Explain the role of microRNA and RNA interference

What Are MicroRNAs?

Key properties:

Non-coding: miRNAs are transcribed from DNA but never translated into protein
Endogenous: Encoded in the organism's own genome (unlike viral siRNAs)
Sequence-specific: Each miRNA has a "seed region" (nucleotides 2-8) that determines which mRNAs it targets
Conserved: Many miRNAs are identical across distant species (e.g., let-7 in worms and humans)

The RNAi Mechanism: Step-by-Step Derivation

Let's derive how a miRNA silences a gene, starting from the DNA blueprint.

Step 1: Transcription of pri-miRNA

What happens: A primary miRNA (pri-miRNA) transcript is made by RNA Polymerase II from a miRNA gene. This is a long (~1000 nt) RNA with a hairpin structure.

Why this step? miRNA genes are scattered across the genome (often introns of other genes or intergenic regions). The cell treats them like regular genes initially—transcription creates the raw material.

\text{DNA (miRNA gene)} \xrightarrow{\text{RNA Pol II}} \text{pri-miRNA (long hairpin)}

Step 2: Processing to pre-miRNA (Nuclear Cropping)

What happens: An enzyme called Drosha (an RNase III endonuclease) cuts the pri-miRNA hairpin ~11 nucleotides from the base of the stem.

Why this step? The full pri-miRNA is too bulky to export from the nucleus. Drosha trims it to a ~70 nt precursor miRNA (pre-miRNA) with a 2-nucleotide 3' overhang (a molecular "barcode" for export).

\text{pri-miRNA} \xrightarrow{\text{Drosha + DGCR8}} \text{pre-miRNA (stemloop, ~70 nt)}

Key detail: Drosha works with a partner protein DGCR8 that recognizes the double-stranded stem.

Step 3: Export to Cytoplasm

What happens: Exportin-5 (a transport protein) recognizes the 3' overhang and shuttles pre-miRNA through nuclear pores into the cytoplasm.

Why this step? The protein synthesis machinery (ribosomes) is in the cytoplasm. The miRNA must reach its mRNA targets where translation occurs.

\text{pre-miRNA (nucleus)} \xrightarrow{\text{Exportin-5 + GTP}} \text{pre-miRNA (cytoplasm)}

Step 4: Dicer Cleavage

What happens: Dicer (another RNase III enzyme) cuts the pre-miRNA loop, leaving a ~22 bp double-stranded RNA duplex.

Why this step? Dicer measures exactly 22 nucleotides from the Drosha cut—this precision ensures the correct "seed region" is preserved. The loop is discarded as junk.

\text{pre-miRNA} \xrightarrow{\text{Dicer}} \text{miRNA duplex (22 bp)} + \text{loop fragment}

Derivation detail: Dicer has a PAZ domain (binds the 3' overhang) and two RNase III domains spaced ~22 nt apart. It's a molecular ruler.

Step 5: RISC Loading and Strand Selection

What happens: The miRNA duplex is loaded onto RISC (RNA-Induced Silencing Complex), a protein complex containing Argonaute (AGO) proteins. One strand (the guide strand) is kept; the other (passenger strand) is degraded.

Why this step? Only one strand is needed to find the target. The cell picks the strand with the less stable5' end (lower base-pairing energy) as the guide.

\text{miRNA duplex} + \text{AGO proteins} \rightarrow \text{RISC (mature miRNA loaded)} + \text{passenger strand (degraded)}

Selection rule:

\Delta G_{5'\text{-end, guide}}< \Delta G_{5'\text{-end, passenger}} \implies \text{guide selected}

Step 6: Target Recognition and Silencing

What happens: The mature miRNA in RISC scans mRNAs for complementary sequences (usually in the 3' UTR). Binding triggers one of two outcomes:

Perfect complementarity (rare in animals): AGO cleaves the mRNA like scissors → degradation
Imperfect match (typical): RISC blocks ribosome binding or recruits deadenylases → translation repression or slow decay

\text{RISC} + \text{target mRNA} \rightarrow \begin{cases} \text{mRNA cleavage} & \text{(perfect match)} \\ \text{translation block} & \text{(imperfect match)} \end{cases}

Why the 3' UTR? The ribosome reads 5'→3', but doesn't need to translate the 3' UTR. miRNAs bind there to avoid disrupting the coding sequence while still interfering.

