Chromatin, Epigenetics & Gene Regulation

Chromatin architecture

The nucleosome. About 147 bp of DNA wraps ~1.7 turns around a core octamer of histones (two each of H2A, H2B, H3 and H4). Lysine- and arginine-rich histones carry a positive charge that grips DNA's negatively charged phosphate backbone. H1 lies outside the core on linker DNA and draws nucleosomes into a tighter fiber.

Heterochromatin vs euchromatin. Heterochromatin is condensed, dark on electron microscopy and closed to RNA polymerase, so its genes are silent; examples are centromeres, telomeres and the inactive X (Barr body). Euchromatin is loosely packed, pale and transcriptionally active. Epigenetic marks, heritable changes in expression that leave the DNA sequence unchanged, decide which state a region adopts.

Epigenetic marks: DNA methylation and histone modification

DNA methylation. DNA methyltransferases methylate cytosine in CpG dinucleotides to form 5-methylcytosine: DNMT3A/3B write new patterns and DNMT1 copies them onto the daughter strand after replication. The methyl donor is S-adenosylmethionine (SAM), regenerated through methionine synthase (methyl-tetrahydrofolate, methylcobalamin). Methylated promoter CpG islands recruit methyl-CpG–binding proteins such as MeCP2, which bring in HDACs and silence the gene; MECP2 loss causes Rett syndrome (girls who regress after normal early development, with hand-wringing). TET dioxygenases begin demethylation by oxidizing 5-methylcytosine.

Histone modification. Histone acetyltransferases (HATs) acetylate lysine tails, neutralizing their charge and loosening DNA so transcription rises; histone deacetylases (HDACs) reverse this and repress. Histone methylation depends on the residue: H3K9me3 and H3K27me3 (the latter written by EZH2 of Polycomb repressive complex 2) mark silent chromatin, whereas H3K4me3 marks active promoters.

Methylation in disease: fragile X and imprinting

Fragile X syndrome. FMR1 (Xq27.3) carries a CGG repeat in its 5′ untranslated region. Above 200 repeats the promoter CpG island is hypermethylated, so FMRP is not made. It is the most common inherited cause of intellectual disability, with a long face, large jaw and ears, post-pubertal macroorchidism and autistic features. Premutations (55–200 repeats) tend to expand through the mother and can cause late-onset tremor/ataxia (FXTAS) in men or primary ovarian insufficiency in women via toxic repeat RNA, unlike the toxic polyglutamine protein of CAG-repeat diseases such as Huntington.

Genomic imprinting. At imprinting control regions, methylation set during sperm or egg formation silences one parental allele; the marks are erased and reset in each generation's germ cells. Losing the only active copy causes disease: at 15q11–q13, paternal loss gives Prader-Willi and maternal UBE3A loss gives Angelman syndrome. Uniparental disomy mimics deletion (maternal UPD 15 → Prader-Willi; paternal UPD 15 → Angelman).

X-chromosome inactivation

Mechanism. Early in female embryogenesis each cell silences one X at random. The chosen X's inactivation center (Xq13) transcribes XIST, a long noncoding RNA that coats that chromosome in cis and recruits repressive histone marks and CpG methylation, condensing it into the Barr body at the nuclear membrane. The choice is inherited by all daughter cells, so every woman is a mosaic.

Counting Barr bodies. Barr bodies = number of X chromosomes minus one: none in 45,X or 46,XY, one in 46,XX and 47,XXY (Klinefelter), two in 47,XXX. Pseudoautosomal and other escape genes explain why a missing or extra X still alters the phenotype.

Skewed lyonization. If by chance most cells inactivate the X bearing the normal allele, a heterozygous woman expresses an X-linked recessive disease: manifesting carriers of Duchenne dystrophy (weakness, dilated cardiomyopathy), hemophilia A (low factor VIII, menorrhagia), G6PD deficiency or Fabry disease.

A symptomatic obligate carrier of an X-linked recessive disorder most often has skewed X-inactivation.

Prokaryotic regulation: the lac operon

Structure. In E. coli, lacZ (β-galactosidase), lacY (permease) and lacA (transacetylase) share one promoter and are read as a single polycistronic mRNA. The separate lacI gene is expressed constitutively and makes the repressor.

Negative control. Without lactose, LacI sits on the operator and blocks RNA polymerase. Entering lactose is partly converted by β-galactosidase to allolactose, which binds LacI and releases it.

Positive control. When glucose is scarce, cAMP rises and binds catabolite activator protein (CAP); CAP–cAMP binds upstream of the promoter and recruits RNA polymerase. Plentiful glucose keeps cAMP low, so transcription stays weak even without the repressor (catabolite repression).

Mutations: cis vs trans. Loss-of-function lacI mutants and operator mutants that cannot bind repressor are both constitutive. A normal lacI gene on a plasmid rescues the repressor mutant because the protein diffuses (acts in trans); an operator mutation acts only in cis and cannot be rescued. A repressor unable to bind allolactose keeps the operon off.

RNA silencing and nucleic acid–based drugs

miRNA vs siRNA. Endogenous miRNAs pair imperfectly (seed region) with the 3′ UTRs of many mRNAs, repressing translation and destabilizing transcripts. siRNAs, usually synthetic or viral, pair perfectly with one target that Argonaute-2 cleaves. Single-stranded antisense oligonucleotides (ASOs) either recruit RNase H to degrade RNA (inotersen for hereditary ATTR) or mask splicing signals (nusinersen; eteplirsen, exon 51 skipping in Duchenne).

Spinal muscular atrophy. Autosomal recessive SMN1 loss causes anterior horn cell degeneration (hypotonia, weakness, tongue fasciculations). SMN2 skips exon 7 in most transcripts, yielding unstable protein, and its copy number modifies severity. Intrathecal nusinersen blocks an intronic splicing silencer so SMN2 keeps exon 7; alternatives are onasemnogene abeparvovec (AAV9 SMN1 gene replacement) and oral risdiplam (splicing modifier).

Hypomethylating agents are myelosuppressive: counts often fall further in the first cycles, and MDS responses usually take several cycles.