|Home | About | Journals | Submit | Contact Us | Français|
MYC homeostasis is critical for major cellular and organismal processes. The physiological and pathologic patterns of c-myc transcription are programmed by a large number of cis-elements and transfactors (RNAs and proteins). These elements and factors receive inputs from a multitude of intracellular and extracellular pathways. Because c-myc regulation has customarily been dissected element by element and factor by factor, it has been difficult to appreciate how the c-myc promoter and regulatory sequences operate as a system. A full accounting of the regulation of c-myc transcription will require an understanding of the dynamic interplay of these factors and elements with one another, with chromatin, and with the changes in DNA structure and topology that are inevitably coupled with gene activity.
MYC function and the regulation of c-myc expression are inseparable. Because the proto-oncogene may also be the oncogene when abnormally expressed, setting the level of MYC is crucial.1,2 Everything regulates MYC. From stem cells3–5 and proliferation to senescence and cell death, MYC participates in almost every crucial decision of almost every cell. Hundreds of extracellular and intracellular signals, operating through an array of transcription factors, chromatin modifiers/remodelers, and regulatory RNAs (recruited to or synthesized at the c-myc locus), are brought to cis-elements vicinal to the promoter or strewn across a still poorly delineated chromosomal domain and are all somehow integrated to set the physiological levels of c-myc mRNA.1,2 The degree to which the complexity of the c-myc regulatory circuitry confers robustness, on one hand, or makes the system brittle and vulnerable to pathologic deregulation, on the other, will remain a matter of speculation without quantitative models6–8 and theoretical insights to explain or at least rationalize c-myc regulation.
In some situations, MYC is essential, and its levels must be precisely set. Because MYC drives promoter escape and the transcription of most active genes, even small changes in MYC levels may have a global impact on the cell.5,9–11 Cell cycle length and organism size within a species scale inversely and directly with MYC levels, respectively.12,13 In Drosophila developmental compartments containing cells making normal and twice-normal dMYC or cells haploinsufficient and normal for dMYC, the cells making less dMYC in each case are induced to apoptotic death, whereas the winners of this supercompetition take over the compartment and complete morphogenesis.6,7 When cocultured, Drosophila S2 cells with a metallothionein promoter-driven dMYC transgene coexist with their nontransgenic counterparts until metal is added to trigger transgene expression.8 Induction of the transgene for 3 hours, the shortest treatment interval tested, suffices to elicit supercompetition in tissue culture. Thus, temporal and quantitative fluctuations of MYC levels are associated with cellular unrest and turnover; after a prolonged period of noisy MYC expression, those cells making the most MYC take over the population. This outcome seems likely to generate a preneoplastic field primed for progression through multistep carcinogenesis.
In other circumstances MYC is entirely dispensable. Deletion of a conditional c-myc by liver- or intestine-specific Cre-recombinase in adult tissue has surprisingly little effect despite the requirement for steady-state proliferation in the intestine14,15; moreover, the c-myc knockout hepatocytes still proliferate in response to injury. Hence, MYC is not a mandatory component of the cell’s proliferation machinery. Human fibroblasts haploinsufficient for c-myc are refractory to immortalization by telomerase.16 In Drosophila, too, a surprising amount of development may occur in the absence of MYC; for example, following deletion of dMYC in the wing imaginal disc, morphogenesis yields a well-formed but diminutive structure.17 Small differences in MYC levels yield large consequences.
Tumors employ a variety of devices to increase MYC expression, including translocations that bring in strong enhancers (Burkitt lymphoma), amplification (colon cancer), viral insertions (human papillomavirus cervical cancer), mutation of cis-elements, mRNA stabilization, and protein stabilization via Ras.18–21
c-myc mRNAs are spliced from primary transcripts containing 3 exons (Figure 1). MYC protein is encoded in exons 2 and 3, although a minor MYC protein variant initiates at a CUG-leucine codon at the end of exon 1. c-myc primary transcripts mostly initiate at the 2 major promoters, P1 and P2, accounting for approximately 25% and 75% of c-myc mRNA, respectively. A lesser amount of transcription initiates at multiple minor promoters; the physiological roles and significance of these transcripts are not known. Pathologically, some Burkitt lymphoma translocations decapitate P1 and P2 from MYC and juxtapose the immunoglobulin enhancer with the minor P3 promoter in intron 1, just upstream of exon 2.1,2 This juxtaposition activates P3. P0 is a minor promoter that initiates several hundred bases upstream of P1. As with many genes, genomic methods of RNA analysis exhibit a menagerie of species.1,18,19,22 Which of these molecules may prove to be important biological species and which may turn out to be irrelevant in vivo transcripts is a matter for future research.