Seed region rule: Nucleotides 2-8 of the miRNA must base-pair with the target. Positions 9-22 can have mismatches ("wobles").

Quantitative Impact: How Much Silencing?

A single miRNA typically reduces target protein levels by 30-80%—not a complete knockout. The math:

Let $P_0$ = protein synthesis rate without miRNA, $P$ = rate with miRNA, $k_{\text{on}}$ = RISC binding rate, $k_{\text{off}}$ = RISC unbinding rate.

At equilibrium, fraction of mRNA bound by RISC:

f_{\text{bound}} = \frac{[\text{RISC}] \cdot k_{\text{on}}}{[\text{RISC}] \cdot k_{\text{on}} + k_{\text{off}}}

If bound mRNA has translation efficiency $\epsilon$ (0 to 1), then:

\frac{P}{P_0} = 1 - f_{\text{bound}} + \epsilon \cdot f_{\text{bound}} = 1 - f_{\text{bound}}(1 - \epsilon)

Example: If 70% of mRNAs are bound ( $f_{\text{bound}} = 0.7$ ) and bound mRNAs are 90% silenced ( $\epsilon = 0.1$ ):

\frac{P}{P_0} = 1 - 0.7(1 - 0.1) = 1 - 0.63 = 0.37 \quad (63\% \text{ reduction})

Worked example Worked Example 1: let-7 in Development

Scenario: In C. elegans larvae, let-7 miRNA is absent. After the final larval stage, let-7 is transcribed. Its target is lin-41 mRNA (codes for a protein that blocks adult cell fates). Step-by-step:

Larva: No let-7 → lin-41 mRNA translated → LIN-41 protein accumulates → cells remain in larval program
Adult transition: let-7 gene transcribed → pri-let-7 → Drosha cuts → pre-let-7 → Exportin-5 → Dicer cuts → let-7 duplex → RISC
Silencing: let-7-RISC binds to seven sites in lin-41 3' UTR (imperfect matches, seed regions align) → ribosomes blocked → LIN-41 protein drops 80% → adult cell fates activated

Why each step?

Drosha/Dicer precision ensures the seed region (GAGUAG) is exactly positioned
Multiple binding sites amplify silencing (each RISC complex is independent, multiplicative effect)
Timing: let-7 is only made when the animal reaches a threshold size (hormone signal), so the switch is developmental

Worked example Worked Example 2: miR-122 in Hepatitis C

Scenario: Hepatitis C virus (HCV) RNA has two binding sites for miR-122 in its 5' UTR. Paradoxically, miR-122 stabilizes HCV RNA instead of silencing it.

Step-by-step:

HCV enters liver cell → viral RNA released
Liver cells express high miR-122 (70% of all miRNAs in hepatocytes)
miR-122-RISC binds HCV 5' UTR (near the ribosome entry site)
Instead of blocking translation: RISC binding protects the 5' end from degradation by Xrn1 exonuclease
HCV RNA half-life increases 10-fold → more viral proteins → more virions Why this step? Normally, cellular mRNAs have a 5' cap that blocks exonucleases. HCV RNA lacks a cap, so Xrn1 would chew it from the 5' end. miR-122-RISC acts as a "prosthetic cap." This is a viral hijack of the RNAi machinery.

Clinical relevance: Drugs that block miR-122 (like Miravirsen, antisense oligonucleotide) reduce HCV RNA by ~2 log₁₀ in patients.

Small Interfering RNAs (siRNAs): The Exogenous Cousins

Mechanism difference:

\begin{align} \text{dsRNA (viral)} &\xrightarrow{\text{Dicer only (no Drosha)}} \text{siRNA duplex} \\ &\xrightarrow{\text{RISC (same AGO)}} \text{perfect match} \rightarrow \text{mRNA cleavage} \end{align}

Why the cell cares: Viruses inject dsRNA during replication. The RNAi pathway evolved to chop up viral genomes.

Biological Roles: Why Does RNAi Exist?

