Introductory Chapter: Gene Expression in Eukaryotic Cells

1. Introduction

“Central Dogma” explains how information from genes to proteins flows. Genes should be transcribed into messenger ribonucleic acids (mRNAs) in nuclei, then they are processed and delivered to the cytoplasm where they are translated into polypeptides (proteins). We, molecular biologists, know that gene expression in mammalian cells is controlled at multiple stages. First, gene expression is epigenetically regulated by chromatin structures, which depend on deoxyribonucleic acid (DNA) methylation and histone modifications. Then, transcription initiation, elongation, and termination occur, and RNA could be maturated in nuclei. Additionally, non-coding RNAs (ncRNAs), including miRNAs, affect gene expression. Moreover, loop structures of DNAs also play roles in gene expression. One of the recent striking topics is the identification of extrachromosomal circular DNAs (eccDNAs), and R-loop formation that is mediated by the interaction between DNAs and RNAs. In summary, the regulation of gene expression is a very much complicated system. In this chapter, I would review how gene expression controlling systems in mammalian cells are presently understood. I hope that we would be inspired to think over essential problems to be dissolved toward the progress of medical sciences.

2. Gene expression, which is regulated by multiple steps

It has been widely known how general transcription factors (GTFs) execute RNA pol II dependent transcription in mammalian cells [1]. I have ever reviewed transcription control systems in mammalian cells, especially focusing on the possibility of the application of transcription-regulating mechanisms on gene therapy [2, 3]. This time, we would challenge the most fundamental problems that should be addressed before discussing the practical use of the transcription system. Generally, gene expression is defined as the producing rate of the mature mRNAs that are to be utilized for the translation of polypeptides. Transcription begins with initiation, elongation, and ends with the termination process. In eukaryotes, the premature RNAs are to be processed by splicing [4], 5′-cap modification [5, 6], and polyadenylation or poly A tail addition [7]. The matured RNAs are transported to the cytoplasm to be used for translation. RNAs are unstable and they are easily degraded by ribonucleases (RNases), which are ubiquitously present in all kinds of cells. In human, it has been known that at least 13 belong to hRNase A superfamily [8]. Some of them play roles in the host defense system [9], and others are required for host and mitochondrial DNA replication [10, 11, 12]. Thus, as same as other biologically synthesized polymers, amounts of matured RNAs are appropriately regulated by the balance between synthesis and degradation.

3. Loop Structures that regulate transcription

Generally, it is thought that gene expression is considerably regulated by the initiation step that is dependent on sequence-specific TFs [2, 3]. The loop structure formation by the interaction between enhancer and core promoter [13, 14, 15, 16] is thought to be essential for GTFs to start transcription in the right direction. The enhancer-promoter loop can be made by two double-stranded DNAs (dsDNA) and the most 5′-upstream RNA strand. The formation of chromatin loop clusters may be mediated by specific proteins, including CCCTC-binding factor (CTCF) [17]. The loop formation between dsDNAs might be associated with chromosome-wide spreading gene-silencing mechanism as that has been well studied for Xist [18]. The possible molecular model is that RNA strands may play a role as bridges between DNAs. On the other hand, R-loop structure, which enables the recycling of TFs and RNA pol II, is made by communication between transcription initiation and termination sites [19, 20, 21, 22]. The R-loop, which is composed of separated complementary DNA strands or “transcription bubble” and elongating RNA, could induce antisense transcription at enhancer elements, and transcription initiation and termination sites [23]. Thus, the DNA-RNA triplet R-loop can also play an essential role in the bidirectional transcription on eukaryotic chromosomes. The R-loop resembles to the replication bubble, which is made by separated complementary strands and a short RNA that is required for the leading strand synthesis. Therefore, the generation of R-loops needs to be accurately controlled. If the dysregulation occurred, it may lead to conflicts between transcription and replication machinery, which will cause DNA damage and cell death. The problem might be dissolved by several key regulators, including ATR, CHEK1, and BRD4 [24, 25]. The promoter regions of eukaryotic genes have been mainly studied by the reporter assay system with plasmid vectors that express such as luciferase. However, usually, it has not been paid attention to the 3′-untranslated regions (3′-UTRs). It has been presumed that 3′-UTRs play important roles in the regulation of stability, localization of transcripts, and translation [26, 27, 28]. If the R-loop has a considerable effect on the production of immature transcripts, it should not be ignored.

