When Europe revoked a patent on every GM soybean: questions and answers

In May 2007 a board of appeal at the European Patent Office in Munich revoked European patent 301 749. In its broadest form, that patent had covered every genetically engineered soybean, however it was made and whatever gene it carried. Campaigners celebrated. Others asked whether a 13-year fight won on technical grounds counted as much of a victory.

Seventeen years later, with the European Union stuck over whether gene-edited plants should be patentable at all, the case reads differently. These are the questions worth asking about it.

What did the patent claim?

EP 0301749, "Particle-mediated transformation of soybean plants and lines", was filed in 1988 by Agracetus, then a subsidiary of the chemical company W.R. Grace, and granted in March 1994. The title describes a method: coating microscopic metal particles with DNA and firing them into plant tissue with a "gene gun". The claims went further. In their broadest form they covered transgenic soybean plants and seeds as such, not only those produced with the gene gun, and the patent as granted reached beyond soybeans to other plants.

Monsanto bought Agracetus in 1996 and the patent came with it. Accounts of the case later pointed out an irony: before it became the owner, Monsanto had been among those objecting to the patent.

Why did a "species patent" alarm people?

Because a product claim does not care how you got there. A patent on the gene-gun process can be avoided by using a different method, such as Agrobacterium, the soil bacterium that became the workhorse of plant transformation. A claim to all transgenic soybeans cannot be avoided. Anyone in Europe engineering soy, for any trait, by any method, would need a licence from one company.

When the patent was granted, no GM soybean was on the market anywhere. The first herbicide-tolerant varieties went on sale in the United States in 1996. The claim therefore reached across an entire crop before the technology had been commercialised: a monopoly on a future. Agracetus had obtained a similarly sweeping patent on GM cotton, and India revoked its version in 1994 after public criticism.

Who fought it, and why did it take 13 years?

The opposition was filed in 1994 by RAFI, the organisation that later renamed itself ETC Group, and other campaign groups backed it over the years. By the time the case was decided, the opponents of record also included two of Monsanto's commercial rivals, Syngenta and Pioneer Hi-Bred. That is worth remembering. Broad patents hurt competitors as well as farmers, which is one reason they get challenged.

The timeline is the real scandal. The EPO's opposition division did not decide the case until 2003, when it kept the patent alive in a narrower form limited to soybeans. The case then went to appeal, and the board ruled on 3 May 2007 in case T 1165/03 (the decision). The patent, filed in 1988, had about a year of its term left. For nearly all of its life, it stood.

Was the revocation a technical win or a principled one?

Technical. The board revoked the patent because the claims Monsanto defended were not new; reports of the hearing also cited insufficient disclosure, the rule that a patent must teach a skilled person how to carry out the invention. The board did not rule that patents covering a whole plant species are wrong, or contrary to morality, or barred by the exclusion of plant varieties from patent protection.

It could hardly have done so. In December 1999 the EPO's Enlarged Board of Appeal had decided, in a case known as G 1/98, that a claim to plants is allowable as long as it is not confined to a single plant variety, even if it covers many varieties. Broad claims across a species are, in principle, fine. What sank the soybean patent was the ordinary test of novelty, which applies to every patent, not any new rule about plants.

That does not make the case pointless. It removed one very broad claim, it put the EPO's slowness on the record, and it showed that persistent opponents can win. But as a precedent on the principle of species patents its value was limited, and the sceptics who said so in 2007 were right.

Then what were the broccoli and tomato cases?

They were where the argument over principle actually took place, and it took another 13 years.

One patent covered a broccoli bred to contain more of a compound linked to protection against cancer; the other, a tomato bred to contain less water. Both plants came from crossing and selection, helped by molecular markers, with no genetic engineering at all. The European Patent Convention excludes "essentially biological processes" for producing plants. The question was what that phrase meant, and how far it reached.

• December 2010 (G 2/07 and G 1/08): the Enlarged Board held that crossing whole genomes and selecting plants is an essentially biological process, and so excluded, even when marker technology helps with the selection.
• March 2015 (G 2/12 and G 2/13): the same board held that the exclusion did not extend to the products. Plants, seeds and fruit could still be patented. A process you could not patent produced a tomato you could.
• November 2016: the European Commission published a notice stating that the EU's biotech patents directive was meant to exclude such products as well.
• July 2017: the EPO's member states changed the convention's implementing rules so that plants obtained exclusively by essentially biological processes are not patentable.
• December 2018: a technical board, in a case about a pepper, held that the new rule conflicted with the convention as interpreted in 2015 and refused to apply it.
• May 2020 (G 3/19): the Enlarged Board changed course and accepted the rule. Plants exclusively obtained by essentially biological processes are not patentable, though not retroactively: patents granted, and applications filed, before 1 July 2017 are unaffected (the opinion).

So are conventionally bred plants now safe from patents?

Mostly, but not cleanly. G 3/19 excludes plants obtained exclusively by conventional breeding. Plants whose traits come from a technical step, such as random mutagenesis or gene editing, remain patentable, and claims are often drafted around a trait rather than a method. If a conventional breeder finds the same trait in existing material, the line between "my variety" and "their claimed plant" can be hard to see. Patent attorneys add disclaimers to carve out natural occurrences. Campaigners argue that those disclaimers give breeders little real certainty.

The seed industry's own answer has been licensing platforms. Vegetable breeders set one up in 2014, and a platform for arable crops followed in 2023, offering access to members' patented traits on standard terms. These help. They also make access depend on joining a club whose rules the largest patent-holders helped write, which is the kind of arrangement a market dominated by a few firms tends to produce, as our survey of the top ten seed companies shows.

Why did patents become the sticking point in the NGT debate?

Because the proposed EU law on new genomic techniques creates a contradiction. Under the Commission's 2023 proposal, a "category 1" gene-edited plant is one that could also have arisen naturally or through conventional breeding, and is therefore treated like a conventional plant. Patent law treats the same plant as a technical invention. A breeder could be told by one law that a plant is conventional and by another that it belongs to someone else.

The European Parliament's answer, adopted in February 2024 and confirmed when it closed its first reading on 24 April 2024, was a full ban on patents for NGT plants, their material and their genetic information, as explained in our piece on the Parliament's NGT vote. Member states are split. Successive Council presidencies have floated softer alternatives, such as obliging applicants to disclose relevant patents, publishing them in a database and encouraging licensing on fair terms. As of November 2024 the Council still has no agreed position, and patents are the main reason.

Can the EU simply ban the patents?

Not with a stroke of a pen. The EPO is not an EU body; it applies a treaty whose members include countries outside the EU. The 2016–2017 sequence shows one route: an EU-level position on what the biotech patents directive means, followed by the EPO's member states amending the implementing rules. But NGT plants, unlike the broccoli, are clearly the product of a technical process, so an interpretive notice would not stretch that far. The directive itself would probably have to change, and that would reopen the same arguments about where "technical" ends that took the broccoli and tomato saga a decade to settle.

There is an alternative already on the shelf. Plant variety protection, the system described in our UPOV 91 explainer, rewards breeders while leaving others free to use a protected variety for further breeding. Many breeders say that is the model they would prefer for edited plants too.

What did the soybean case teach?

Two things. First, that a patent office can take 13 years to decide whether a claim was valid, by which point the claim has done most of its work. Second, that technical victories remove individual patents but do not change the rules. Changing the rules took a decade of political pressure, a Commission notice and an amendment adopted by the EPO's member states. The NGT fight is the same lesson again, applied to a newer technique and a bigger market.