Fine-tuning gene expression: Most miRNAs reduce their targets by 2-5 fold, not ON/OFF. This allows graded responses (e.g., miR-1 tunes heart contraction strength by adjusting calcium channel mRNA).
Noise buffering: If a transcription burst makes too much mRNA, miRNAs trim the excess. Math: If mRNA noise $\sigma_m^2$ is high, miRNA-mediated decay adds damping term:
$\frac{d[\text{mRNA}]}{dt} = k_{\text{transcription}} - k_{\text{decay}}[\text{mRNA}] - k_{\text{RNAi}}[\text{RISC}][\text{mRNA}]$
The $k_{\text{RNAi}}$ term flattens fluctuations.
Developmental switches: let-7, lin-4, miR-125 are ancient timers that trigger metamorphosis, differentiation, or cell cycle exit.
Antiviral defense: siRNAs destroy viral RNA. In plants and invertebrates, this is the primary immune system. Mammals evolved interferon, but RNAi still exists.
Transposon silencing: Transposons (jumping genes) are kept in check by piRNAs (another small RNA class) via RNAi-like mechanisms.

Common mistake Common Error: "miRNAs Delete Genes"

Wrong idea: "If a miRNA targets a gene, that gene is turned off like a knockout." Why it feels right: The term "silencing" sounds absolute. Cleaved mRNAs are destroyed.

The fix: Most animal miRNAs cause partial repression (30-80% reduction), not elimination. The target mRNA is still present and translated, just less. Think dimer switch, not light switch. Additionally, miRNAs often target multiple genes (one miRNA can have 100+ targets), so the effect is distributed. A knockout mutant (e.g., let-7 deletion) has dramatic phenotypes, but overexpressing let-7 only modestly reduces each target.

Proof: RNA-seq of cells overexpressing miR-200 shows target mRNAs drop to 40-60% of normal, not zero. Ribosome profiling shows some translation still occurs.

Formula Key Equations

Seed region binding energy (determines target specificity):

\Delta G_{\text{seed}} = \sum_{i=2}^{8} \Delta G_{\text{base-pair}, i} \quad(\text{kcal/mol})

Typical threshold: $\Delta G_{\text{seed}} < -5 \, \text{kcal/mol}$ for functional binding.

Target repression (assuming imperfect match, translational block):

\text{Protein output} = P_0 \left(1 - \frac{n \cdot [\text{RISC}]}{K_d +[\text{RISC}]}\right)

where $n$ = number of miRNA binding sites, $K_d$ = dissociation constant.

mRNA decay rate (with RNAi):

\frac{d[M]}{dt} = k_{\text{tx}} - \left(k_{\text{basal}} + k_{\text{RNAi}} \cdot f_{\text{bound}}\right)[M]

Steady-state mRNA level: $[M]_{\text{ss}} = \frac{k_{\text{tx}}}{k_{\text{basal}} + k_{\text{RNAi}} \cdot f_{\text{bound}}}$

Recall Feynman: Explain to a 12-Year-Old

Imagine your cells are like a factory making toys (proteins) from instruction manuals (mRNA). Sometimes the factory makes too many of one toy, which wastes materials. So the factory has tiny robots called miRNAs. Each robot has a scanner that looks for a specific barcode (the seed region) on certain manuals.

When a miRNA robot finds its matching manual, it puts a sticker on it that says "slow down" or "throw this away." The toy-making machines (ribosomes) see the sticker and either skip that manual or work slower. Over a few hours, the pile of unwanted toys shrinks.

Here's the cool part: the miRNA robots are made from the same DNA cookbook as everything else, but in a special way. First, a long piece of RNA is folded into a hairpin shape (like a bobby pin). An enzyme called Drosha cuts it shorter, then another enzyme called Dicer cuts it even shorter, making a tiny double-stranded piece. One strand becomes the robot's scanner, and it teams up with other proteins to form the RISC team.

The factory uses these robots to respond quickly to changes—if the environment says "stop making this toy," turning on a miRNA is faster than rewriting the DNA instructions. That's why miRNAs are so important for growing and adapting.

Seed region is positions 2-8 → "28 days later, the miRNA found its target" (zombie movie reference to remember the numbering).

Connections

DNA Methylation and Histone Modifications: Chromatin state determines which miRNA genes are transcribed
Transcription Factors: Many transcription factors regulate pri-miRNA expression (e.g., p53 induces miR-34)
mRNA Splicing: Some miRNAs are embedded in introns and released during splicing
Post-Translational Modifications: RNAi reduces mRNA, PTMs regulate existing protein—complementary strategies
Viral Replication: Many viruses encode anti-RNAi proteins (e.g., NS1 from influenza blocks Dicer)
Cancer Biology: Tumors often downregulate tumor-suppressor miRNAs (e.g., let-7 targets RAS oncogenes)
Epigenetics: DNA methylation can silence miRNA genes → loss of RNAi → disease

#flashcards/biology

What are the two key enzymes that process miRNA, and where does each act? :: Drosha (acts in the nucleus, cuts pri-miRNA to pre-miRNA) and Dicer (acts in the cytoplasm, cuts pre-miRNA to mature miRNA duplex).