In addition to mRNAs, antisense transcripts in the vicinity of exon 1 and a variety of short RNAs are derived from the MYC locus. Although transfection studies have indicated a potential regulatory capability for these molecules in tissue culture, the in vivo cellular and organismal significance of these molecules is not known.23,24 The case for the regulation of c-myc mRNA levels by microRNAs is more clear.25,26
At the earliest stages of development, embryonic stem cells maintain high levels of MYC. c-myc expression has long been known to support proliferation and to antagonize cellular differentiation—properties that help to sustain physiological function in stem cells but exacerbate pathology in cancer. Along with Nanog, Sox2, Oct4, and Klf4, MYC is among the cadre of transcription factors capable of converting fibroblasts into induced pleuripotent stem cells.27 The transfactors and the binding sites that maintain the expression of c-myc in embryonic stem cells have not been enumerated (in contrast, the targets of MYC have been better studied). Later during embryogenesis, at various stages of development, high c-myc expression ensures the proliferative potential of precursor populations until c-myc downregulation licenses terminal differentiation. The developmental signals that elicit the induction or downregulation of MYC at later stages include Wnts, Notch, TGF-beta, and BMPs.1,2 These signals are conveyed to Tcf, Cbf, Smad, and many other families of transcription factors.
Although c-myc is broadly expressed in developing tissues, constitutive and globally expressed transcription factors alone may be insufficient to generate and maintain this pattern. For example, although MYC is found in all regions of the developing nervous system, loss of a single factor, CNBP, in the forebrain is enough to eliminate MYC from just those regions where CNBP is expressed. In immediately adjacent areas that do not express CNBP, MYC is unperturbed when CNBP is lost.28 Thus, general MYC expression may be pseudo-constitutive; its broad expression may be a mosaic where the expression within each tile is controlled by a customized set of transcription and chromatin factors.
In mature tissues, c-myc expression is modulated by growth factors such as PDGF and EGF, interferons, interleukins, the NF kappa B pathway, retinoic acid, vitamin D, and estrogen. These signals are conducted to members of the Stat, ets, NF kappa B, RAR, VDR, and ER. In addition, the E2Fs, Sp1-3, Rfx’s, AP1/fos-jun, octamer, and so on, have all been reported to regulate MYC expression.1,2 The binding sites for the transcription factors that regulate c-myc are often poor matches with their optimal consensus sequences or are noncanonical sites altogether.1
Many of the transcription factors that regulate c-myc bind to multiple elements scattered through hundreds of kilobases. Perhaps the best example of this is TCF. Through the use of ChIP (including ChIP-chip and ChIP-Seq) and 3-C (chromatin conformation capture), TCF has been found to bind at multiple sites—ranging from 330,000 base pairs upstream of the major start site to 30,000 base pairs downstream—and it is generally agreed on the basis of 3C experiments that TCF that is bound at these various sites loops to the promoter to stimulate c-myc expression.29–34 The degree to which these individual sites are redundant or are components of independently acting cis-regulatory modules (CRMs) is not known.35 The biological relevance of these sites is attested to by the presence of a colon cancer–associated SNP, 450 kb upstream of the c-myc transcription start site, that converts a weaker TCF4-binding site into a stronger one. TCF binding is enriched at the cancer-associated SNP in vivo and in vitro, and the TCF bound at this site loops to the c-myc promoter as demonstrated with 3C. Moreover, in a cell line whose 2 c-myc alleles are distinguishable by a SNP within the transcribed region, the c-myc allele in cis with the high-binding TCF-SNP is expressed at appropriately higher levels than those of its low-binding partner.32 It is difficult to understand how this polymorphism at this one remote element overrides or modifies the inputs of the other TCF4-binding elements. One possibility is that each binding site resides within a different CRM so that a colonic CRM would have maximal influence in the colon but not elsewhere. In this scheme, CRM selectivity could be achieved in various ways. At one extreme, only active CRMs would be loaded with transcription factors, whereas inactive CRMs would be unbound; at the other extreme, all CRMs would be fully loaded with transcription factors, but CRM looping to the promoter would be regulated so that only particular CRMs would have access to the promoter and the transcription machinery. Still another possibility is that all the TCF-binding sites are redundant; when TCF levels are high, all the binding sites are occupied. When TCF levels are low, only the highest-affinity binding sites would be filled, and so the tight-binding SNP-associated site would drive its c-myc allele in cis under conditions where its vacant partner allele would not be TCF-driven. As such, lingering low-level c-myc transcription would sustain the cancer diathesis.