4. Transcriptional direction might be epigenetically regulated

Epigenetic regulation is mainly executed by DNA methylation and histone modification [29] that control chromatin structure to regulate genomic imprinting [30] and cellular senescence [31]. The direction of transcription might be dependent on DNA methylation, which is regulated by DNA methyltransferases (DNMTs) and the Ten-eleven-translocation (TET) enzymes [29, 32, 33]. CpG islands, which can be a methylation target [34], are commonly present at bidirectional transcription loci in human chromosomes [35, 36]. The bidirectional promoter regions have more GC-rich sequences but less TATA boxes than unidirectional promoters [37]. The majority (>80%) of CpGs in the human genome of the somatic cells are methylated, apart from actively transcribed regions, including promoters and enhancers [38]. However, because whole-genome methylome analyses identified differentially methylated regions (DMRs) in the human chromosomes [39], not all GC-rich sequences are the targets for methylation. Notably, specific TFs preferentially bind to the methylated CpGs [40]. ETS family protein PU.1 (SPI1) forms a complex with Dnmt3a/b to bring site-specific methylation to cause down-regulation of transcription [41]. The GC-box recognizing Sp1 can interact with DNMT1 in human cells [42]. Moreover, C/EBPα, Klf4, and Tfcp2l1 can affect Tet2 to demethylate specific promoters to induce pluripotency of cells [43]. In summary, site-specific DNA methylation/demethylation, modulating affinities with specific TFs, might determine which unidirectional or bidirectional transcription would be preferred.

5. Nutrients and metabolites dependent gene expression

Prokaryotic lactose operon system has been generally known. Nutrients or metabolites must be monitored to control transcription adequately in mammalian cells [44]. Glucose regulates the transcription of genes that encode lipogenesis-associated proteins through activation of the LXR (NR1H) factors [45]. HDL metabolism functioning protein-encoding genes are induced by glucose [46]. A glucose derivative molecule, 2-deoxy-D-glucose (2-DG) upregulates promoter activities of the TERT and WRN genes in HeLa S3 cells [47].

Fatty acids can affect transcription mediated by PPARs [48], SREBP-1 [49], and other TFs [50]. The n-butyrate (butyric acid), which is produced by gut bacteria, or sodium butyrate regulates gene expression in human cells [51]. The transcriptional regulation by butyrate has been explained by an inhibitory action on histone deacetylase [52] and increasing stabilities of mRNAs [53]. Notably, Sp1 that recognizes a GC-rich sequence [54] is hyper-acetylated by butyrate in human colon cells [55]. Therefore, transcription of specific genes, if their promoter contains a Sp1 binding element(s), could be affected by a lipophilic acid, which stops proliferation [56]. Other TFs, including ETS family ETV1 (ER81) [57], ETV4 (PEA3) [58], and p53 [59], are also activated by butyrate-induced signals.

Amino acids also regulate transcription. For example, glutamine responsive genes have been identified [60]. Leucine starvation induces promoter activity of the CHOP gene, encoding the CCAAT/enhancer binding protein, through binding of ATF-2 to an amino acid response element (AARE) [61]. Some amino acids are to be metabolized to acetyl-CoA and a methyl-group donor S-adenosylmethionine. They are the substrates for histone acetylation and DNA methylation, respectively. Tryptophan could be metabolized to a nicotinamide adenine dinucleotide (NAD⁺), which is also produced from niacin (nicotinic acid and nicotine amide) or vitamin B₃, is not only required for mitochondrial functions, but also for poly (ADP-ribosyl)ation that is catalyzed by poly(ADP-ribose) polymerases (PARPs) [62, 63]. It has been suggested that NAD⁺ regulates transcription in mammalian cells [64, 65]. In mice, the hepatic and cytosolic NADH/NAD⁺ ratio alters the circulating α-hydroxybutyrate level, which could be an early biomarker for diabetes [66, 67]. Decrease in NAD⁺ concentration has been suggested to be associated with aging or age-related diseases [68]. Notably, 2-DG and trans-resveratrol, which up-regulate NAD⁺/NADH ratio in HeLa S3 cells [69], can activate the human WRN and TERT promoters [47, 70].

