The Journey of CRISPR: A Gene-Editing Revolution and its Implications for IP Protection

The Patent Race Hidden Inside the Discovery

On June 28, 2012, Jennifer Doudna and Emmanuelle Charpentier submitted a paper to Science demonstrating that a bacterial immune sequence called CRISPR-Cas9 could be programmed to cut specific DNA strands in a test tube. The biology was extraordinary. The IP move that followed was even more consequential — and most founders building on CRISPR today do not know what it was or why it cost UC Berkeley a patent portfolio worth billions.

Within months of that publication, Feng Zhang at the Broad Institute filed a patent application demonstrating CRISPR-Cas9 working inside eukaryotic cells — the cells that make up every plant, animal, and human being on earth. Berkeley's earlier filing covered the mechanism in a biochemical context. Broad's later filing covered the mechanism in a living human cell. When the US Patent and Trademark Office declared an interference proceeding in 2016, the question was not who discovered CRISPR first. It was who first reduced the invention to practice in a specific biological context that was not already described in the prior art. The Federal Circuit upheld Broad's claims in 2018. That single strategic distinction — the gap between a demonstrated mechanism and its first proven application inside a eukaryotic cell — has since anchored dozens of issued patents, licensing structures generating hundreds of millions in annual royalties, and the IP architecture of companies including Editas Medicine and Intellia Therapeutics.

For any founder building a CRISPR-derived platform today, understanding why that eukaryotic distinction mattered — and how to replicate the strategic logic in your own filing — is more valuable than any tutorial on how the Cas9 enzyme operates.

The §101 Problem Every CRISPR Founder Must Solve First

Before drafting a single claim, every CRISPR innovator collides with the same wall: 35 U.S.C. §101 bars patents on products of nature, natural phenomena, and abstract ideas. The Supreme Court's 2013 ruling in Association for Molecular Pathology v. Myriad Genetics made explicit what patent examiners were already applying — isolating a naturally occurring DNA sequence does not make it patent-eligible. Because CRISPR systems exist natively in bacterial genomes, any claim framed around the isolated mechanism in its natural form is dead on arrival.

The solution is not to avoid CRISPR's natural origins. It is to claim the engineered application in a specific context that does not exist in nature. The before/after contrast is concrete:

The second claim survives §101 scrutiny because it recites a specific cell type, a specific gene target, a specific delivery format, and a measurable outcome. None of that combination exists in nature. Each specificity element is also a potential invalidity vulnerability — which is why claim architecture in CRISPR requires a portfolio of claims layered from narrow (eukaryotic cell type + gene target) to slightly broader (class of hematopoietic cells + gene family), never a single broad claim that tries to capture the mechanism itself.

The Context Compression Trap

The Broad/Berkeley interference illustrates a pattern that repeats across every CRISPR startup: the patent-eligible claim space in gene-editing technology compresses from all sides simultaneously, not merely from the front.

In most technology categories, a founder's primary IP risk is competitors filing before them. In CRISPR-derived platforms, that risk is real — but the more insidious threat is the founder's own publication record combined with the broader literature. Call this the Context Compression Trap: every peer-reviewed paper demonstrating the CRISPR mechanism in a new cell type, delivery format, or therapeutic context permanently closes one patent-eligible application window. Because the mechanism itself is nature-derived and §101-ineligible in isolation, the only territory where durable claims can be driven is the gap between the mechanism's natural occurrence and its first engineered demonstration in a specific biological context. Each paper — yours or a competitor's — fills one of those gaps irreversibly.

Feng Zhang's team understood this intuitively. They raced to demonstrate eukaryotic function precisely because the prokaryotic context was already closing, and they filed before the eukaryotic context appeared in the literature. Founders who wait to publish until their platform is "complete" — a common instinct in academic spinouts — often discover that by the time they file, three adjacent context gaps have been closed by conference preprints they did not track. The 12-month US grace period does not protect you from a competitor's publication that closes the same context gap you were planning to claim. It only protects you from your own disclosures becoming prior art against your own application. Those are two entirely different problems, and conflating them is one of the most expensive mistakes a CRISPR-adjacent biotech can make.

Co-Inventorship and the Multi-Lab Problem

CRISPR research rarely happens in a single lab. The Doudna group collaborated with Charpentier's group across two continents; Zhang's team included collaborators across multiple MIT-affiliated institutions. Co-inventorship in patent law is not about who contributed labor — it is about who contributed to the conception of at least one claim. A graduate student who optimized transfection efficiency as directed is not a co-inventor. A postdoc who independently suggested the sgRNA architecture that made the eukaryotic demonstration possible almost certainly is.

Misidentifying inventors on a CRISPR patent application is not a technicality. Under Pannu v. Iolab Corp., a patent with incorrect inventorship is unenforceable until corrected — and if the omission was intentional, it can constitute inequitable conduct that renders the patent permanently unenforceable. For CRISPR platforms developed across academic-industry partnerships, where IP assignment agreements are often incomplete and lab notebooks span multiple institutions, inventorship audits should happen before the provisional application is filed, not during prosecution when the record has already hardened.

Practical rule: any person who contributed a non-obvious technical solution to a problem identified in any claim should be treated as a potential inventor until counsel rules them out in writing. The cost of that analysis is negligible against the cost of an enforceability challenge during a licensing negotiation.

