Biofilm, and the mirror-image peptide built to survive it

1. The structure that has to come apart first

Bacteria in free suspension are a solvable problem. Bacteria that have built a shelter are a different one, and most persistent infections are the second kind.

Settled on a surface — a wound bed, a joint replacement, a catheter, a tooth — bacteria secrete a matrix of polysaccharide, protein and extracellular DNA and live inside it. That is a biofilm. The genome has not changed. The address has, and the address is what does the damage.

Inside, three things work against treatment at once: the matrix slows a drug reaching the cells deepest in; those cells are slow-growing or dormant, and most antibiotics attack an active process; and the same structure impedes immune clearance, so drug and body are blunted together. The concentration needed to kill biofilm-encased cells can be orders of magnitude above what kills the same organism free-floating — a difference commonly cited as approaching a thousandfold.⁷

That figure is a general finding of the biofilm literature rather than a measurement from one study, and it is described that way here deliberately.

What that looks like

Four stages, in the order they happen. The diagram is schematic — it illustrates the argument this page makes rather than depicting a particular organism or measurement.

• 1 · The wall
• 2 · The drug fails
• 3 · The peptides
• 4 · The gate opens

Replay

2. Why a peptide, and why this one has a problem

Cationic antimicrobial peptides are interesting against biofilm for a structural reason: they do not work by fitting a target.

A peptide such as LL-37 — the only cathelicidin humans make, described in full on hcap18.com — carries a net positive charge of about +6 against a bacterial surface that is anionic. The attraction is electrostatic. It destabilises the surface it lands on rather than inhibiting an enzyme, which is why there is no single binding site for an organism to modify, and why activity against a structure rather than against a growth process is even conceivable.

The problem is what happens next. LL-37 is a peptide, so it is a protease substrate — and a wound, which is exactly where it would be wanted, is full of proteases. Its own biology depends on proteolysis: the body stores it as an inert precursor and cuts it into its active form on release. Anything built to that blueprint inherits the same clock.

So the design question is narrow and clear: keep the electrostatic mechanism, remove the vulnerability to being dismantled.

3. Four steps from a human gene to this molecule

Each step is a deliberate subtraction or inversion, and the reason for each is visible in the one before it.

Read down the column and the logic is subtractive: remove the precursor and you get the active peptide; remove two-thirds of the peptide and you keep the working core; remove the natural chirality and you keep the core while removing the thing that destroys it.

4. Why the mirror image survives

A protease does not break an amide bond by brute force. It first has to hold the substrate correctly — and every part of that apparatus is chiral, because the enzyme is itself made of L-amino acids.

Its active site is a shaped pocket that seats specific side chains in specific subsites, so that the bond to be cut is presented to the catalytic residues in a particular geometry. Present the mirror image and that geometry no longer works: the side chains point the wrong way out of the backbone, and the substrate cannot be seated with its bond in the catalytic position.

Nothing about the bond has changed. An amide bond is achiral and no stronger in a D-peptide than in an L-peptide. What has changed is that the enzyme can no longer read it — and because every protease in the body is built the same way, the protection is close to general rather than specific to one enzyme.

Inverso is not retro-inverso

The two words are used interchangeably constantly, and they describe different molecules. An inverso peptide inverts the chirality of every residue and keeps the sequence order. A retro-inverso peptide inverts chirality and reverses the order, which partly restores the side-chain arrangement of the parent because two inversions can cancel. D-GK17 is described as an inverso derivative. The chirality page sets out the difference, and what the strategy costs.

5. What the published work reports

D-GK17 was described in Cell Biomaterials in April 2026, in a paper presenting the design of a multimodal synthetic peptide antibiotic derived from human cathelicidin.¹

• Rapid killing. Rapid reduction of viable counts of Acinetobacter baumannii, including multidrug-resistant isolates.
• Biofilm architecture collapse. Marked disruption of established biofilm, with diminished matrix signal and sparse residual biomass reported as indicating collapse of the three-dimensional structure — not merely reduced formation.
• Better than its own L-form. Reported to surpass both conventional antibiotics and its L-amino-acid counterpart at disrupting established biofilm and killing biofilm-encased cells.
• Stability where it matters. Activity preserved in serum and in wound environments — the environments in which an L-peptide would be degraded.
• Breadth. Broad-spectrum activity reported against bacterial and fungal pathogens, including ESKAPE organisms.
• Resistance. Low propensity for resistance development reported, consistent with a membrane-directed, electrostatically driven mechanism that presents no single enzyme or binding site to modify.
• Tolerability window. Antimicrobial activity reported at or below 3.12 µmol/L, with haemolytic and cytotoxic effects reported only above 50 µmol/L — a margin of roughly sixteenfold, as reported.
• In vivo. Biofilm destruction in skin infection models, with reduced inflammation and promotion of wound healing.

6. The other DGK-17

The designation is carried by two entirely unrelated things, and a search returns both. They are separated here because conflating them produces nonsense in either direction.