6. Euchromatin and heterochromatin

Euchromatin and heterochromatin represent specific structures of eukaryotic chromosomes, which are transcriptionally active and inactive, respectively [71]. Generally, heterochromatin represents a state where chromosomes are attached to a nuclear membrane with nuclear lamina [72]. The formation of the heterochromatin is thought to be under epigenetic regulation, during both the development and aging processes of mammalian cells. Heterochromatin does not only affect transcription but also protects chromosomes from mechanical stresses [73, 74]. Relationships between heterochromatin and DNA-repair systems have been proposed [75]. The DNA-repair systems could dominantly work on the euchromatin, and that is enabled by the stabilization of chromosomes by heterochromatin structure. It might be controversial that DNA repair factors cause heterochromatinization, but PARP-1 and 2, which play roles in the DNA-repair system, can contribute to the maintenance of heterochromatin [76]. The telomeric region where PARP or tankyrase is located plays a role in the maintenance of the ends of chromosomes [77]. The shortening of telomeres may lead to severe chromosomal instability that accelerates cellular senescence and cancer generation [78], suggesting that PARP enzymes protect telomeres by heterochromatinization. Moreover, the PARP modulates chromatin structure when it functions at centromeres [79]. Overall, PARP-dependent DNA repair systems are not only required for the conservation of nucleotide sequences of functional proteins but also for the maintenance of chromosomal structures that are constructed by specific or repetitive sequences. That would partly explain the reason why telomeres and centromeres might have excluded translocations of protein-encoding genes and transposons. This might also explain why PARP-1 does not have a preference for specific DNA sequences, surely it can find DNA breakage to load poly(ADP-ribose) to chromatin-associating proteins, including histones [80] and p53 [81]. The introduction of poly(ADP-ribose) on TFs may suppress transcription [82] and the repair system will work well in the promoter regions. Activation of PARPs consumes NAD⁺ to synthesize poly(ADP-ribose), which is required for indicating the DNA damaging sites. That will cause a reduction in NAD⁺-dependent transcription of mitochondrial protein-encoding genes [65]. Taken together, poly(ADP-ribosyl)ation plays a role in keeping a balance between DNA-repair, energy production, and transcription, maintaining chromosomal structures.

7. Concluding remarks

“Gene Expression” can be discussed from many points of view. Because it includes many biological events that are executed by various proteins and RNAs. Although I have not reviewed the recent progress in studies of non-coding RNAs, they play essential roles in transcriptional regulation [83]. “Gene Expression” is regulated by many stresses that can modulate DNA structures, loop formations, and epigenetic states. Dysregulation of “Gene Expression” will cause aging-related diseases. Hopefully, artificial transcription controlling systems will be developed and clinically applied to cancer and neurodegenerative diseases in the future. Nevertheless, we have not yet reached a conclusion or even a hypothesis on how gene expression system has been established and how it developed through a long evolution process. All organisms, including prokaryotes, archaea, and prokaryotes, would not live without accurate execution of the transcription system. The exceptions are viruses that just utilize the infected host cell system. Among them, some retroviruses are unique in carrying oncogenes to cause cancer and lymphoma [84]. Their genes are RNAs to be reverse transcribed to DNAs, which can be integrated into host cell chromosomes. The composition of the genome is characteristic, having long terminal repeats (LTRs). Interestingly, many retrovirus-like elements or transposons, including LINEs and SINEs, have been suggested to regulate gene expression, by both transcriptional and post-transcriptional mechanisms [85]. Every protein-encoding gene has a transcription start and termination site. How have genes acquired promoters and terminators, which are present on the 5′-upstreams and 3′-downstreams, respectively? The loop structures and extrachromosomal circular DNAs (eccDNAs), which are frequently identified when DNA amplification occurs, might give us a hint [86].

The dsDNA loops are thought to be formed when DNA damage was induced [87] or when a rearrangement of genes occurs in immune cells [88]. Although it is a hypothesis, the generation of multiple DNA replication initiation sites in eukaryotic chromosomes suggests that linear chromosomes might have been evolved from the fusion of multiple circular chromosomes (Figure 1). To prove this hypothesis, hot spot junction sites, which are the same as tentative dsDNA break sites, should be identified (Figure 1C). Loop or circular DNAs are formed at the time when dsDNAs integrated in or released from chromosomes. Therefore, there are both chances to gain or lose DNAs. If it occurred at chromosomal crossover during meiosis, the acquired or lost DNA sequences would be inherited to descendants. Elucidation of the biological meanings of transposition and amplification of genes will answer the question of how genes acquired promoters and terminators through evolution.

Figure 1.

DNA replication of eukaryotic cells. (A) Multiple DNA replication initiation sites (yellow circles) are distributed all over chromosomes. (B) Leading strands (red) and lagging strands (blue) are synthesized from the initiation sites when S-phase started. (C) At the sites where opposite direction proceeding leading strands collide, the temporary generated dsDNA breaks would be generated to be ligated. Taken together, multiple circular DNAs, which are to be ligated with, might be generated at the end of the S-phase.