For sake of metaphor, I want to include a plant that bears fruit that is poisonous, with the intent that when an animal eats it, it will walk away and die within a few hours, such that their corpse fertilizes the growing plant.
It seems there would be a rather strong evolutionary pressure for animals to avoid this fruit, so I'm looking for ways to mitigate this.
The Fruit Contains an Opioid
There's already a plant compound that is both highly addictive to numerous animal species and kills millions per year.
The fruit could be stronger than a typical Opium Poppy. Its ingestion causes addiction in the animal who eats it. First ingestion comes with a string high, with them returning because they crave the high and the painful withdrawals.
This eventually causes them to eat more and more until they eventually overdose and die with seeds still in their gut.
It wouldn't be the first plant to live off of corpses, but usually they find the corpses instead of making them. Here are a few ideas for biological strategies.
In addition to animals avoiding it, there seems to be a continuity conflict with an animal that both walks away and fertilizes the plant.
Have the plant look almost exactly like other plants. This works the wrong way in the viceroy butterfly, who looks exactly like the monarch, but isn't toxic. Birds avoid both of them. Maybe even a rare species of a healthy plant.
Seasonal toxicity: Yummy and healthy most of the time, but randomly toxic due to a chaotic metabolism. Now you only have to figure out how to kill something quickly. The standard fictional answer is "neurotoxin."
Blowdart plants. An old staple of fiction, like with the Borgia plant in Star Trek TOS. Or, even better razor-grass. Strike that, you specified fruit.
real 3. Anything that goes through the digestive system usually takes 20 minutes to be absorbed, so maybe you want it to go the other way. Maybe it has a strongly addictive quality that makes animals hang out around it subconsciously, and be less aware when they do it. Let's further say that some animals know that the plant is a good source of easy prey. The plant just lives off of the carrion -- it doesn't care who does the actual killing.
The poison won't actually kill the eater when the fruits are eaten in the right amount, but only produces an intestinal irritation which causes a laxative effect, leading to the expulsion of both the seeds and a conspicuous amount of fertilizer.
The killing only happens when the eater exceeds in eating of those fruits, resulting in massive loss of fluids and alteration of the saline equilibrium of the body.
This mechanism ensures that more animals eat the fruit and spread the seeds in various location, instead of a single animal concentrating the seeds in a single place.
For the animals there is a benefit in eating small amounts of the fruit because of the laxative effect.
The plant/tree can't kill every animal... to do so would make them extinct and its descendants wouldn't have fertilizer within a few centuries.
It needs to only kill maybe one in fifty (if even that many).
More so, when it kills them, it needs to kill them quickly. They shouldn't even have the time to wander more than ten feet away. This ensures that the animal dies over the top of where the plant's root are.
All that is needed is a mechanism where one or two fruits per season are poisonous. Plant toxins can be powerful, ricin for instance. Animals that die after eating the fruit still die infrequently enough that other animals watching may not make the connection between the feeding and the death, so long as the fruit aren't visually or fragrantly distinct from the safe varieties.
If the environment is such that few other sources of food exist, all the better. Even those animals that might make the connection may be forced to gamble and play Russian roulette with these things.
Assuming the plant bears fruit seasonally you could have it run on a cycle.
Yearly the plant produces bunches of safe to eat, sweet fruit with no / infertile seeds inside. Then, when the plant senses a shift, perhaps more of its fruits are being eaten Or some time has passed. The plant produces a far smaller crop of the same fruits with the toxin in them.
Assuming there's many of these plants in an ecosystem, animals would go from relative abundance to scarcity in a season. Making them more likely to eat the poisonous fruit.
There are plants with specific neurological effects.
Say your plant damages the brain of male victims in a specific way that causes the victim to progressively obsess on having sex, more and more, devoting all their efforts toward reproduction. Rape, murder, it doesn't matter! In a day, before the seeds have been expelled as waste, the poison kills them. But have a much greater chance of having impregnated a female to reproduce their vulnerability to the fruit of this plant.
It isn't so much that the animals learn and pass down avoidance of poisonous plants, but that those naturally dislike and avoid the smell or taste or more likely to survive and reproduce than those that like it.
So reverse the script, consuming the plant increases the chances of reproduction, by sending victims on a mission to reproduce before they die. It can affect women in the same way as men, making them sex crazed, but they obviously cannot reproduce before eliminating the seeds. Men can.
As others have mentioned mimicry is possibly one way it would work. But even then it would probably only be plausible within some vary strict parameters. This is because as noted previously there is strong evolutionary pressure on 'prey' species to detect and avoid toxic food sources. So that said I would suggest the following;
Any of these strategies severely limits the mimic. So the solution is to be rare. Then if the mimic copies a common, widely distributed food source but reproduces slowly or in limited numbers the pressure on the prey species to adapt is also limited and the plant can go on 'doing it's thing' without risk.
