predator/pestlock.txt

file: pestlock.doc
Derived from File: Azadirac.doc (alpha version)
<modified 20/11/1999>


Bigger IS better : why it is harder to evolve resistance against a complex 
poison molecule than it is to evolve resistance against a simple one.
----------------------------------------------------------------------------

Since before the start of the 20th century, there's been an "arms race" 
between pesticide manufacturers and their new killer chemicals, and the 
target pests who eventually learn how to tolerate them. It always seems to be 
that these synthetics are hailed as a silver bullet, but soon enough the 
target organism learns to dodge it. Why might this be the case? And more 
pertinently what might be the solution?

This doesn't just happen down on the farm, either. It occurs at all biological 
scales. The physical size of the pest animal is irrelevant, since the war is 
fought at a molecular level. The wars are being lost : there's plenty of 
antivirals to which viruses are now resistant, bacteria which eat multiple 
antibiotics for breakfast and survive, fungi which are not killed by 
antifungal agents, insects which can happily metabolise insecticides all day 
long, and plants which manage to survive despite an onslaught of herbicides. 

(It is important that this happens. Some of the things we kill with our 
 nonspecific poisons are actually our allies, and we need every ally we 
 can get, but that's another issue.)

Many of the agents employed in the quest to kill various organisms are 
extremely effective in their initial application, but less effective with 
repeated use. All those drums of "Kill-O" in the shed which did great work
last year will underperform next year and be useless the year after that. 
Why? The pests literally engineer a way out. But how do they do it? Why
can they do it? How do we stop them?

To define this problem further we will have to go down to the molecular
arena where these battles are fought out, and first gain an understanding 
of what a poison actually does.

Enzymes, poisons, and the art of the evolutionary molecular locksmithing
------------------------------------------------------------------------
A useful aid to understanding the toxicological concepts without having to 
drown oneself in the agonies of biochemistry is to use an analogy. Most of us
have a bit of a familiarity with locks, and although the analogy isn't exact  
it can give you a good idea of what's going on.

Locks permit gates to be opened and closed by specific keys. In biochemistry
the gates have to open and close at specific times or, amongst other things, 
nutrients and raw materials can't get where they need to go. As in real life 
the the keys control the state of the locks, and the locks control the state 
of the gates. Enzymes often combine the "lock" and "gate" in the one, 
dual functional package. 

As with locks, in biochemistry, you can have the locks and keys set up in 
particular ways. If you have one gate and two locks in tandem, opening one
lock will open your gate even if the other lock is still locked. On the
other hand, you can have a gate with two locks in parallel, each on separate 
hasps, so you need to unlock both locks at the same time to open the gate.

In nature, although you will occasionally find a setup where only one lock in
several needs to work for the gates to open and close appropriately, the 
set-up is usually parallel, in the sense that all the locks must work or
the gate can't be opened and closed at the right times. 

There is one significant difference in biochemistry: you CAN'T change the 
keys, because the keys also happen to be very same nutrients and raw 
materials that the gate will permit through it! 

Locks are constructed a particular way, and will admit only certain types of
key - round keys on vending machine locks, U-shaped keys on Bi-lock locks,
your front-door lock takes a familiar brass Yale key into its keyhole. 
Then, of the keys that fit, then only the one with the right wiggles on it 
will open the lock. 

It's a similar thing with the enzymes which run living things. They are 
shaped a particular, specific way, will only let particular substances into 
their gaps and crevices, and they are very choosy. Just as you can't fit a 
round key into a lock with a U-shaped keyhole, you can't fit molecules into a 
given enzyme unless they are shaped just right.

Nature would prefer that she could open and close her molecular locks and 
biochemical gates as she sees fit. If she can't do it, certain gates are shut 
or open when they shouldn't be, so valuable things escape, or nutrients can't 
come in. Things die, simple as that.

It is useful to think of poisons as a kind of a dud key. Whereas normal keys 
enable you to open or close a door by unlocking or locking a lock, the poison 
key still fits the lock, but has to gum up the lock's working somehow so the 
gate can't be opened ever again, or is locked open when it should be shut, 
or whatever. 

Poisons look similar to the usual stuff a protein interacts with, but are 
different in some critical way which happens to ruin the protein. There are
many different interactions. To continue with our lock and key analogy,
it's as if a key has been filed in such a way that it jams against the pins 
and won't come out, kind of like a dynabolt: it changes once it is inserted 
so you can't pull it out again. This consequently means you lose control of
your gate - it is open or closed at inappropriate moments.

