Module 6

Network Dynamics & Mitophagy

Note on terminology.“Fusion” and “fission” in this module refer to membrane events — two mitochondria merging their lipid bilayers into one, or a single mitochondrion dividing into two. This is notthe nuclear-physics meaning (atoms joining or splitting with release of binding energy). Mitochondrial dynamics is a cell-biology process driven by dynamin- family GTPases; no nuclear reactions are involved.

Mitochondria form dynamic networks that continually fuse, divide, move along microtubules, and are selectively degraded. The balance between these processes — mitochondrial network dynamics — controls ATP production, quality control, apoptosis susceptibility, and differentiation. Defects in the core machinery (MFN1/2, OPA1, DRP1, PINK1, Parkin) cause Charcot-Marie-Tooth, dominant optic atrophy, encephalopathies, and early-onset Parkinson’s disease.

1. The Fusion Machinery

Fusion couples two mitochondria into one by sequentially merging their outer and inner membranes. Three dynamin-family GTPases execute the process:

MFN1 / MFN2 (mitofusin 1 and 2) on the outer membrane. Each is a large (86 kDa) GTPase with two transmembrane helices and a heptad-repeat coiled-coil domain. MFN1 on one mitochondrion trans-heterodimerises with MFN1 or MFN2 on the other via its HR2 coiled coil. GTP hydrolysis pulls the membranes together; membrane zippering follows. Cryo-EM (Nakamura 2022 Nat. Struct. Mol. Biol.) resolved the dimer architecture.
OPA1 on the inner membrane. OPA1 is expressed as 8 isoforms, each cleaved by OMA1 and YME1L into long (L-OPA1) and short (S-OPA1) forms. Fusion requires both forms plus an inner-membrane proton gradient; OMA1 sensitively depolarises OPA1 under stress, switching it off. OPA1 also maintains cristae structure via its mitochondrial-contact-site role.

Disease: MFN2 mutations cause Charcot-Marie-Tooth disease type 2A (axonal neuropathy), demonstrating the specific requirement for long-range axonal mitochondrial transport and fusion. OPA1 mutations cause autosomal- dominant optic atrophy (DOA) — visual loss from retinal ganglion cell mitochondrial failure.

2. The Fission Machinery

Mitochondrial fission is executed by DRP1(dynamin-related protein 1), a soluble cytosolic GTPase recruited to the outer mitochondrial membrane by receptor proteins: Fis1, MFF (mitochondrial fission factor), MiD49, and MiD51. DRP1 oligomerises into helical filaments encircling the constriction site; GTP hydrolysis tightens the helix and severs both membranes simultaneously.

\[ \text{DRP1}_{cytosol} \;\xrightarrow{\text{MFF/Fis1}}\; \text{OMM} \;\xrightarrow{\text{GTP hydrolysis}}\; \text{membrane scission} \]

Fission sites are marked by the ER: tubules of the endoplasmic reticulum wrap around the mitochondrion and pre-constrict it to ~140 nm diameter before DRP1 recruitment (Friedman 2011 Science). Actin filaments nucleated by INF2 at these sites assist the pre-constriction.

DRP1 activity is modulated by post-translational modifications: CDK1 phosphorylation at S616 activates DRP1 during mitosis (ensuring daughter-cell mitochondrial inheritance); PKA phosphorylation at S637 inhibits DRP1 during fasting/starvation (promoting fusion and preserving network). Calcineurin dephosphorylates S637 during Ca²⁺ spikes, enabling stress-induced fragmentation. DRP1 loss-of-function mutations cause lethal encephalopathy.

Simulation: Steady-State Network Morphology

Monte-Carlo simulation of fusion/fission events on a population of discrete mitochondria. Balanced rates give a steady-state size distribution; fusion- dominant collapses the population into few large objects; fission-dominant produces many small fragments.

