How king cobra venom became a toxin cocktail

Start with a fraction. In a 2013 study of the king cobra genome, about two thirds of the toxin transcripts in the venom gland, 66.73% to be exact, came from a single family: the three-finger toxins. (A transcript is a working copy of a gene, the kind a cell reads to build a protein.) That is a lot of eggs in one basket, and it is a clue to how snake venom evolved. Venom is not one poison. It is a cocktail, and the recipe was written over millions of years by copying genes, changing the copies and putting them to work in one gland.

Where did the ingredients come from? One answer says spit: venom as modified saliva. Another says genes were drafted from all over the body. The king cobra (Ophiophagus hannah), which the Reptile Database describes as the longest venomous snake, at up to 5.5 meters, is one of the best-studied examples, so this article follows both stories, one that starts in saliva and one that borrows genes from many tissues. We flag where a result rests on a single animal, where scientists disagree, and where the evidence comes from other reptiles. This article is not a guide to handling king cobras or any venomous snake. They should be watched from a distance.

Meet the king cobra

The Animal Diversity Web (ADW), a University of Michigan database, places the king cobra from northern India east to southern China, south through the Malay Peninsula and east to western Indonesia and the Philippines. It lives near streams in dense or open forest, in bamboo thickets, near farmland and in mangrove swamps. ADW gives an average length of 10 to 12 feet, with a maximum of about 18 feet, and the Reptile Database says up to 5.5 meters, which would make it the longest venomous snake. The genus name, Ophiophagus, means snake-eating, and ADW confirms that the diet is mainly other snakes. ADW also says that in a threat display the snake can raise the front of its body about one meter off the ground.

That last row is worth a note. The Reptile Database says population genetics point to four main lineages that may be separate species, and National Geographic says the animal was recently reclassified into four species. The Reptile Database page does not say whether it should be split. The IUCN and GBIF records we use below are for Ophiophagus hannah as one species, and this article treats it that way.

A cocktail, not a poison

What is in the cocktail? The best look at one animal comes from Tan and colleagues in 2015, who studied a single adult king cobra, about 190 cm long, caught near Seremban, Malaysia. Its venom gland made 23 toxin families, and the venom itself held 45 toxins grouped into 16 protein families. Three-finger toxins were the biggest group, at about 43% of the venom’s protein, followed by snake venom metalloproteinases at 24.4%. The authors link the neurotoxins to rapid paralysis of nerves and muscles, and the metalloproteinases and some other enzymes to local tissue inflammation and damage.

One animal cannot speak for a snake that lives across half of Asia. Tan and colleagues compare their results with data from Balinese, Thai and Chinese king cobras and report differences in toxin sequences and amounts. From a mouse test, they argue that antivenoms made against the monocled cobra, a different genus, are a poor match, and they call for one that covers the whole region.

Here is a puzzle that shows how careful the numbers must be. Tan’s team found three-finger toxins made up 84.9% of the venom gland’s toxin transcripts, but only about 43% of the venom protein, and the two measures did not line up well across the families. The 2013 genome study reported 66.73% for transcripts in another animal. Different animals, different methods, different answers. All three agree on the main point: this one family dominates.

Two ideas for where the ingredients came from

The 2013 genome paper, by Vonk and colleagues, reports that toxins were recruited from non-toxin genes found in many tissues. It names two routes: a gene is hijacked and modified, or a gene is duplicated, and then the copy is switched on mainly in the venom gland. The three-finger toxins show what that looks like: at least 21 loci (places in the genome where a gene sits), a large expansion with much duplication and signs of positive selection, which means changes that were kept because they helped.

Some hints come from the controls, not the toxins. The most abundant microRNA in the venom gland, miR-375, is a typical pancreas microRNA, and the venom gland’s overall microRNA profile resembles known pancreas profiles. The authors say this is consistent with a gland whose controls were borrowed from the pancreas, which is a suggestion, not a proof, and it rests largely on one microRNA. A second gland, the accessory gland, tells a different story: lectin genes are the top toxin-like family there, at least 40 times more active than in the venom gland, yet no lectin turned up in the venom. The authors suggest the lectins do not help envenom prey.

The second idea is the spit one. In 2014, Hargreaves and colleagues argued that the older “recruitment” story was treated as fact on thin evidence. Their alternative is restriction: a gene is already active in several tissues, including the salivary gland, then gets duplicated, and one copy keeps its broad job while the other is restricted to the venom gland. They sequenced five reptile species and found that many proposed toxin families, such as complement C3 and nerve growth factor, are active in the salivary glands of non-venomous reptiles. They describe venom as a modified saliva in which typically 6 to 14 gene families have expanded. The limits are large: very small samples, pooled data, and only one venomous species. The authors say more species and tissues are needed.

These two papers do not fully agree, and neither closes the question.

How old is venom?

A separate question is when venom first appeared. In 2006, Bryan Fry and colleagues reported venom toxins in monitor lizards and in iguanians, groups not known to be venomous before, and found that all lineages with toxin-secreting oral glands form one branch of the family tree. They called this strong support for a single early origin of the venom system in lizards and snakes. The venomous lizards they mention include the Gila monster. The 2014 restriction paper mostly asks how toxin genes arose, not when venom first appeared, and our sources do not settle whether venom arose once or several times.

What this looks like in practice

Here is the king cobra venom from the Tan study, by share of venom protein. Treat it as a snapshot of one animal.

Common mistakes and myths about snake venom

Mix-up: venom is one chemical. In the animal Tan studied, it was 45 toxins in 16 protein families.

Myth: the king cobra has the most potent venom. National Geographic says it does not. Size is its record, not potency.

Mix-up: venom is an invention with no past. Both camps in the sources describe it as built from existing genes, whether from saliva, the pancreas or other tissues.

Mix-up: every venomous snake kills the same way. The Western diamondback rattlesnake has long, tubular fangs, according to ADW, while the Reptile Database describes the king cobra’s genus as having short fangs. ADW describes the king cobra’s venom as a strong neurotoxin that affects the brainstem’s breathing centers. The Boa constrictor uses no venom at all and kills by cutting off blood flow.

Mix-up: one species, one venom. The regional differences above show one name can cover several recipes, and the taxonomy may yet change.

Conservation status

The king cobra is listed as Vulnerable on the IUCN Red List of Threatened Species. That is the category as distributed by GBIF, the Global Biodiversity Information Facility, and we checked it on November 9, 2026. Vulnerable means the species was evaluated and meets the criteria for a threatened category, with a high risk of extinction in the wild if conditions do not improve.

The Animal Diversity Web names habitat destruction and persecution by people who fear the snake. National Geographic adds heavy deforestation in Southeast Asia, harvest for skin, food and medicine, and collection for the international pet trade, and states a decreasing population trend. We did not see the IUCN assessment text itself, so we repeat that trend as National Geographic’s report. If the species is split, the category may change, since the current record covers it as one.

Categories are reviewed and can change, and the copy GBIF distributes can trail the IUCN website by some months, which is why we print the date we checked. GBIF’s copy does not give the year of the IUCN assessment, so we do not state one.