This sort of stuff happens when poisons interact with biochemical systems,
but nature can't change the keys!

It's worth noting that historically some locks were made with detector levers 
in them... enabling them to be easily `poisoned' or made unopenable. If you 
tried the wrong key, relockers were engaged and then NO key would open the 
lock, including the correct one. 

It seems now that a lot of our dud keys are in fact no longer jamming the
targetted locks. How do bugs get resistant to our dud chemical keys? 

Nature changes the locks.
-------------------------
Nature isn't conscious in the conventional sense. It doesn't say, "Hmmm,
yeah, if I rip off a chlorine atom here I can neutralise this poison."

Instead, routinely, nature's organisms make hundreds of slightly different 
versions of their locks - in this case, many versions of target enzymes in a 
pest's biochemistry. All of these will still perform their usual biochemical  
job, and most of these versions are messed-up by poison. However, because 
organisms have twenty different types of amino acids to play with, in each of 
several hundred positions in the target protein, they have an amazing range of 
lock versions to potentially construct, and chances are that they can come up 
with one which will still work with the original key, but which now won't 
admit the dud key (poison) which jams up the lock.  

The rate at which an organism comes up with a solution is related to a couple 
of things, mainly how flexible the organism's improvisational locksmithing is, 
and also how often the organism reproduces. Each member of the target 
species has a slightly different plan for their own personal locks, which 
still use the original key but varies in some other way, which might happen
to make it un-poisonable. Each new member gets a crack at accidentally 
inheriting the lucky new lock variety, which still uses the original key
but which won't be wrecked by the dud one. What this means is that the more 
often the bug species reproduces, the more bugs there are trying to figure
out what the work-around lock version should be, with each generation of 
surviving bugs. 

When this biochemical locksmithing problem is solved, the bug that solves it 
reaps an enormous benefit. It not only is it now immune to the poison key but 
almost all of its progeny have the design for the new locks encoded in their 
DNA - resistance is hereditary - so they are immune too. 

It all sounds wonderful, but there is a caveat. 

If the dud key is complex, and very subtly made to simultaneously interact 
with many parts of the lock, or worse still, interacts with many different 
kinds of locks at the same time, nature has a much harder time of it and has 
to devote serious, often unaffordable resources to build the new locks so it 
can run its biochemistry again. It is then that other approaches tend to be 
tried, such as systems which recognise dud keys and chop'em up, or which 
pump the dud keys out of the organism. 

It is here that the lock analogy breaks down a bit and we have to return 
into the real world for a little while. There is another analogy which will
be useful, but I'll get to that when I come to it.

Humans make simple poisons, nature makes complex ones. 
-------------------------------------------------------
So back to the molecular machinery of resistance in insects. Insects have
been under attack from many organisms for millennia, the most recent being 
h.sapiens, which fancies itself a bit of an organic chemist, but we're nowhere
near as clever as Nature at this molecular art. Humans have synthesised and 
sprayed all sorts of stuff around to kill insects, and other things. 

Maybe some of the names will be familiar... alachlor, aldicarb, aldrin, 
atrazine, benomyl, amitrole, 2,4-D, chlordimethiform, carbaryl, carbofuran,
chlordane, chlordimethiform, chlorvenifos, chlorpyrifos, chlorotoluron, 
cyclodiene, DBCP, DDT, dicamba, dieldrin, dicrotophos, dimethoate, disulfoton,
endothall, fenthion, glyphos, heptachlor, hexazinone, lindane, malathion,
mancozeb, monocrotophos, oxychlordane, paraquat, permethrin, primicarb,
simazine, thiocarb, trifluralin, zineb. 

You might notice a few sounds repeated. For example, chlor- means there
is one or more chlorine atoms in the stuff. It is interesting that halogens
don't show up very often in plant toxins. Phos- and fos- suggest a phosphorus
which is another atom which doesn't tend to show up in natural poisons either. 

You might notice a few sounds are repeated frequently. For example, 
chlor- appears several times. So does -phos, -azi, -thio/sulf. Thio and sulf 
imply a sulfur, which is another uncommon atom in plant poisons, unless you 
look at relatives of the onion and garlic familes which tend to use 
non-protein sulfur compounds a lot.  Pyr- suggests one of several rings with 
nitrogen and carbon in them. Carb- suggests a member of a family of the 
carbamate family. 

A lot of these chemical "Leggo-blocks" show up time and again in humanity's
artificial synthetic pesticides. 