With regulatory elements so widely distributed, it is not surprising that the full spectrum of c-myc regulation has been difficult to recapitulate in transgenic or transfection systems. The presence of multiple CTCF sites around c-myc and the utilization of alternative chromatin remodeling complexes suggest that proper organization, folding, and looping of the region are important for normal regulation, although as discussed below, MYC expression is surprisingly resistant to inactivation when the CTCF-binding sites and other binding sites are removed.36–38
On top of the myriad conventional transcription factors regulating MYC by binding with cognate elements in a conventional double helix (i.e., most members of larger families), a curious collection of unconventional DNA-binding, or nucleic acid–binding, proteins recognizing non–B DNAs has been proposed to regulate c-myc.1,39–42 The roles of these factors and non–B DNA structures may prove to be critical components to tune the performance and response of the system to environmental, physiological, and pathological stresses. The non–B DNA structures proposed to participate in c-myc regulation include single-stranded bubbles—loops of unpaired bases—Z-DNA, quadruplex, and triplex.1,39–42 Of course, the transcription bubbles found at transcription start sites and promoter proximal regions owing to open complexes and paused RNA polymerases are also foci of non–B DNA. To exist on a biologically relevant timescale, any non–B DNA must successfully compete with standard duplex to form a metastable, if not stable, conformation. So the problem becomes how to selectively destabilize the double helix while stabilizing the alternative structures. In vivo all non–B DNA segments must coexist with normal right-handed B-DNA, so there are only limited intranuclear options to shift the equilibrium to the alternative structures. Although ionic conditions, DNA concentration, pH, and temperature may be each be manipulated in vitro (often within narrow windows) to stabilize particular DNA conformations, it is difficult to imagine how these same parameters might be differentially adjusted at the level of individual chromatin domains, genes, or particular cis-elements. Yet, a variety of single-molecule experiments have proven that DNA structures are sensitive to applied mechanical forces. Of the 2 modes of applying force in single-molecule experiments—namely, stretching and twisting—twisting is likely to be the more prevalent in vivo because all DNA transactions associated with unwinding of the double helix generate torsion that is then transmitted through the DNA fiber, partitioning between twist and writhe.43–45 If chromatin is tethered to nuclear structures or filaments, then in principle, stretching and bending forces could be focally applied to DNA fibers and could locally alter DNA conformation. Because almost all non–B DNA is associated with persistent segments or foci of unpaired bases, unwinding of the double helix in vivo would assist DNA melting to accelerate the formation and stabilization of non–B DNA.
The untwisting supercoiling in mammalian DNA is partitioned into 2 types: restrained and unrestrained. Restrained supercoils arise when DNA is tightly wrapped around nucleosomes. Because DNA is a stiff polymer, considerable potential energy is stored in these spools; but for this energy to do work, this restrained DNA must be liberated from the nucleosome. The DNA between the nucleosomes is more accommodating to torque, temporally and spatially. The response of free DNA to torque generates unrestrained supercoils. When torque rises to high-enough levels, the double helix buckles via strand separation that relieves the torsional stress and allows the separated strands to adopt alternative DNA structures.43–45 Because topoisomerases drain unrestrained supercoils from DNA fibers, the existence and potential significance of unrestrained supercoils to genetic activity in mammalian cells has been debated, as has the extent, if not the existence, of non–B DNA structures. But during transcription, long segments of DNA are screwed through the active site of RNA polymerase, dynamically generating and transmitting torque through DNA fibers.43–45 Torque-driven changes in DNA structure may thus serve as a mechanosensor measuring the intensity of nearby genetic processes in real time. In c-myc, the far upstream element (FUSE) is just such a mechanosensor.46
Melting is the first structural transition that occurs in response to unwinding torque. Melting at any sequence within a closed topological domain lessens the stress everywhere within that domain, making it more difficult for other sequences to melt.47,48 Thus, at equilibrium, supercoil-induced melting is not simply a function of local sequence but a function of the meltability of all sequences within a domain; the same sequence that melts at a given level of supercoiling in one context may be rock solid in another. Dynamically, it may be a different story. Torque is highest near the torque generator (most often, RNA polymerase); then, it decays as it is transmitted via twist or writhe strain to less-stressed regions.43–45,49 In an open domain, the conformational response of a sequence depends on its location relative to the torque source and on the sort of DNA interposed between the source and the responding sequence; sequences further upstream than FUSE cannot absorb the supercoiling needed for FUSE melting, whereas downstream sequences can in a dynamic situation. FUSE melting is important because it exposes the binding site for the single-strand sequence-specific FUSE binding protein family (FBP1–3) and then for the FBP interacting repressor (FIR).50