What is the seed region of a miRNA, and why is it critical?

The seed region is nucleotides 2-8 of the mature miRNA. It must base-pair perfectly with the target mRNA's 3' UTR for RISC to bind and silence the target. Mismatches in the seed abolish silencing.

Distinguish miRNA from siRNA in origin and complementarity.

miRNAs are endogenous (encoded in the genome), processed from hairpin precursors, and have imperfect target complementarity (usually repress translation). siRNAs are exogenous (from dsRNA, viruses, or lab), have perfect complementarity, and trigger mRNA cleavage.

What complex loads the mature miRNA and executes gene silencing?

RISC (RNA-Induced Silencing Complex), which contains Argonaute (AGO) proteins as the catalytic component. AGO binds the guide strand and scans for complementary mRNA targets.

Why do most miRNAs bind the 3' UTR of mRNAs?

The 3' UTR is not translated, so miRNA binding there doesn't disrupt the coding sequence. Yet it's still close enough to the ribosome binding region to interfere with translation initiation or recruit deadenylases for mRNA decay.

In the let-7 example, what happens when let-7 is absent vs. present?

Absent: lin-41 mRNA is fully translated → LIN-41 protein keeps the cell in a larval state. Present: let-7-RISC binds lin-41 3' UTR → translation blocked → LIN-41 drops adult differentiation proceeds.

How does Drosha recognize pri-miRNA for cleavage?

Drosha partners with DGCR8, which recognizes the double-stranded stem of the hairpin structure. Drosha then cuts ~11 nucleotides from the base of the stem, leaving a 2-nt 3' overhang on the pre-miRNA.

What determines which strand of the miRNA duplex becomes the guide strand in RISC?

The strand with the less stable 5' end (lower base-pairing free energy) is selected as the guide. The other strand (passenger) is degraded.

Concept Map

Hinglish (regional understanding)

Intuition Hinglish mein samjho

Dekho beta, is note ka core idea samajhna bahut simple hai agar hum ise ek "volume knob" ki tarah socho. Cell ke andar jab DNA se RNA banta hai aur phir protein banta hai, toh usually hum sochte hain ki control sirf DNA level pe hota hai (transcription rokna). Par RNA interference ek alag trick hai — yeh mRNA banne ke baad usko silence kar deta hai, protein banne se pehle. Aur ye kaam karti hai chhoti si microRNA jo bas ~22 nucleotides lambi hoti hai, koi protein nahi banati, par kaam bahut important. Ye ek "address label" ki tarah cell ko batati hai ki kaunsi mRNA ko band karna hai.

Ab why it matters? Socho, agar cell ko kisi protein ki quantity thoda kam karni ho — poori tarah ON/OFF nahi, bas volume thoda dheema — toh DNA rewrite karna ya transcription factors badalna slow aur mehenga hota hai. miRNA is kaam ko fast aur precise tareeke se kar deti hai. Aur ek bonus ye ki ek hi miRNA hundreds of alag-alag mRNAs ko target kar sakti hai agar unke sequence milte-julte hain. Isliye ye cell ke liye ek smart, economical fine-tuning system ban jaata hai.

Mechanism yaad rakhne ke liye ek assembly line jaisa socho: pehle DNA se lambi pri-miRNA banti hai (RNA Pol II se), phir Drosha enzyme use trim karke pre-miRNA banata hai nucleus mein, Exportin-5 use cytoplasm mein bhejta hai, wahan Dicer usko cut karke 22 bp ka duplex banata hai, aur finally RISC complex (Argonaute proteins ke saath) usme se ek guide strand rakh leta hai jo target mRNA dhoondh ke silence kar deti hai. Har step ka apna logic hai — jaise Drosha isliye trim karta hai kyunki lambi RNA nucleus se bahar nahi ja sakti. Bas yehi flow yaad rakho, toh puri kahani clear ho jayegi.

Test yourself — Mutations & Gene Regulation

Connections