Also smaller animals are more numerous than large ones so the plant automatically raises the chances of having it's seeds eaten by selecting for animals it has the greatest chance of encountering i.e. the most numerous. Selecting smaller animals also maximizes the amount of nutrients it can recover from the corpse as it decays. If the plant only needs X amount of nutrients why select to kill larger animal Y when the excess remains will simply be wasted. It's not efficient.
Finally once it's been eaten and has killed it's host and started germinating it would help if the seed gave off a clear chemical signal don't eat this warning to other animals so they don't scavenge the body - and it's much easier to poison a smaller corpse in its entirety than a very large one. If the poison is slow acting and spreads throughout the body before death occurs the 'do not eat' message (perhaps the scent of the toxin) will spread at the same time.
EDIT: one last special condition I just thought of: The plant should perhaps be found in swampy low nitrate soils. These soils generally provide little of the nitrogen essential for plant growth and it's where you find the majority of the worlds carnivorous plants like Venus Fly Traps etc. This would perhaps justify a plant evolving to kill in the way described because it's seeds would germinate in a rich source of nitrogen, assuming all other factors remained unchanged.
I think what you want isn't so much about poison as it is about the strategy to make the poisoning system work. Given enough evolutionary pressure, defenses will evolve against anything. So don't make the pressure too high, and have a carrot along with the stick.
Your plant is a mimic, attaching itself to other plants and using the other plant for support (like grafting an apple branch of one variety to another). It has berries that look identical to the host plant. This would work even better for something like a neurotoxic fungus that could transform normal berries into secret poison berries (a little like ergot). An animal that eats one, eats the other. When the animal dies, it actually fertilizes the HOST plant's seedlings, while the graft plant grows along for the ride on the new host plant.
This way, the animals eating the host plant are eating healthy food (except for that one berry...). The host plant is actually benefiting from the arrangement since the host plant is the one most fertilized (although their seed-bearing species are killed in the process). Predators eating the dead berry-eating animal may potentially die themselves and provide even more opportunities for fertilizer (and reducing pressure by predators on the seed-bearing species).
The fruit is sweet; it can even have an addictive property as other answers have mentioned, to ensure enough of it is ingested by its victims.
The thing is, the seeds of the fruit are non-digestible, and when eaten they get "attached" to the digestive system (the stomach wall, the intestines) and begin to grow there, getting nutrients from both/either the body itself and/or other food the victim eats.
Eventually, the plant grows enough to completely block/replace the digestive system, killing its victim after a few days/weeks; then the plant begins to grow and feed from the corpse.
Active Parasite: What if the seedling is active-- it's more like a nematode-hunting fungus than a plant at the early stage. Once it gets eaten, it comes out of hibernation and quickly sends shoots/hyphae that digest the animal from the inside out. These take a few hours before they begin seriously affecting the host animal, perhaps even days, but soon the unsuspecting host will have tendrils of digestive roots crisscrossing their body, dissolving tissue and eventually killing the animal.
Once the animal dies, the hyphae roots continue digesting the animal until the plant erupts from the animal and unfurls its first leaves. Once the initial digesting has finished, the plant ends the active phase of its life and becomes a more passive photosynthesize.
There is a problem with this strategy, however. Dead animals in the wild are often eaten by other animals very quickly. That means the parasite plant must act quickly. Ideally, it only kills the host once it has spread through most of it's body, such that when the animal dies, the plant can monopolize on the nutrients and energy of the animal before it gets picked apart by a scavenger.
The toxin is always present. Young eaters have gut enzymes though that nutralize this toxin which older members lack. Over time their diet or increasing age causes them to no longer be produced. Within a few hours after after eating this fruit without this safe guard in your gut, the toxin is absorbed within the intestines and kills the host via neurological shutdown.
It may have to be a day or so to not make the immediate connection to the source of danger, but it would allow a population to continue thriving on it, with the occasional seed dispersal here and there.
The Pacific yew does not make taxol - but its endophyte Taxomyces andreanae does. Taxol is not tremendously effective as a poison, being mostly noted for killing cancer cells. However, it could be replaced by some more toxic poison, such as an opioid. We could also select the fungus to be able to infect a more nutritious plant.
In this instance, the fungus produces the potent toxin with the figurative "intent" that the eater will die from the poison, and fertilize the plant, which will become infected with the fungus. The fungus therefore promotes its own survival. It can also infect new plants directly. Would-be victims may include some addicts who learn to transplant the bushes - while this defeats the fatal "intent" of the fungus in the short term, in the long term it expands the number of plants and hence the number of victims.
The fungus would not compete if it lost its toxic product. The plant would replicate less if it came to resist the fungus, since addicts would no longer transplant it. Even in the absence of addicts, it would face an arms race with continued infections by ever-evolving strains of fungus, while resistance would be selected for only on the basis that if many plants are infected, herbivores might stop spreading them. As a result, some rate of endophyte infection would persist indefinitely.