Despite possessing powerful activation domains that act through TFIIH to help to advance and release c-myc’s paused RNA polymerase, FBP1 and FBP3 are excluded from the c-myc promoter until ongoing transcription generates enough dynamic supercoiling to surmount the threshold needed to melt FUSE. Thus, the FBPs and FIR act at later stages of the c-myc initiation cycle to program a pulse of expression. As with most c-myc cis-elements and transfactors, the FBP–FUSE system is not essential for c-myc expression. In the absence of the FBPs and FIR or with defective TFIIH, c-myc expression proceeds but is not closely supervised. Only the promoter proximal FUSE has been intensively investigated for c-myc. Whether there are additional FBP-binding sites in the chromosomal region of c-myc and, if so, what functional significance these sites have remains to be investigated.
Z-DNA is a left-handed double helix formed from repeating, alternating purine–pyrimidine (Pu–Py)n dinucleotides—with (CA)n–(GT)n among the most preferred—and the helical repeat is 12 base pairs corresponding to a pitch of 4.6 nm. To accommodate the right- to left-handed reversal in winding, a short B–Z junction includes a nidus of unpairing where the bases project into the solvent. Whereas Z-DNA is stiffer than B-DNA, the focal base unpairing at the junction may serve as joint augmenting, bending and twisting the Z-DNA segment relative to the embracing B-DNA.51,52 Because the conversion of one turn of B-DNA to one turn of Z-DNA absorbs two negative supercoils, it is the most efficient structural transition for accommodating the torsional stress associated with unwinding. The proteins thus far characterized that bind to Z-DNA in vitro recognize the geometry of the phosphodiester backbone but are insensitive to the specific sequence of the Z-DNA. Whether Z-DNA-binding proteins are equipped with the molecular tools to regulate genes directly is an open question. Other than for nucleic acid binding, these proteins have not been extensively interrogated for function: Do they possess activator or repression domains? Do they interact with the general transcription machinery? with chromatin-remodeling complexes? with chromatin-modifying complexes?
In the human c-myc gene, antibodies against Z-DNA detect 2 upstream regions as well as a segment near the intron 1–exon 2 junction.41 The reactivity of the antibody with these 3 regions depends on ongoing transcription or replication and is affected by topoisomerase inhibition. Neither cis-element activity nor protein binding for any of these sequences has been reported. So whether the structural change at the Z-DNA-forming regions of c-myc represents a physiologically significant transition or a bona fide but physiologically irrelevant in vivo event has not been established.
Perhaps, the most complex region of the c-myc gene is the segment that includes DNase I hypersensitivity site II2, also called the NHE (nuclease hypersensitive element), which comprises 4 direct imperfect repeats of the sequence CCCTCCCCA on the nontemplate strand and a fifth, more promoter proximal repeat separated from the others by 7 base pairs. The CT-NHE1 region is always nucleosome free, a feature that undoubtedly contributes to its nuclease hypersensitivity.40 Besides adopting B-DNA, this region in vitro has been reported to adopt triplex and several overlapping G-quadruplex structures on the purine-rich strand, with I-motif forming on the pyrimidine-rich strand. Other structures, such as slipped mismatch, may occur here. G-quadruplex is assembled from intrastrand runs of 3 or more guanosines, looping back and forth in several alternative arrangements that stack squares of 4 Gs, which hydrogen bond to one another in a plane, with each G hydrogen bonding to one partner using its Watson-Crick functional groups and to another partner using Hoogstein base pairing. Cations along the central axis of the stack are held within an electronegative cage. Potassium, the predominant intracellular cation, is favored over sodium within this cage. The individual runs of Gs may all be parallel, or they may run in alternative arrangements of antiparallel pairs. The backbones of the stacks are connected by loops of variable length (a minimum of 3). The complementary pyrimidine-rich strand of quadruplex may also minimize its conformational energy by forming an I-motif with interdigitated bases. This structure is weaker, reflecting the smaller surfaces of pyrimidines available for stacking. Although quadruplex structure is stable when folding from a single strand, it must compete for its existence with B-DNA. So the critical energetic parameter here is ΔΔG, the difference between the energies associated with the formation of B-DNA and those associated with the quadruplex from the purine- and pyrimidine-rich complementary strands in random coil. Note that when quadruplex forms in RNA, it does not have to compete with a double helix but only against a variety of alternative structures likely to be less energetically favored.
