The naive question: if Cas9 already works, why does the field keep inventing new CRISPR enzymes? Because Cas9 is one tool with one set of trade-offs, and not every job suits it. Type V enzymes are a different family chosen for different strengths — and the 2025 patent record shows them being claimed for real therapeutic use, not just bench curiosity.

Think of it like cutting tools in a workshop. Cas9 is the well-known utility knife. Type V enzymes — Cas12 and its relatives — are like a smaller, differently-shaped blade: in some cases more compact, in some cases leaving a different kind of cut, in some cases reaching spots the bigger tool can't.

“In one aspect, embodiments disclosed herein are directed to engineered CRISPRCas effector proteins that comprise at least one modification compared to an unmodified CRISPR-Cas effector protein that enhances binding of the of the CRISPR complex to the binding site and/or alters editing preference as…”— U.S. Patent No. 12,305,204 source

That quote, from the Broad Institute's 2025 grant US12305204B2 on CRISPR enzymes and systems, captures the two levers engineers pull on a Type V cutter: how tightly it binds its target, and what it prefers to edit. But the granted claims of that patent are about something more procedural and, arguably, more interesting — they cover a method for designing the guide, not just the enzyme. The claim walks through selecting candidate target sequences for a population, then deliberately throwing out any candidate that carries common variants in that population, and any whose near-matches ("off-target candidates") appear in haplotypes present in even 0.1% of people. What's left is a target set chosen to cut cleanly across genetic diversity. Dependent claims fold in PAM type and length, cleavage efficiency, and even sequencing the individual patient's genome before designing their guide. That is population-aware, almost regulatory-grade design discipline written into a patent claim — and it explicitly applies to both Type II (Cas9) and Type V (Cas12) guide molecules.

Size matters more than it sounds, and it's a recurring reason to reach for Type V. Gene-editing therapies often have to be packed into a delivery vehicle with limited room, such as a viral capsid that can only hold a small payload. A more compact enzyme leaves more space for the rest of the cargo. Caribou Biosciences' publication US20250325698A1 covers therapeutic applications of CRISPR Type V systems — these enzymes aimed squarely at treating disease rather than demonstrating a mechanism.

The guides change too, and that's a genuine point of difference from Cas9. Caribou's US20250027078A1 covers DNA-containing guides for Type V systems — a different guide chemistry from the all-RNA guides Cas9 uses. Building part of the guide from DNA can change how it is made, how stable it is, and how it behaves once inside a cell. Different enzyme family, different instruction format: the same job done with a re-engineered set of parts.

It is worth noting how the foundational and applied players line up in this 2025 window. The Broad's grant reads like an anchor estate — a design-method claim broad enough to span both major enzyme families and to recite a list of disease-associated loci, from tumor antigens to targets like transthyretin (TTR) and PCSK9. Caribou's two publications read like applied bets on the Type V family specifically: one staking the therapeutic use, one staking the guide chemistry. Read together, they show a field that is no longer standing on Cas9 alone but is actively claiming the next family of cutters and the methods to aim them.

The design-method claim in the Broad grant deserves a second look, because it reframes what "a CRISPR patent" even covers in 2025. Earlier estates often claimed an enzyme or a guide as a composition of matter. This one claims a procedure: gather candidate target sequences, discard those overlapping common population variants, discard those whose near-matches appear in even 0.1% of haplotypes, and only then build the guide — optionally after sequencing the specific individual. That is a workflow you would associate with careful, safety-minded engineering, and claiming the workflow rather than a single molecule is a deliberately broad move, since it reaches any Type II or Type V guide designed that way.

Caribou's two publications show the applied counterpart to that breadth. One stakes therapeutic use of the Type V family directly; the other stakes a re-engineered guide that incorporates DNA, not just RNA. The DNA-containing guide is a concrete example of why a new enzyme family is more than a relabeling exercise: a Type V cutter can accept an instruction format that Cas9's machinery would not, which in turn changes how the guide is manufactured and how it behaves in a cell. The 2025 record, in other words, captures both halves of a generational shift — broad method claims from the anchor institution and specific, format-changing applications from a focused player.

A practical consequence of all this is that "CRISPR" is no longer a single product category but a toolbox with diverging branches. The Broad's population-aware design method, the Type V therapeutic-use stake, and the DNA-containing guide each address a different bottleneck — choosing safe targets, deploying a smaller cutter, and supplying it with a buildable instruction — and the 2025 filings claim them as separable inventions rather than one bundled platform.

The short version: Type V enzymes are CRISPR's other cutters — often smaller, sometimes leaving a different cut, and easier to deliver. They exist because one enzyme can't do every job. The 2025 patents show them being pointed at real therapies, paired with re-engineered DNA-containing guides, and designed with population-aware, off-target-screening methods that treat guide selection as a careful engineering process in its own right.