Lipopeptide–Membrane MD with AMBER (Lipid21 + PACKMOL-Memgen)

A practical, runnable protocol to simulate an N-acylated peptide (lipopeptide) interacting with a POPE:POPG (3:1) bacterial membrane using AmberTools/AMBER, Lipid21, and PACKMOL-Memgen.

Example context. The input example files correspond to a 9-mer helical antimicrobial peptide N-acylated with a C8 fatty acid. The peptide sequence and PDB are intentionally omitted (unpublished data), but all other files (fatty-acid ligand parameters, TLEaP inputs, MD control files) are provided so you can reproduce the system-building and MD pipeline with your own peptide coordinates.

Expected behavior. In reference runs, the lipopeptide anchors at the headgroup region via the fatty-acid tail and the helical segment inserts into the bilayer during equilibration and production. These templates are reusable for analogous lipopeptides (e.g., C8–C10 N-acyl chains) with minimal edits to lipo/LIG.cif and residue indices in TLEaP.

Pre (left): fatty-acid tail anchors at the headgroup region. Post MD (right): the helical segment inserts into the bilayer.

---

Contents

• Overview
• Requirements & environment
• Repo structure
• Commands cheat sheet (exact steps)
• Step-by-step explanation
• Detailed commands & input templates
• Analysis notes
• References
• Author

---

Overview

We will: (1) parameterize the fatty-acid cap as a small molecule (GAFF/AM1-BCC), (2) build the peptide and form the N-acyl amide, (3) pack a POPE:POPG bilayer with PACKMOL-Memgen, (4) assemble everything in TLEaP (Lipid21 + ff14SB + TIP3P), re-solvate and add ions (neutrality and/or ~150 mM), and (5) run MD: minimization → thermalization/heating → equilibration → production using min1.mdin, min2.mdin, term.mdin, eq.mdin, and md.in.

---

Requirements & environment

• AmberTools / AMBER (AmberTools22+ recommended; includes packmol-memgen)
• GPU AMBER (pmemd.cuda) or CPU (pmemd / sander)
• Force fields: leaprc.protein.ff14SB, leaprc.lipid21, leaprc.water.tip3p
• Optional modelers/editors: Chimera(X), PyMOL (Sculpting), DS Visualizer
• Bash + Python (optional helper for ion counts)

Define AMBER install location

Set AMBERHOME and add AmberTools to PATH:

# Linux/macOS (bash/zsh)
export AMBERHOME="/home/cluster/amber22"
export PATH="$AMBERHOME/bin:$PATH"

# Verify:
which antechamber
which tleap

(If your install lives elsewhere, change the path accordingly.)

---

Repo structure

.
├── lipo/                   # fatty acid (lipid cap) parametrization
│   ├── LIG.cif             # standalone fatty acid as -COOH (temporary OH)
│   ├── lig.ac              # antechamber output
│   ├── lig.prepin          # prepgen output
│   └── frcmod.lig          # parmchk2 output
├── peptide/
│   └── lipopep.pdb         # peptide + C8 fatty acid (N-acyl) starting coords
├── membrane/
│   └── system.pdb          # membrane + peptide, cleaned/merged (TLEaP INPUT)
├── tleap/
│   ├── lipopep_membrane.in # run from repo root:  tleap -f tleap/lipopep_membrane.in
│   └── leap.log
├── mdin/
│   ├── min1.mdin           # STEP 1: restrained minimization
│   ├── min2.mdin           # STEP 2: unrestrained minimization
│   ├── term.mdin           # STEP 3: thermalization / heating (NVT)
│   ├── eq.mdin             # STEP 4: equilibration (NPT)
│   └── md.in               # STEP 5: production (NPT)
├── pre-post-md.jpeg
└── README.md

---

Commands cheat sheet (exact steps)

1) **Create structure (peptide + C8 fatty acid)**
   - Save peptide+lipid as `peptide/lipopep.pdb`.
   - Save the fatty acid alone (with extra –OH as carboxylic acid) as `lipo/LIG.cif`.

Note: de novo peptide modeling (Rosetta/AlphaFold/other) is outside the scope of this tutorial.

2. Parametrize the fatty acid (uses $AMBERHOME)

antechamber -fi ccif -i lipo/LIG.cif -bk LIG -fo ac -o lipo/lig.ac -c bcc -at amber
prepgen     -i lipo/lig.ac -o lipo/lig.prepin -m lipo/lig.mc -rn LIG
parmchk2    -i lipo/lig.prepin -f prepi -o lipo/frcmod.lig -a Y -p "$AMBERHOME/dat/leap/parm/parm10.dat"

3. PACKMOL-Memgen: build membrane–peptide complex

packmol-memgen --lipids POPE:POPG --ratio 3:1 --solute peptide/lipopep.pdb --solute_con 1 --parametrize --distxy_fix 50
# --distxy_fix sets the XY side length in Å (e.g., 50 Å per side for a ~50×50 Å patch).

4. Manual clean & reposition (TLEaP-friendly input)

   • Remove all waters/ions from the PACKMOL PDB (keep only lipids + peptide).
   • Reposition the peptide in DS Visualizer / Chimera / PyMOL and save.
   • Because viewers may reorder membrane atoms, copy only the peptide ATOM/HETATM records from the viewer-saved file back into the original membrane PDB.
   • Save as membrane/system.pdb (TLEaP input).