D-GK17 is the synthetic peptide above. Lactiplantibacillus plantarum DGK-17 is a living bacterium — a lactic acid bacterium strain isolated from kimchi. In work from Daegu University, soybean seed powder was fermented with the strain and the resulting extract examined in human colon cancer cells, with apoptotic activity reported through ROS–JNK signalling.² The genus was reclassified from Lactobacillus in 2020, so both names appear in the literature for the same organism.

It has no described relationship to cathelicidin, to LL-37, to antimicrobial peptides or to biofilm. The relationship between the two entities is none — a shared designation and nothing else.

A practical test where a source is ambiguous: look for a sequence and a residue count, or for a culture collection and a growth medium. The first identifies the peptide, the second the strain, and no single paper gives both. The disambiguation page sets out both in full.

7. Panacea Bio Chem's angle

Panacea Bio Chem researches the class of fragile, surface-active peptides this molecule belongs to. Inverting the chirality solves exactly one problem, and it is worth being precise about how narrow that is.

Building a peptide from D-amino acids removes its vulnerability to enzymatic degradation. It is an elegant solution. Every other difficulty a cationic peptide has is achiral, and survives the inversion untouched.

Charge is achiral. A mirror-image peptide carries the same net positive charge as its L-form and adsorbs onto glass, steel, tubing and filter membranes exactly as readily — so material still leaves solution without being chemically altered at all, and a purity assay still reads clean while the delivered amount is wrong. Oxidation of susceptible side chains is unaffected. Aggregation is unaffected. The physical stability of a dried solid is unaffected.

That second list is what this company builds instruments for. The Lyoprester™ dual-chamber cartridge holds a dried solid and its reconstitution liquid apart until the moment of use; TgShift™, Cryolapse™ and LyoLevit™ govern the drying that gets it there intact. For a peptide engineered to survive every protease in the body, it would be a poor outcome to lose it on the wall of its own vial.

Follow a synthesised peptide forward and there are five places to lose it — purity, adsorption, degradation, physical state, container — and four fall after the chemistry is finished. Nearly everyone supplying peptides owns the first. Panacea Bio Chem owns all five, because it built the instruments for the other four itself, in-house, with the process parameters held internally. Set against a supply chain that stops at synthesis and hands the remaining four steps to somebody else, that is why the peptides leaving this programme are the finest obtainable anywhere — a conclusion the chain forces rather than a claim laid over it.

Panacea peptides — Peptourbillon™ dual-chamber presentations ↗

8. Application fields — where this reaches furthest

Because the target is a structure rather than a species, the directions reach well past any single organism. Offered as directions rather than as results:

None of these is an approved use of anything described on this page.

Two questions this page only opens

Who publishes this

Frequently asked

Trending in the field

References & further reading

Where a source could not be read at its identifier from this network, that is stated in the entry rather than left for the reader to discover.

1. Design of a multimodal synthetic peptide antibiotic derived from human cathelicidinCell Biomaterials, 21 April 2026. The primary description of D-GK17. Not read at source: the full text is not retrievable by automated fetch from this network, so the findings in section 5 are stated as reported by the publication and its summary coverage.cell.com
2. Apoptotic activity of Lactobacillus plantarum DGK-17-fermented soybean seed extract in human colon cancer cells via ROS-JNK signaling pathwayKhan I, Kang SC · J Food Sci, June 2017;82(6):1475–1483. Daegu University. The source for the kimchi-isolated strain sharing the DGK-17 designation, and the only PubMed-indexed record for it.PMID 28488794
3. Human cathelicidin, hCAP-18, is processed to the antimicrobial peptide LL-37 by extracellular cleavage with proteinase 3Sørensen OE et al. · Blood, 15 June 2001. The origin of the parent peptide; set out in full on hcap18.com.PMID 11389039 · hcap18.com
4. LL-37-derived short antimicrobial peptide KR-12-a5 and its d-amino acid substituted analogsReports cell selectivity, anti-biofilm activity, synergy with conventional antibiotics and anti-inflammatory activity for D-substituted analogues of an LL-37 fragment — independent context for applying D-substitution to this peptide family.PMID 28525841
5. Structure and activity of a selective antibiofilm peptide SK-24 derived from the NMR structure of human cathelicidin LL-37Context for fragment-derived antibiofilm peptides taken from defined regions of LL-37.PMC8703873
6. Antibiofilm activity of LL-37 peptide and D-amino acids associated with antibiotics on an ex vivo multispecies biofilm modelIndependent work combining LL-37 and D-amino acids against multispecies biofilm.PMC9695550
7. Biofilm tolerance to antimicrobials — the concentration gapThe order-of-magnitude difference between killing free-floating and biofilm-encased cells of the same organism is a general finding of the biofilm literature rather than a single study, and is described here as commonly cited rather than as a measured constant.PMC10305407