Python

script.py63 lines

import numpy as np, matplotlib
matplotlib.use('Agg')
import matplotlib.pyplot as plt

rng = np.random.default_rng(0)
def simulate(k_fus, k_fis, steps=6000, n_init=200):
    sizes = [1]*n_init
    trace_frag = []
    for step in range(steps):
        if rng.random() < k_fus and len(sizes) > 1:
            i, j = rng.integers(0, len(sizes), 2)
            if i != j:
                sizes[i] += sizes[j]
                sizes.pop(j)
        if rng.random() < k_fis:
            i = rng.integers(0, len(sizes))
            if sizes[i] > 1:
                new = rng.integers(1, sizes[i])
                sizes.append(sizes[i] - new)
                sizes[i] = new
        if step % 50 == 0:
            trace_frag.append(len(sizes))
    return sizes, trace_frag

bal, tr_bal   = simulate(0.30, 0.30)
fus_h, tr_fus = simulate(0.55, 0.08)
fis_h, tr_fis = simulate(0.08, 0.55)

fig, axes = plt.subplots(1, 2, figsize=(14, 5), facecolor='#0a0a1a')
for ax in axes:
    ax.set_facecolor('#111827'); ax.tick_params(colors='#fdba74')
    for s in ax.spines.values(): s.set_color('#78350f')

bins = np.arange(1, 30)
axes[0].hist(bal, bins=bins, alpha=0.5, color='#fb923c',
             edgecolor='#fdba74', label='Balanced')
axes[0].hist(fus_h, bins=bins, alpha=0.5, color='#f87171',
             edgecolor='#fdba74', label='Fusion-dominant')
axes[0].hist(fis_h, bins=bins, alpha=0.5, color='#60a5fa',
             edgecolor='#fdba74', label='Fission-dominant')
axes[0].set_xlabel('Mitochondrion size (units)', color='#fdba74')
axes[0].set_ylabel('Count', color='#fdba74')
axes[0].set_title('Steady-state size distribution',
                  color='#fed7aa', fontweight='bold')
axes[0].legend(facecolor='#1c1917', edgecolor='#78350f', labelcolor='#fed7aa')
axes[0].grid(True, color='#78350f', alpha=0.3)

t = np.arange(len(tr_bal))*50
axes[1].plot(t, tr_bal, color='#fb923c', lw=2, label='Balanced')
axes[1].plot(t, tr_fus, color='#f87171', lw=2, label='Fusion-dominant')
axes[1].plot(t, tr_fis, color='#60a5fa', lw=2, label='Fission-dominant')
axes[1].set_xlabel('Step', color='#fdba74')
axes[1].set_ylabel('# discrete mitochondria', color='#fdba74')
axes[1].set_title('Network fragmentation dynamics',
                  color='#fed7aa', fontweight='bold')
axes[1].legend(facecolor='#1c1917', edgecolor='#78350f', labelcolor='#fed7aa')
axes[1].grid(True, color='#78350f', alpha=0.3)

plt.tight_layout()
plt.savefig('output.png', dpi=120, bbox_inches='tight', facecolor='#0a0a1a')
print('Fusion-dominant -> interconnected network (few large units)')
print('Fission-dominant -> fragmented (many small units, apoptosis/quality-control state)')
print('Mfn2 KO fibroblasts show near-permanent fission state; OPA1 KO also')

Click Run to execute the Python code

Code will be executed with Python 3 on the server

3. Axonal Transport & Miro/TRAK

Long neuronal axons (up to 1 m in humans) require continuous mitochondrial trafficking. Miro1/2 (RHOT1/2) is a Rho-family GTPase anchored in the outer mitochondrial membrane with two Ca²⁺-binding EF hands. It tethers mitochondria to TRAK1/2 adaptors, which in turn bind kinesin-1 (anterograde) and dynein (retrograde) motors for microtubule transport.

When a local Ca²⁺ spike (e.g., at a synaptic terminal) exceeds ~1 µM, Miro EF hands bind Ca²⁺ and release the motor, arresting the mitochondrion at the active site — matching ATP supply to demand. This calcium-gated arrest is central to synaptic plasticity and is disrupted in early PINK1/Parkin-linked Parkinson’s disease (Wang 2011 Cell).

