restriction-modification system

Deep research

Generated by just research-trait; source: research/traits/genomics/restriction_modification_system-deep-research-falcon.md

# Curation report: restriction–modification system

## Trait record and recommended scope

- **Trait label:** restriction-modification system
- **Trait identifier:** `traitmech:000095`
- **Category / kind / status:** GENOMICS / CLASS / REVIEWED
- **Parent:** `METPO:1000188`
- **Synonym:** R-M system

### Scope summary

This trait should represent **possession of a functional, sequence-specific self/non-self DNA-discrimination module** in which a DNA methyltransferase marks cognate sites in cellular DNA and a restriction endonuclease attacks cognate sites in DNA that lacks the protective methylation state. Canonical Type I–III systems fit this definition. The minimal causal chain is:

> cognate methyltransferase → methylated host recognition sites → protection of self DNA; and cognate restriction endonuclease + unprotected foreign recognition sites → foreign-DNA cleavage → reduced establishment of phages or other mobile genetic elements.

Type II systems are especially clear examples: an REase cuts double-stranded DNA at specific 4–8-bp targets, often palindromic, while the paired MTase methylates the same targets. Comparative work reports that 83% of surveyed prokaryotic genomes encode at least one R-M system and that Type II systems occur in 39.2% of bacterial genomes, at approximately 0.5 system per genome. These are database-dependent estimates, not a universal biological constant. (shaw2023restrictionmodificationsystemshave pages 1-2)

### Boundary cases

1. **Orphan/solitary methyltransferases:** insufficient by themselves. They can regulate replication, repair, or transcription without a cognate REase and therefore should not automatically instantiate `traitmech:000095`. Solitary MTases may be represented as related epigenetic traits or optional downstream regulators. More than 90% of solitary R-M-component hits found in phages in one comparative analysis were MTases, illustrating why component detection cannot substitute for a functional-system call. (oliveira2014theinterplayof pages 11-12)
2. **Standalone restriction or nicking endonucleases:** insufficient unless a cognate self-protection mechanism is demonstrated.
3. **Type IV modification-dependent restriction:** these enzymes preferentially recognize modified DNA, unlike the methylation-blocked restriction performed by canonical Types I–III. Type IV should therefore be modeled as a distinct subtype or neighboring trait, not forced into the core “unmethylated foreign DNA” branch.
4. **CRISPR–Cas, abortive infection, CBASS, BREX, DISARM, phosphorothioate defense:** separate antiviral traits even when colocated in defense islands.
5. **Methylome evidence alone:** a modified motif does not establish a complete R-M system. A cognate nuclease, genomic linkage, loss-of-function phenotype, biochemical cleavage, or strong curated annotation is needed.
6. **Inactive, pseudogenized, or phase-OFF loci:** genomic possession and current activity should be separately represented. Phase-variable systems can alter methylation and gene expression without being constitutively active. (vasu2013diversefunctionsof pages 13-14, vasu2013diversefunctionsof pages 14-15)
7. **Anti-restriction and phage DNA modification:** these are modifiers of penetrance, not evidence that the host lacks the trait. Phages and plasmids may acquire host-compatible methylation, encode anti-restriction proteins, alter target abundance, or hypermodify DNA. (loenen2014typeirestriction pages 2-3)

## Mechanistic classes

| Class | Defining organization and action | Curation implication |
|---|---|---|
| **Type I** | `hsdR`, `hsdM`, and `hsdS` encode a pentameric R₂M₂S complex. HsdS recognizes a bipartite motif; HsdR uses ATP-driven translocation and cleaves at variable distances. ATP, Mg²⁺, and SAM are required. | Model HsdS specificity, MTase protection, ATP-dependent translocation, and distant cleavage as a Type-I-only branch. (loenen2014typeirestriction pages 2-3) |
| **Type II** | Usually separate REase and MTase activities recognize the same short motif; cleavage occurs at or near a defined position. Mg²⁺ supports cleavage and SAM supplies the methyl group. | Best basis for the compact universal graph, but architecture can include fused or unusual enzymes. (shaw2023restrictionmodificationsystemshave pages 1-2, heitman1993ontheorigins pages 1-4) |
| **Type III** | Mod and Res activities form a complex; modification is strand-specific and restriction is ATP-dependent. | Keep subtype-specific details outside the universal core unless the individual system is experimentally typed. (heitman1993ontheorigins pages 1-4) |
| **Type IV** | Modification-dependent enzymes attack methylated or otherwise modified DNA. | Do not assert “methylation protects DNA” for Type IV; create a distinct modification-dependent branch. |

## Candidate nodes grouped by type

### Trait and module nodes

- restriction-modification system — `traitmech:000095`
- Type I R-M system — label-only candidate
- Type II R-M system — label-only candidate
- Type III R-M system — label-only candidate
- Type IV modification-dependent restriction system — label-only boundary/subtype
- cognate restriction–modification recognition motif — label-only
- host methylation pattern / methylome — label-only

### Genes, proteins, enzymes, and complexes

- DNA methyltransferase — candidate grounding **GO:0009008** (DNA-methyltransferase activity)
- restriction endonuclease — candidate grounding **GO:0009036** (Type II site-specific deoxyribonuclease activity) only when Type II-specific; use a label-only general REase node otherwise
- Type I HsdR restriction/motor subunit — label-only
- Type I HsdM methylation subunit — label-only
- Type I HsdS specificity subunit — label-only
- Type I R₂M₂S restriction complex — label-only
- Type III Mod subunit — label-only
- Type III Res subunit — label-only
- controller protein / C protein — optional, system-specific label-only node; no general edge should be curated without locus-specific evidence

Showing the first 60 of 224 lines of findings; the linked file also carries the run's front matter and the prompt it was given — read the full report.