There are others, but it doesn't matter that I omit them. I'm using the 
phonetic similarity in the names to illustrate a structural similarity in the 
pesticide molecules. If you looked at structural drawings of them, or even 
had to wrestle with their special chemical names, you'd see similarities 
there too.

The "dud" keys we use to jam nature's molecular locks have some commonalities.

They're simple, small and structurally fairly similar. Firstly, they 
generally aren't very big, as far as molecules go. Also, since they are made 
of heavy atoms, weight for weight, they aren't very complex compared to 
equivalently heavy molecules made of lighter atoms. Look at something like 
heptachlor - it's basically a loop of carbon atoms where molecular weight 
is gained by bolting on a few fat chlorine atoms. The molecule has a lot
of similar and simple branches on it. Which raises a third point: synthetics
often they tend to have similar and simple structural backbones. Our 
synthetic pesticides are all simple variations on the same themes, childish 
molecular Leggo structures compared with the amazingly complex pesticidal 
sculptures nature comes up with.

Complexity is determined by how much stuff you have to build with, and 
also how configurable all the bits are. You can only build so much with five
bits of leggo, but nature dictates that by doubling the pieces of leggo, you
get far far more than double the number of ways of putting them all together.
You can, weight for weight, get many more permutations and combinations out
of a given mass of "light" C, O, H and N atoms than you can out of the same 
mass of atoms like S, P, Cl and related "heavies". The total mass of the 
leggo is not the issue - it is the complexity of its configuration.

Some of the reasons for this are that humans simply haven't been doing 
chemistry for several million years and simply cannot cheaply make these
complex backbones which nature seems to do so easily and cheaply. So our 
approach is, yeah, let's synth this, then drown it in nitriles or halogens or
something else amenable to synthesis by the bulk chemical synthetic methods 
we humans tend to use.

In contrast, poisons plants make and use against bug attack are made naturally
and most of them are made out entirely of carbon, hydrogen, oxygen and 
to a lesser extent nitrogen. These elements are also the main ingredients in 
plant toxins with other atoms in them, like sulfur or bromine.

The reason for this is that probably N, P and S are environmentally scarce 
and metabolically not worth the price of manufacture for defense purposes.
Phosphorus is so rare and presumably so precious to the organism's energy 
(ATP) and information (DNA) metabolism, that it will not be allocated to 
other tasks, because these energy and information metabolism functions are 
so critical to the system that there would be a selection pressure against 
an organism that didn't allocate P only to these critical tasks. Same for 
sulfur, which is a critical component of many proteins but which is 
relatively rare in the environment. From a plant's point of view, compared 
to N, P and halogens, there's a stack of "cheap" carbon and oxygen around 
with which to build complex stuff, so the plant making a toxin to defend 
against attack is less pressured not to deplete these elements by using them 
to make defensive chemicals. 

On the other hand nature might just be better at complex carbon oxygen and 
hydrogen chemistry than she is at complex sulfur phosphorus and nitrogen 
chemistry. But that's not really central to the issue. The central issue is 
the complexity. 

Nature seems to rely more on taking whatever is lying around and building a 
really complicated pest-repellent molecule, instead of building heavy, but 
simple, molecule. The molecules which nature uses as pest repellents, if they 
are heavy, get this way by being complicated artworks of light atoms, rather 
than being structurally simple molecules with heavy atoms attached to them.

Simple vs Complex Dud Keys
--------------------------
So what? Why should the complexity of a poison matter? It's the interactions.

A large, complex poison molecule will necessarily interact with many parts
of its target enzyme at once. The ultimate poison key is something which 
interacts with a lot of the lock components and renders them useless, e.g. a
squirt of adhesive from a hot glue gun, all the way up the inside of the lock, 
will jam up that lock in a much more irreparable way, than will a wad of 
chewing gum stuck shallowly in the keyhole. 

Putting a bubble-gum shield on keyhole is easy: add-on a strip of teflon, and 
the gum can't stick to the lock, but you can still use the original keys. 

Compare this simple bubble-gum-repulsion problem, to the problem of 
redesigning a lock to keep liquid epoxy out of the keyhole, the broach, all 
the little pins and springs, and out of the surface where the lock barrel 
turns inside the lock  body- it's a screaming nightmare if you need to 
continue to use the existing keys, which demands that there remains a open 
hole in the lock through which the existing key (or the deadly hot glue) can 
be inserted. 