The in vivo existence and significance of G-quadruplex DNA has been a matter of conjecture owing to a paucity of in vivo studies and methods to definitively discriminate among alternative DNA structures. The development of chemical probes to interrogate quadruplex formation is an active area of investigation. Several G-quadruplex-binding compounds have been developed, most of which act through porphyrin-like rings that cap the guanosine stacks.39,40 Notwithstanding the objection that such capping might drive the formation of otherwise unstable G-quadruplexes, these agents have the potential to become valuable tools to study DNA in vivo, as well as therapeutic agents.
What is the significance of the various structures of the CT-NHE1 region? First it is possible that some structures directly influence the propagation of supercoiling to the FUSE. Torsional stress partitions between twist (how tightly one strand of DNA is entwined around its partner, as reflected by the pitch and by the number of residues/ helical turn) and writhe (undulation of the helical axis in 3 dimensions). Recent results and calculations predict that a nidus of alternative structure upstream of and near the transcription start site might promote energetically efficient extrusion of an interwound spur of plectonemic DNA in response to dynamically applied torque.53,54 As transcription proceeds, this spur would grow, acting as a capacitor storing strain as writhe, until redistributed to other sites. Note that the CT-NHE1 region is always nucleosome free, simplifying these polymer dynamics.1 Most of the regulatory significance of the CT-NHE1 region is likely to be attributable to the proteins and regulatory complexes that it binds.1,2,39,40 Among the proteins reported to bind here are hnRNP A1, A2, and B1, hnRNP K, NM23-H2/nucleoside diphosphate kinase B, CNBP, NSEP-1, Sp1, Sp3, MAZi, THZif-1, and c-MYB. Sp1, Sp3, and MYB are card-carrying duplex-binding gene regulatory factors, and converting B-DNA to other structures reversibly inactivates the binding sites of these factors. Most other proteins bind non–B DNAs with varying degrees of sequence specificity; by stabilizing alternative DNA conformations, they antagonize the action of the duplex binders. hnRNP K has been repeatedly identified as a transcription factor and has been shown to interact with the general transcription machinery.1 Nucleolin has emerged as the candidate mostly likely to be a bona fide in vivo quadruplex-binding and regulatory protein here.55 CNBP and nm23 have been proposed to be nucleic acid chaperones that facilitate structural transitions of the DNA.56,57 The molecular mechanisms and structural hardware employed by the other proteins to modify gene expression remain largely unstudied.
To functionally dissect the roles of individual factors and cis-elements, it has been necessary to exploit transfected reporters bearing deleted or mutated sequences, as well as overexpress or knockdown individual transcription factors. Although these sorts of experiments have greatly contributed to our understanding of transcription and gene expression, this approach is subject to certain limitations, especially when analyzing a complex regulatory system such as the c-myc promoter. The factor-by-factor analysis has led to a confusing, if encyclopedic, literature with little illumination of the operation of the system as a whole. Certain problems have limited our ability to dissect complex genes such as c-myc—for example, redundancy. Scattered and repeated cis-elements and families of related transcription factors whose members are often at least compensatory, if not flat out redundant, may mask an important role for particular elements and factors. Many experimental schemes to assess the biological importance of each component have assumed singular and independent activity for each cis- and transcomponent of the c-myc promoter. But to assign the relative importance of the regulatory sequences and transfactors for the regulation of c-myc is almost a meaningless question, a bit like asking what is the most important part of a car. The same brakes that are lifesaving as the car careens downhill are irrelevant without the keys. The car functions as a dynamic system assembled from submodules, the functions of which are customized for particular situations. Because many submodules share identical or related parts, their specialized features arise as emergent properties of the assembly and may not be understood in terms of the features of individual components. Although protein complexes may display catalytic and regulatory properties (e.g., cooperativity) that cannot arise in the separated subunits, whether and how cross-dependence and emergent properties arise from sets or subsets of cis-elements in a vast complex promoter has been less well evaluated. The standard factor-by-factor, element-by-element analysis of regulatory sequences may not be an efficient strategy to elucidate the rules governing the expression of genes such as c-myc.