4. Mitophagy: PINK1 / Parkin

Damaged mitochondria lose membrane potential (ΔΨ_m). In healthy mitochondria, the serine/threonine kinase PINK1 is imported via TOM/TIM, cleaved by matrix processing peptidase (MPP) and by inner- membrane protease PARL, and rapidly degraded by the proteasome. When ΔΨ_m drops, PINK1 import stalls; PINK1 accumulates on the outer membrane instead:

\[ \text{Healthy:} \ \text{PINK1} \xrightarrow{\text{TOM/TIM}} \text{matrix} \xrightarrow{\text{PARL/MPP}} \text{degraded} \]

\[ \Delta\Psi_m \downarrow \ \text{:}\ \text{PINK1 stabilised on OMM}\ \xrightarrow{\text{autophosphorylation}} \text{active kinase} \]

Active PINK1 phosphorylates ubiquitin at Ser65 and activates Parkin (an E3 ubiquitin ligase) by phosphorylating Parkin’s own ubiquitin-like domain. Activated Parkin synthesises K48- and K63-polyubiquitin chains on outer- membrane proteins (VDAC, MFN, MIRO, TOM20). These chains recruit the autophagy adaptors OPTN, NDP52, and p62, which anchor an LC3-decorated autophagosome. The whole damaged mitochondrion is engulfed and fused with a lysosome for degradation.

The process was first resolved by Narendra 2008 (J. Cell Biol.). PINK1 and Parkin mutations cause autosomal-recessive juvenile-onset Parkinson’s disease — the clearest monogenic link between mitochondrial quality control and neurodegeneration.

Simulation: PINK1/Parkin Cascade Timeline

Python

script.py44 lines

import numpy as np, matplotlib
matplotlib.use('Agg')
import matplotlib.pyplot as plt

# PINK1/Parkin mitophagy timeline after CCCP-induced depolarisation
t = np.linspace(0, 240, 2000)

# Membrane potential drops
dpsi = 180 * np.exp(-t/5) + 5
# PINK1 accumulation on OMM
PINK1 = np.where(t < 5, 0, 1 - np.exp(-(t-5)/10))
# Ubiquitin phosphorylation
pUb = PINK1 * (1 - np.exp(-t/30))
# Parkin recruitment
Parkin = pUb * (1 - np.exp(-t/40))
# Ubiquitin chain synthesis
chains = Parkin * (1 - np.exp(-t/60))
# LC3 / autophagosome
LC3 = chains * (1 - np.exp(-t/120)) * 0.9
# Lysosomal fusion / degradation
degraded = LC3 * (1 - np.exp(-(t-60)/60)) * (t > 60)

fig, ax = plt.subplots(figsize=(12, 6), facecolor='#0a0a1a')
ax.set_facecolor('#111827'); ax.tick_params(colors='#fdba74')
for s in ax.spines.values(): s.set_color('#78350f')
ax.plot(t, dpsi/180, color='#f87171', lw=2.4, label='Membrane potential (fraction)')
ax.plot(t, PINK1, color='#60a5fa', lw=2.4, label='PINK1 accumulation')
ax.plot(t, pUb, color='#a78bfa', lw=2.4, label='Phospho-ubiquitin')
ax.plot(t, Parkin, color='#fbbf24', lw=2.4, label='Parkin recruitment')
ax.plot(t, chains, color='#34d399', lw=2.4, label='Ub chains (K48/K63)')
ax.plot(t, LC3, color='#38bdf8', lw=2.4, label='LC3 autophagosome')
ax.plot(t, degraded, color='#fb923c', lw=2.4, label='Mitochondrion degraded')

ax.set_xlabel('Minutes after CCCP', color='#fdba74')
ax.set_ylabel('Relative level', color='#fdba74')
ax.set_title('PINK1/Parkin mitophagy cascade timeline',
             color='#fed7aa', fontweight='bold')
ax.grid(True, color='#78350f', alpha=0.3)
ax.legend(facecolor='#1c1917', edgecolor='#78350f', labelcolor='#fed7aa',
          fontsize=9, loc='center left', bbox_to_anchor=(1.02, 0.5))
plt.tight_layout()
plt.savefig('output.png', dpi=120, bbox_inches='tight', facecolor='#0a0a1a')
print('Healthy mitochondria: PINK1 is imported and degraded by PARL/MPP (rapid turnover)')
print('Depolarised: PINK1 stabilises on OMM, recruits Parkin, drives ubiquitin chains')