Hot glue is a hell of a poison for locks, because it gets intimate with so 
much of the guts of just about any mechanical lock you can build. Once inside
it forms a complex shape which happens to match all the inner surfaces of the
lock guts. To get around this, the design of the locks must be radically 
changed to keep the glue out. This change is so radical, it means you also 
need a kind of key which you don't have to actually insert into the lock. 

There are locks immune to hot glue. They lack keyholes and their key is a 
specially constructed blade of plastic, which contains embedded magnets.
The magnetic field passes through the wall of the lock directly, and needs no
keyhole. You can drown the magnetic lock in as much glue as you want but it
will still work. Magnetic locks are immune to destruction by hot glue guns.

The price we paid for locks immune to a hot-glue poisons, was thet we had to 
change not only the lock, but also change all the keys too, because all the 
old brass keys don't work in the new locks. When locksmiths first made 
magnetic locks they had to start using unfamiliar materials like plastics 
(they used to work with metals and ceramics) and they had to learn about 
magnetism, which was a considerable lot of new stuff to learn. The magnetic 
locks were expensive to construct because the tools needed to make them were
very different to the tools via which the usual metal locks were made. 
Of course, the new magnetic locks didn't work with all the old brass keys so 
they keys all had to be changed too.

But nature can't change keys, she is constrained to continue to build
locks which are susceptible to ruin by complex poisons. The very nature of 
the existing keys render the locks vulnerable to a complex attack.

This means, from an evolutionary point of view, that to get around a complex 
poison, MANY changes need to be made to the target enzyme, all at once. On 
top of this is the need to maintain the ability to use the existing key. This 
is a much bigger ask, just like the design of a lock immune to hot glue.

Each interaction adds itself to the list of problems which need to be solved
to enable the lock to work again, and they *ALL* need to be solved together.

It can take the target insects or plants (or whatever) decades, even 
centuries to solve such a problem - sometimes they don't ever solve the 
problem (basically they run out of time) and slide into extiction.

[An alternative strategy is the messing-up of more than one lock at the same 
 time. Sure enough, you find multiple toxins in the same plants. This is an 
 even bigger ask, because the pest has to evolve several new locks all at 
 once. Look at plants like barley, onions, horseradish, carrots, tomatos. 
 They have at least four phytotoxins in them. Look at the common spud, got 
 about 9 of them too. We usually get around them by cooking the food or 
 otherwise destroying the toxicity. Most pests don't do this.]

Well if nature is so smart, it probably knows that complex poisons are more
useful and give a better return on the biological resources used in their 
development. Does nature tend to use simple or complex poisons? What sort of 
pesticides do plants use against the bugs which suck their sap and eat their 
leaves?

Nature makes complex poisons
-----------------------------
The hypothesis that the pesticide companies would need be unable to falsify, 
in order to prove that their stuff is as difficult to get resistant 
to as the sort of complex agents nature has taken millions of years to 
patiently evolve, is that 

"natural complex pesticides exhibit the same resistance problems as our 
 simple synthetic ones."

I think the hypothesis has already been falsified anyway, however, in the
course of Nature's ordinary problem-solving. Nature presumably knows about 
resistance, after all, various organisms have been fighting chemical wars 
against each other long before we ever came down from the trees. The bacteria 
and fungi have, particularly, been fighting for aeons - we use the weapons 
that the fungi provide in our wars against bacteria, most of our antibiotics 
are derived from moulds and other organisms in the fungal realm.

If nature "thinks" big molecules are harder to get resistance too, then they
should be more common in her armament of poisons, than small and simple 
molecules. The payoff for designing a poison is then greater, because it
defends the designer for a longer period in evolutionary time. The payoff is
greater than the cost of developing it.

Nature also knows that it takes considerable effort to evolve these things, 
and tends to not go over the top by simply bolting on more complexity than
is absolutely warranted in keeping the pests guessing.

So what to expect? Well, few simple poisons, many complex poisons, and a few
really complex nightmares. Such a profile will reflect two things ... 

 1) nature CAN synthesise complex poisons against pests, when it is worth the 
    effort to prevent resistance over evolutionary time, and 
 2) will reach a plateau of complexity when the chemistry becomes too 
    metabolically expensive or synthetically intractable. 
    
It also has to be remembered that it does the defending organism no good to 
get poisoned by its own defensive chemicals, which further constrains its 
scope for engineering poisons against pests.