How can we conceptually organize the complex biology of the c-myc promoter? Many of the features of the c-myc promoter can be rationalized as a system to control noise—that is, the cell-to-cell variation in MYC levels. The architecture of some transcription networks has evolved in some cases to secure a uniform and reliable response—for example, in gut development in Caenorhabditis elegans.58 Considering that many observations indicate that small differences in MYC levels yield quantitative and qualitative changes in cell, tissue, or organismal biology, physiology, and pathology, it would be surprising if MYC levels were permitted to fluctuate widely. Requiring several factor-accelerated steps to ratchet the transcription machinery from initiation through promoter escape reduces the intrinsic noise—stochastic fluctuations—when the output of the promoter is low59–61 (Figure 2). By responding to many signals via multiple transcription factors dynamically binding and unbinding at c-myc regulatory sequences, the fluctuations in one pathway (extrinsic noise60 propagated onto the c-myc promoter) are offset by the fluctuations in another, thereby stabilizing the output at steady state. A single signal has to be strong and sustained to rise above the fluctuating noise, or several separate and less-intense signals concurrently received might act in concert to promote c-myc transcription (Figure 3). Once transcription achieves sufficient intensity to melt FUSE, then the FBP–FIR–FUSE system may override the input of other factors to program a pulse of expression. Thus c-myc transcription is regulated in stages and layers.
This scheme is robust. For example, the promoter will work without FBP–FIR–FUSE; it just will not work as well, and expression would be expected to become noisy. The noise in c-myc expression is not a problem in the majority of cells in a tissue. But because cancer is a rare disease at the cellular level, despite its prevalence at the organismal level, the safeguards that have evolved to constrain MYC expression must operate on even the outliers of the population. The simplest scheme to achieve this fine-tuning is to superimpose separate regulatory modules. No one transfactor is critical for the c-myc promoter in general, although in special situations, physiological levels of MYC may depend on high-intensity signaling through a single pathway. Because of redundancy and parallel inputs by compensatory factors, no one cis-element is critical either. Mutation and deletion analysis of the c-myc promoter using stable transfections has shown that promoter was difficult to inactivate.38,62
Although the c-myc promoter is resistant to full inactivation, it is exquisitely sensitive to genetic disturbances, pathologic and experimental. In Burkitt lymphoma, c-myc expression survives the many t(8:14) translocations that decapitate the P1 and P2 promoters from the body of the gene, but it is also deregulated by translocations hundreds of kilobases away in t(2:8) and t(8:14) translocations. An exciting new approach for the analysis of c-myc is the targeting of human chromosome 8 introduced into the hyperrecombinogenic chicken B-cell line DT40 by cell fusion. These targeted chromosomes can then be introduced into mouse embryo fibroblasts, also by cell fusion, yielding a precisely altered c-myc gene in its proper chromosome 8 context (albeit in a heterologous species).38 Introducing small deletions removing the CTCF sites at P2 and/or situated at HS1, as well as a larger deletion between –1.5 and –3.0 kb upstream of the transcription start site, reduced but did not eliminate c-myc expression at steady state. Serum starvation of these cells followed by repletion effectively induced c-myc mRNA but modulated the temporal profile of the engineered alleles relative to the normal allele. Note that these alleles were not tested across a spectrum of physiological challenges, so larger deficits in c-myc regulation may occur under other conditions in these cells.
Reverse engineering the c-myc promoter promises to highlight how different layers of regulation are integrated to determine the final output of the gene. To understand this integration, it will be important to monitor expression in single cells to visualize and measure the noise profile of c-myc under a variety of physiological and pathological conditions. Embedded within the population statistics, transcriptional noise, and expression coherence between single cells is information concerning alternate states (i.e., bistability), alternative biological programs, and details of molecular mechanisms. Understanding the complexity of the c-myc promoter may highlight where and when to intervene biologically or pharmacologically and how to work around malfunctioning switches in a faltering transcription network.
This research received no specific grant from any funding agency in the public, private, or not-for-profit sectors.
The author declares no conflicts of interest with respect to the publication of this article.