Click Run to execute the Python code

Code will be executed with Python 3 on the server

5. Alternative Mitophagy Pathways

Beyond PINK1/Parkin, several receptor-mediated mitophagy pathways exist for specific physiological contexts:

BNIP3 / NIX: outer-membrane receptors containing LIR motifs that bind LC3 directly. Essential for erythrocyte maturation (removal of all mitochondria) and for hypoxia-induced mitophagy.
FUNDC1: OMM receptor dephosphorylated during hypoxia, exposing its LIR motif. Links hypoxic stress to mitophagy.
Cardiolipin externalisation: oxidative damage flips cardiolipin from IMM to OMM, where LC3 binds it directly, targeting damaged mitochondria even without PINK1/Parkin.
Mieap / Lonp1-mediated: matrix quality control degrading specific damaged proteins within the mitochondrion rather than the whole organelle.

6. Functional Consequences of the Dynamic Balance

The fusion/fission ratio sets mitochondrial network state, which determines cellular phenotype:

Fused network: increased ATP output, tighter OXPHOS coupling, content mixing (mtDNA complementation, sharing of substrates), resistance to apoptosis. Dominant during G1-S cell cycle and nutrient abundance.
Fragmented network: facilitates quality control (isolates damaged units for mitophagy), apoptosis (Bax/Bak pore formation at fission sites), distribution to daughter cells during mitosis. Dominant during stress and M phase.

Many kinases tune the balance: AMPK (starvation), CaMK (Ca²⁺signals), ERK1/2 (growth), CDK1 (mitosis). Together they make the mitochondrial network a dynamic sensor-effector of cellular metabolic state.

7. Therapeutic Targeting

Drugging mitochondrial dynamics is an emerging therapeutic area:

Mdivi-1: small-molecule DRP1 inhibitor, preclinical in heart failure and neurodegeneration (off-target concerns limit clinical translation).
P110: peptide inhibitor of DRP1-Fis1 interaction; protective in Parkinson’s and Alzheimer’s models (Qi 2013).
Urolithin A (Amazentis): gut-microbiome-derived metabolite that induces mitophagy. Phase 2 trials in sarcopenia / aging (Andreux 2019 Nat. Metab.).
UDCA (ursodeoxycholic acid): clinical trial for DOA (OPA1 mutations). Stabilises mitochondria, improves vision.

Key References

• Chan, D. C. (2006). “Mitochondria: dynamic organelles in disease, aging, and development.” Cell, 125, 1241–1252.

• Youle, R. J. & van der Bliek, A. M. (2012). “Mitochondrial fission, fusion, and stress.” Science, 337, 1062–1065.

• Nakamura, T. et al. (2022). “Cryo-EM structure of the human mitofusin 2 dimer.” Nat. Struct. Mol. Biol., 29, 555–562.

• Narendra, D. et al. (2008). “Parkin is recruited selectively to impaired mitochondria and promotes their autophagy.” J. Cell Biol., 183, 795–803.

• Pickrell, A. M. & Youle, R. J. (2015). “The roles of PINK1, Parkin, and mitochondrial fidelity in Parkinson’s disease.” Neuron, 85, 257–273.

• Friedman, J. R. et al. (2011). “ER tubules mark sites of mitochondrial division.” Science, 334, 358–362.

• Andreux, P. A. et al. (2019). “The mitophagy activator urolithin A is safe and induces a molecular signature of improved mitochondrial and cellular health in humans.” Nat. Metab., 1, 595–603.

• Wang, X. et al. (2011). “PINK1 and Parkin target Miro for phosphorylation and degradation to arrest mitochondrial motility.” Cell, 147, 893–906.

← Module 5 Module 7: Mitochondria in Disease →