A rough guide, a fingerprint to look for, is the preponderance of carbon in 
the sorts of molecules which plants tend to use as poisons against various 
pests.

I happened to pick up an expensive book at a half price sale some years ago,
called the Dictionary of Plant Toxins. It happens to list in the back the 
molecular formulas of the molecules in the whole dictionary, in increasing 
numerical order, starting with the number of carbon atoms in the poison. 

Some of the molecules in this count are not toxic to things against which the
plant has had to compete - for example, there are plant toxins here which 
kill tumor cells in mice, and plants don't have to compete against mouse 
tumor cells. But most of these are toxins made to help the plant survive 
attacks by insects, fungi, parasites, plant viruses, bacteria, grazing 
animals, and even nearby competing plants.

I counted 'em up. What do we see?

# of Carbons : 1 2 3 4 5 6  7  8  9  10 11 12 13 14  15 16 17 18 19 20
Listed toxins: 2 5 2 9 6 16 14 25 15 51 51 36 34 51 169 80 78 52 66 114

# of Carbons : 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
Listed toxins: 75 68 28 21 17 16 35 10 34 32 17 25  8 13 19 21 10 12 5  9

# of Carbons : 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Listed toxins: 19  7  4  1  8 10  9  7  3  3  2  0  1  1  3  1  1  1  1  2

Summary: a number moderately simple toxins (less than 10 carbon atoms) 
   A hell of a lot of complex toxins (Between ten and forty carbon atoms) 
  Very few extremely complex toxins (more than forty carbon atoms)        

Pretty much what you might expect. It's a trade-off between effectiveness and
the molecular engineering difficulty associated with making a really complex
poison. Hey, YOU try and synthesise a complex molecule with 40 carbon atoms
in it, starting with sunlight, water and carbon dioxide! There is a bit of
bias in the low end, you just can't make much complex stuff with three carbon 
atoms. You can make plenty of things with five, and more with oxygen and 
nitrogen 
thrown in.

The data has been available for years for anyone to look. It probably has 
some sample biases (like, protein poisons are very complex but not hard to 
make) but I don't think this matters : it was just a bunch of plant poisons  
listed in a toxicological dictionary. It happens to fit what we might have 
expected if the evolutionary economics of natural synthesis of plant 
pesticides were subject to the sorts of trade-offs 1) and 2) outlined a few 
paragraphs above.

Ag-pesticide companies tell us they know their chemistry, we know they have 
business acumen. You might want accuse the pesticide companies of knowing 
this trend and deliberately only designing simple poisons so you have to go 
and buy another one when the last simple one you got became worthless due to 
the appearance of resistance. 

It's a kind of inbuilt obsolescence at the molecular level. It happens to 
benefit the chem companies that this is the case. But I never attribute to 
malice what can adequately be attributed to stupidity. In this case, it's
stupidity. We just don't yet know how to cheaply make really complex 
pesticides to which it is hard for the target organisms to get resistant.

Nature has, incidentally, solved the complexity-of-synthesis issue in a 
novel way : modularity. It knows how to synthesise twenty or so amino acids;
but since these amino acids can be daisy-chained by a single, uniform   
mechanism, it can make an unlimted number of possible proteins simply by 
bolting the amino acids together in different sequences. There is no need to
come up with new chemistry for each new protein, it is simply a matter of 
changing the order in which the well-known reactions occur. Like a Rubik's
Cube, you only have six colours to choose from, but depending on the way
you configure the cube you can have billions of combinations of colours, and
getting them is a simple matter of twisting the faces - any child can do it.
Protein synthesis still remains a fairly tricky feat of peptide biochemistry,
we generally employ recombinant bacteria to do it for us because it's 
something we humans just can't very easily or successfully do in a test tube.

I'm a synthetic organic chemist, and I know it is terribly, terribly hard 
to synthesise complex molecules. Its possible, but the cost in unwanted  
byproducts is just too much to make the final pesticide affordable. There is
another advantage. Biological poisons generally biodegrade, and don't 
become long term stable environmental contaminants like most of the 
organochlorines and organophosphates used in the last five decades. Throw in 
the requirement for biodegradeability and we're synthetically and 
economically pretty well sunk. By comparison, all of nature's poisons are 
ultimately biodegradeable.

So what to do? Use nature's chemicals against pests
----------------------------------------------------

I think the way of the future is clear - stop using simple synthetics and 
instead, extract complex pesticides from natural sources. Nature is a much 
better pesticide chemist than humanity, after all.

-Mike Carlton