The terms, per the grant. On April 22, 2025, the Broad Institute was issued US12281303B2, “Methods and compositions for prime editing nucleotide sequences.” Prime editing is the next-generation editing approach positioned beyond first-generation CRISPR cutting, and a foundational institutional grant on it is the kind of asset a platform license is built around.

The independent claim is worth reading because it tells a deals desk exactly what the right is over. Claim 1 covers "a prime editing guide RNA (PEgRNA)" — the engineered guide that distinguishes prime editing from a standard CRISPR edit — comprising three named parts: "a spacer sequence"; "a gRNA core that is capable of complexing with a nucleic acid programmable DNA binding protein (napDNAbp) configured to generate a cut site"; and "an RNA extension arm" that itself carries "a DNA synthesis template comprising an edit template that encodes one or more edits" plus "a primer binding site." In other words, the grant claims the molecule that programs the edit, not just the abstract idea of editing.

Why a business desk values a foundational method grant: in gene editing, the deals that move money are platform licenses, and platform licenses are written around foundational method-and-composition IP that applies across many disease targets. An institutional grant on a next-generation modality is the kind of asset a company in-licenses to give its programs freedom to operate and a defensible base.

“Compositions and methods are provided herein for conducting prime editing of a target DNA molecule (e.g., a genome) that enables the incorporation of a nucleotide change and/or targeted mutagenesis.”— U.S. Patent No. 12,281,303 source

The dependent claims define the breadth of edit the platform can claim — and breadth is what determines how wide a licensee's freedom-to-operate becomes. The encoded edits "comprise one or more nucleotide insertions, one or more nucleotide substitutions, one or more nucleotide deletions, or a combination thereof" (claim 12). That single claim spans the three basic classes of genetic change, which is the feature that lets prime editing be marketed as a general-purpose tool rather than a point solution. The claims also fix engineering parameters that a licensee's chemistry would have to live inside: a 20-nucleotide spacer (claim 5), a primer binding site "from 7 to 17 nucleotides in length" (claim 7), a DNA synthesis template "from 5 to 58 nucleotides" (claim 15), and an edit that "disrupts an endogenous PAM site" to prevent re-cutting (claim 19).

The grant also ties the PEgRNA to the protein machinery, which matters for how the IP estate connects to existing CRISPR positions. The napDNAbp can be "a CRISPR-Cas effector protein" (claim 27), can comprise "an HNH domain, a RuvC domain, or both" (claim 28), and can be "a Cas9 protein" (claim 29) — and the guide can recruit additional effectors through "an aptamer," including "a MS2 aptamer" (claims 26, 30). For a model, a guide-RNA composition claim that explicitly reaches Cas9-based systems is a position that interlocks with, rather than sits apart from, the broader editing IP a licensee already needs.

The structure point: a foundational-method grant supports a platform royalty that collects across a licensee's whole pipeline, plus milestones per program, rather than a single-product royalty. For a model, that distinction is the difference between a narrow and a broad royalty base — and a claim that covers insertions, substitutions and deletions in one molecule is the kind of breadth that supports the broad base.

What the grant does not promise: a product, a clinical result, or a clear path around the other foundational editing patents held by various institutions. It is an exclusivity claim on a guide-RNA composition and the associated method, and it still has to coexist with the broader editing IP thicket.

To see why a guide-RNA composition claim anchors platform value, it helps to locate prime editing relative to the editing approaches that came before it. First-generation CRISPR makes a double-strand cut and lets the cell's repair machinery finish the job, which limits the precision of the change. Prime editing instead carries the intended edit on the guide itself: the PEgRNA's "RNA extension arm" supplies "a DNA synthesis template" encoding the edit plus "a primer binding site," and a fused polymerase writes that template into the genome. Because the engineered guide is the component that programs the edit, a composition claim over the PEgRNA is a claim over the thing that makes the modality work, not over a downstream use of it. That is the structural reason an institution can build a broad licensing platform on this single claim type.

The dependent claims are where a licensing desk reads the platform's reach and its engineering boundaries. Breadth comes from claim 12, covering edits that "comprise one or more nucleotide insertions, one or more nucleotide substitutions, one or more nucleotide deletions, or a combination thereof": one molecule spanning all three classes of change is what supports marketing prime editing as a general tool and therefore what supports a pipeline-wide royalty rather than a per-edit license. The architecture claims then fix the parts a licensee's design would have to respect: a homology arm "5 to 20 nucleotides in length" (claim 11), the 5'-to-3' order of homology arm, edit template and primer binding site (claim 13), and the PAM-disruption feature that "disrupts an endogenous PAM site associated with the spacer sequence" (claim 19) so the edited site is not re-cut. Claims 20 through 22 cover the single-molecule format in which spacer, gRNA core and extension arm are one species, with the extension arm "flanked by a 5' fragment of the gRNA core and a 3' fragment of the gRNA core." For a model, the more of the realistic design space the claims occupy, the larger and more defensible the platform royalty base, and these claims occupy a great deal of it.

The takeaway: when pricing a next-generation editing platform deal, read the foundational PEgRNA grants as the load-bearing assets. The Broad Institute's April 2025 prime-editing grant is a dated, concrete example of the institutional IP that platform licenses are built around — with claims broad enough to span the three core edit types in a single molecule.