5. TLEaP (one-liner; run from repo root)

tleap -f tleap/lipopep_membrane.in
# Outputs: system.prmtop, system.rst7, system_solv.pdb, tleap/leap.log

6. Minimization, heating, equilibration (CPU with sander)

sander -i mdin/min1.mdin -p system.prmtop -c system.rst7     -r systemmin1.rst7 -o systemmin1.mdout -ref system.rst7
sander -i mdin/min2.mdin -p system.prmtop -c systemmin1.rst7 -r systemmin2.rst7 -o systemmin2.mdout
sander -i mdin/term.mdin -p system.prmtop -c systemmin2.rst7 -r systemterm.rst7 -o systemterm.mdout -x systemterm.nc
sander -i mdin/eq.mdin   -p system.prmtop -c systemterm.rst7 -r systemeq.rst7   -o systemeq.mdout   -x systemeq.nc

7. Production

# either your script:
sh run.xxx
# or direct with GPU:
pmemd.cuda -O -i mdin/md.in -o md.out -p system.prmtop -c systemeq.rst7 -r md.rst -x md.nc

---

Step-by-step explanation (what each stage does)

1. Fatty-acid cap parametrization (lipo/)
   The fatty acid is not covered by protein/lipid FFs, so we parametrize it as a small molecule with GAFF/GAFF2 and AM1-BCC charges (Antechamber/parmchk2). Build it temporarily as –COOH to get stable atom types/charges; the extra O/H are removed when forming the amide.

2. Peptide build & N-acyl bond (peptide/)
   De novo modeling of the peptide can be obtained with Rosetta, AlphaFold, or other tools and is not part of this tutorial.
   Create peptide/lipopep.pdb with your peptide model and a C8 fatty acid linked via an amide (fatty-acid carbonyl C to peptide N-terminal N).

3. Membrane packing + clean/reposition (membrane/)
   PACKMOL-Memgen builds a POPE:POPG (3:1) patch and places the lipopeptide.
   Then remove waters/ions and carefully merge the peptide back into the original membrane PDB to avoid re-ordering artifacts. Result: membrane/system.pdb (TLEaP input).

4. System assembly & re-solvation in TLEaP (tleap/)
   Load Lipid21 (membrane), ff14SB (peptide), TIP3P (water), and your ligand params (lipo/lig.prepin, lipo/frcmod.lig). Re-solvate with solvatebox (TIP3P) and add ions for neutrality and/or ~150 mM. This yields system.prmtop, system.rst7, system_solv.pdb.

150 mM quick recipe (AMBER-style):

• From tleap/leap.log get the box volume in ų → convert to liters: V_L = V_ang3 × 1e-27
• Moles at 150 mM: n_pairs = 0.150 × V_L
• Ion pairs: N_pairs = n_pairs × 6.022e23 → round to nearest integer
• Add equal Na⁺ and Cl⁻ after neutrality, e.g.:
  addIons prot Na+ 19 Cl- 19

5. MD sequence (mdin/)
   • min1.mdin — restrained minimization (relieve clashes; restrain solute heavy atoms).
   • min2.mdin — unrestrained minimization.
   • term.mdin — thermalization/heating (NVT to 310 K) with light restraints.
   • eq.mdin — equilibration (NPT) at 1 atm / 310 K.
   • md.in — production (NPT) at 310 K / 1 atm.

---

Detailed commands & input templates

Ligand (fatty-acid cap) parametrization

antechamber -fi ccif -i lipo/LIG.cif -bk LIG -fo ac   -o lipo/lig.ac   -c bcc -at amber
prepgen     -i lipo/lig.ac -o lipo/lig.prepin -m lipo/lig.mc   -rn LIG
parmchk2    -i lipo/lig.prepin -f prepi    -o lipo/frcmod.lig -a Y \
            -p "$AMBERHOME/dat/leap/parm/parm10.dat"

Note: lipo/LIG.cif is the standalone fatty acid in –COOH form (temporary O/H only for parametrization).

TLEaP (short version using an .in file)

Run from the repo root (so relative paths work):

tleap -f tleap/lipopep_membrane.in

This generates:

system.prmtop
system.rst7
system_solv.pdb
tleap/leap.log

Why remove PACKMOL water and re-solvate with TLEaP?
PACKMOL outputs may include water/ions for packing, but we strip them and keep only membrane + peptide in membrane/system.pdb. Then we solvate with solvatebox in TLEaP because:

• TLEaP uses pre-equilibrated TIP3P and trims overlaps cleanly around membranes.
• Hydration is uniform and reproducible.
• Ion addition and target molarity are controlled after TLEaP solvation.

---

Analysis notes

• Track bilayer thickness, area per lipid, density, and leaflet balance during equilibration.
• Check lipopeptide orientation/insertion; gentle Z-axis restraints at early stages can help if starting above headgroups.
• For long productions, segment runs (restart files) and adjust ntpr/ntwx/ntwr to your analysis/storage plan.
• Consider center-of-mass removal and PBC handling during analysis (e.g., imaging lipids and peptide).

---

References

• PACKMOL-Memgen (Amber community tutorials & paper)
• AMBER ligand (Antechamber/parmchk2) workflows
• Lipid21 (AMBER lipid force field) and usage notes
• Salt concentration calculation from leap.log volume (classic AMBER recipe)

---

👤 Author

Ivan Sanchis, PhD
Laboratorio de Péptidos Bioactivos
Facultad de Bioquímica y Ciencias Biológicas
Universidad Nacional del Litoral
Santa Fe, Argentina
📧 [email protected] / [email protected]