. RNAPII is subsequently dephosphorylated and recycled into a new transcriptional cycle.

Chromatin remodeling. Transcriptional activators bind and induce chromatin remodeling at the promoter region to allow transcriptional machinery access to the transcription start site. Promoters of genes undergoing continual transcription are usually nucleosome-free.

Preinitiation complex assembly. GTFs assemble on the core promoter and recruits RNAPII.

TFIID binds to the core promoter through the TBP subunit; binding may occur at specific elements such as the TATA or Inr. In promoters lacking specific elements, TAFs associate with the promoter and induce nonspecific binding of TBP. TBP binding to the TATA opens the minor groove, bending the DNA at an 80° angle.

TFIIA binds to TBP, stabilizing the interaction between TFIID and DNA.

TFIIB binds to the core promoter through BRE and positions the DNA for entry into the active site. This TFIID/A/B complex recruits RNAPII and TFIIF simultaneously.

TFIIE joins the complex and recruits TFIIH, aiding in opening the DNA double strand, called promoter melting. TFIIF induces torsional strain, forming the transcription bubble.

TFIIH contains a helicase subunit that unwinds the DNA into two strands; it binds to the template strand to ensure transcription of the correct strand. The catalytic activity of TFIIH is dependent upon TFIIE activity, which it interacts with. TFIIH promotes formation of the transcriptional bubble. TFIIH also contains a kinase subunit that phosphorylates RNAPII at the CTD to initiate transcription.

The bridging protein mediator links distal transcriptional activators with the RNAPII complex.

Initiation and clearance. Phosphorylation of the CTD at serine 5 (RNAPIIS5) by TFIIH or a mediator subunit induces RNAPII to clear the PIC and begin transcribing RNA from the DNA template strand; RNAPII may pause approximately 50 bp downstream.

Elongation. Paused RNAPII is phosphorylated at serine 2 (RNAPIIS2) on the CTD by the positive transcription elongation factor b, causing dissociation of the negative elongation factor, allowing RNAPII to proceed into continuous, productive elongation. Chromatin remodeling of nucleosomes occurs as RNAPII transcribes through the coding region.

Termination. Upon reaching the termination signal, the transcribed RNA is released and RNAPII is dephosphorylated, dissociating from the gene to be recycled for subsequent cycles of transcription.

2.5.6.1 Applications of CET to Prokaryotic Cells

Many prokaryotic cells are sufficiently thin to be imaged in toto (i.e., aqueous suspensions of the cells under scrutiny are applied to an EM grid and vitrified as described above). The archeons Sulfolobus and Pyrobaculum were the subject of the first CET studies of whole cells.26 The tomograms revealed the surface layer that is typical for archea and a dense largely homogeneous cytoplasm with a protein concentration of approximately 30–50%. Remarkably, these results were obtained with a 120 keV LaB6 TEM (i.e., at comparably low voltage and without a FEG). Subsequent studies with FEG microscopes operated at higher voltages allowed more detailed insights into archea. For example, in tomograms of ARMAN cells, ultrasmall archea with a diameter of approximately 300 nm, tubular organelles as well as large macromolecular complexes such as ribosomes could be discerned.117

CET has become an increasingly common tool for microbiologists to image bacteria in recent years. In particular, CET has been used to study the bacterial cytoskeleton. In the mollicute Spiroplasma melliferum it has been shown for the first time that long filamentous structures are associated with the cell membrane.118 Similarly, large filament bundles were observed near the membrane, but also in the cytoplasm of Caulobacter crescentus,119 Bdellovibrio bacteriovorus,120 and Treponema pallidum121 cells. In magnetotactic bacteria such as Magnetospirillum gryphiswaldense, it could be shown that magnetosomes, magnetic bacterial organelles, are associated with filaments that span the extended bacteria longitudinally.122,123 In all cases, the filaments are likely assembled either from actin, tubulin, or intermediate-filament homologs, but the precise proteins have not yet been identified.

Analysis of bacterial membranes by CET revealed new insights into bacteria. For example, CET of whole mycobacteria cells of different species revealed differences in the structure of the mycobacterial outer membrane compared to that of other Gram-negative bacteria:124 the cell envelope is composed of several layers, where the outer membrane consists of a lipid bilayer. Moreover, chemoreceptors that are embedded in the inner bacterial membrane have been subject to intensive study using CET. Bacterial chemoreceptors can be resolved as striations orthogonal to the cytoplasmic membrane, located near the pole and a thin line density parallel to the membrane in tomograms of Escherichia coli.125,126 More detailed analysis revealed that chemoreceptors form approximately 12-nm honeycomb-like hexagonal arrays.127 Interestingly, the lateral dimensions of the chemoreceptor arrays seem to be largely invariant for different bacterial species, but the length of the receptors varies depending on subunit sizes.128 Subtomogram averaging has been used to resolve the asymmetric unit of the chemoreceptor arrays to ∼(1/3.3 nm), and tomograms acquired under different chemotactic states revealed conformational changes of the receptors.129 These averages and the distribution of the receptor molecules in the array were used to devise a computational model for chemotactic response, which agreed with fluorescence resonance energy transfer (FRET) measurements.130

Structural studies of flagellar motors in situ have also been accomplished for several organisms, such as Treponema primitia,131 Bdellovibrio bacteriovorus,120 and Borrelia.132,133 The overall architecture of flagellar motors seems to be largely conserved among the different species and the highest resolution subtomogram averages provided the density of the motor in its native membrane to 1/(3.5 nm).

The molecular crowding inside bacterial cells makes it challenging to interpret the interior of bacteria tomograms. Nevertheless, large macromolecules, such as ribosomes can be localized with high fidelity in Spiroplasma and E. coli.88,134 Imaging of starving E. coli cells revealed the structure of hibernating ribosomes.134 Under starvation conditions, 70S ribosomes assemble to a dimeric form (100S ribosomes). Subtomogram averages revealed the quaternary structure of 100S ribosomes from lysates and in whole cells, in which dimerization is mediated by the 30S ribosome subunits (Figure 12). Due to the high conformational variability of these complexes, classification by constrained PCA and ML was essential to capture the prevalent modes of the supercomplex.

In the nucleoid-containing Bdellovibrio bacteriovorus, ribosomes were shown to delineate the nucleoid.135 Even smaller complexes, such as GroEL can be identified in extremely thin cells, such as Leptospira interrogans.90