Make the plant only kill males. Females are immune. You can lose a load of stupid males, providing females and some % of males survive.
If females cluster around the trees because they are a great food source, it'll balance out the propensity to avoid that fruit.
When an animal species finds a fruit to be toxic it can either adapt by not eating it, or by becoming resistant to the toxin; so, the ideal vector of attack for your plant is not voluntary ingestion. But there is a type of fruit that uses animals involuntarily to spread their seeds: burs.
Burs are a kind of dry fruit covered with velcro like spines that hitch rides on animals by clinging to their fur. Most animals find burs unpleasant and try to avoid them, but instead of advertising their presence like edible fruits, these fruits are small and try to camouflage themselves among other plants; so, even animals that try to avoid them still often find themselves covered in these unpleasant little hitchhikers.
The normal strategy for burs is to hitch a ride on an animal until the animal becomes aggravated with its presence, and then picks the seeds out of its fur (usually with its teeth). If you were to make the bur excrete a powerful enough toxin, then when the animal goes pick them out, it could poison the animal killing it before it can remove all the seeds.
Because getting burs is involuntary, host and plant species can evolve in equilibrium in something similar to a predator/prey relationship. The more animals evolve to spot and avoid the toxic burs, the more the plant evolves to hide its spiny little fruits.
Some plants take a VERY long time to produce a viable seedling, like a redwood waiting for a forest fire to drop seeds, and many seeds are viable for many years anyway. Perhaps your seed doesn't kill anything. The fruit is eaten, the seed attaches, either through burs or some initial early-stage growth that may be more just a chemical reaction between something in the hull reacting to stomach acid than actual "growth", and then it waits, embedded in the gut. When the heat drops to a certain level, or the amount of digestive chemicals drops below a certain threshold, indicating a natural death of the organism, then it begins to grow.
Consuming a large number of the fruit may kill creatures anyway, simply through gastric blockage. Some seeds may come loose, and may even still be able to barely germinate in the droppings, but nothing as reliable as the fertilization of a corpse. And the seed has the chance of propagating miles away from its host plant, since the animal may travel for years before dying.
For humans in some especially "green" society, it may even be a common near-death ritual to help the environment. Avoid the plant most of your life, but when you're dying anyway, eat a fruit, enjoy it's wonderful taste, and serve as host to the next tree.
From an evolutionary standpoint, for it to evade being out-evolutioned, it would need to not affect reproduction of it's prey.
So maybe it makes sure to circumvent exactly that. The poison works as a very potent aphrodisiac and causes those who eat it's fruit to go into a mating frenzy. Then, either by being actual poison, or by starving out the prey (because it doesn't eat anymore, it does other things), it kills and can use the body as fertilizer.
The survival of the prey species is not diminished if only one eats the fruit so it will probably not evolve to avoid the plat. It may even form a mutual parasitism, where each needs the other organism to procreate.
It's a symbiotic relationship.
The plant and animals both have a yearly lifecycle because the winter is very, very harsh--so harsh that pretty much everything dies over the winter unless it's in some protected state (like a seed or an egg case).
The plant is only seasonally poisonous. During the spring/summer/early fall it provides an absolutely ideal habitat for the young animals (maybe even mandatory like monarch butterfly larvae and milkweed). Sometime in the fall the animals mate and lay eggs which they carry around with them. Then in the late fall the plants start becoming poisonous.
The adults (carrying their eggs) happen to eat the poisonous plants just before the time when they burrow deep, deep into the ground to create a safe place for the eggs. When the timing is right, the poison kills the animal still in the burrow next to the eggs, in which case the eggs and seed have a nice, safe, comfortable winter under ground.
Then in the early spring the plant starts to grow. A few weeks later, the eggs hatch and the young have a perfect home to live in all year until fall.
Now you have a bunch of evolutionary pressures keeping this stable...
The plants need the animals to carry their seeds away and dig down, so only individuals who have a thriving nest of the animals living with them will be able to reproduce. And the seed poison needs good timing (based on digestive time of the animal) to make sure that the animal and the seed and the eggs are all underground together when the animal dies. If the seed is not poisonous enough then the animal might bury the eggs but return to the surface before dying, if it's too poisonous then the animal doesn't have time to burrow. If the seeds are too early in the year then the animal hasn't laid the eggs yet, too late and no animals are left to eat the seeds.
The animals who's young start off with their own plant habitat are able to grow faster and out-compete the ones that didn't, so the animals have evolutionary pressure to keep eating the seeds even though they're poisonous (plus they provide good nutrition for digging that burrow to protect the eggs). There's no evolutionary pressure against eating the seeds because the adults are already reproductively done anyway.
On earth some animals lay their eggs in other animals to provide nutrition (parasitic wasps) and some animals commit matriphagy -- the mother's body provides the nutrition for the young. This is sort of a combination, the adult's body provides nutrition for the plant which then provides nutrition for the young.