Provisional Filing Strategy in a Fast-Moving Literature

The provisional patent application is the correct first move for most CRISPR innovators, but its value depends entirely on the level of detail it contains. A provisional that describes the mechanism generally — "CRISPR-Cas9 editing in mammalian cells" — does not give you priority date protection for claims directed at a specific cell type or delivery format that you develop in the following 12 months. Under New Railhead Manufacturing v. Vermeer Manufacturing, claims in the non-provisional are entitled to the provisional's filing date only to the extent the provisional's written description supports them. If your eukaryotic delivery optimization is not described in the provisional, a competitor who publishes that optimization nine months later may have closed that context gap before your non-provisional issues.

Given the Context Compression Trap, the correct use of a provisional in CRISPR-adjacent biotech is to file early and broadly across your anticipated context landscape, then use the 12-month window to generate the experimental data that supports the specific dependent claims in the non-provisional. This is the inverse of the intuition most academic founders bring from journal publishing, where you file after the experiments are complete. In patent strategy, the filing date is the asset — the experiments are what converts that asset into an enforceable claim.

Freedom-to-Operate in the Broad/Berkeley Shadow

Any company commercializing CRISPR-Cas9 in the United States operates under a licensing requirement that is not optional. The Broad Institute holds issued patents covering eukaryotic CRISPR-Cas9 use; UC Berkeley holds patents covering the core Cas9-guide RNA mechanism in biochemical contexts; MilliporeSigma, the Vilnius University group, and others hold additional foundational rights. Most commercial applications will require licenses from at least two of these portfolios simultaneously.

A freedom-to-operate analysis for a CRISPR platform should address three distinct questions that generic FTO guidance collapses into one. First, does your specific cell type and delivery method fall within the literal scope of any issued claim? Second, if it does not literally infringe, could the doctrine of equivalents reach your implementation? Third, are the patents you rely on as design-arounds actually valid in light of the 2012 Jinek Science paper and the subsequent literature — because an FTO opinion built on patents that may be invalidated in post-grant review is a false comfort.

Licensing structure matters as much as licensing existence. Broad's non-exclusive licensing program for therapeutics is available at published rates; Berkeley's licensing has historically been more selective. A startup whose business model depends on freedom in a therapeutic indication that Broad has licensed exclusively to a direct competitor is not insulated by a general FTO opinion.

Where the White Space Still Exists

The foundational CRISPR-Cas9 claims are largely settled territory. The patent-eligible white space for new entrants is defined by three concrete technical axes, each rooted in the Context Compression logic:

• Delivery mechanisms: Lipid nanoparticle formulations, AAV serotype specificity for target tissues, and base-editing delivery in post-mitotic cells remain areas where novel engineering combinations can yield application-specific claims that do not read on existing foundational patents.
• Next-generation editors: Prime editing, base editing (cytosine and adenine), and epigenome editors operate on distinct mechanistic principles from Cas9 nuclease activity. Claims anchored to these systems' unique features — the reverse transcriptase domain in prime editing, for instance — do not automatically fall within the Broad/Berkeley claim structures and represent genuine filing opportunities.
• Non-mammalian contexts with commercial relevance: CRISPR applications in plant genomes, agricultural microbiomes, and aquaculture species involve cell types and regulatory frameworks distinct enough from the therapeutic patent thicket that a focused portfolio can establish meaningful exclusivity without navigating the foundational license landscape.

In each case, the strategic imperative is identical to the lesson from the 2018 Federal Circuit ruling: file at the boundary of the demonstrated context, not after the literature has closed it. The Broad did not win because it had the most brilliant lawyers or the largest filing budget. It won because Zhang's team understood that the eukaryotic cellular context was the next unclaimed gap, and they filled it before anyone else published in that space. Every CRISPR innovator operating today faces an equivalent choice — the context gaps are smaller, the literature moves faster, and the cost of waiting has compounded accordingly.

Action Steps: Filing Sequence for CRISPR-Adjacent Founders

1. Map your context landscape before filing anything. Identify every cell type, delivery format, and therapeutic indication your platform could address. This is not a business exercise — it is a prior art scoping exercise. Any context already appearing in the literature is a closed gap.
2. File a detailed provisional that describes each open context. At USPTO's current micro-entity provisional fee (~$320), filing a provisional that describes five distinct context applications costs less than one hour of outside counsel time. The protection it creates is disproportionately large.
3. Conduct an inventorship audit before the provisional is filed. Identify every person who contributed a non-obvious technical solution to any element you intend to claim. Resolve ambiguities in writing, with IP assignment documentation, before the application creates a record.
4. Commission a tri-portfolio FTO before closing a seed round. Broad, Berkeley, and MilliporeSigma represent the three foundational licensing relationships. Knowing your position against all three before investor conversations is table stakes, not due diligence.
5. Track the literature weekly against your open context gaps. Set automated alerts for your target cell types and delivery mechanisms. A competitor's preprint posting is not prior art until published, but it signals that the context gap is about to close — giving you a narrow window to file before it does.

The CRISPR patent landscape is not a barrier that blocks new entrants. It is a topology — complex, mapped in some areas and genuinely open in others — that rewards founders who understand its specific shape. The Broad Institute's eukaryotic victory was not inevitable. It was the product of a deliberate filing decision made at a precise moment in the literature's development. That same decision, applied to the open contexts that still exist in 2025, remains available to founders who move before the compression closes the remaining gaps.

This article is for informational purposes only and does not constitute legal advice. Consult a registered patent attorney for guidance specific to your invention.