=chemistry =explanation =chemical safety
This is a post covering some basic guidelines for estimating how dangerous a chemical is by looking at its molecular structure.
Anything that can react with an amine unspecifically is probably toxic! There are various ways this can be bad, but DNA alkylation in particular is worth noting because even a very small amount is bad.
Aliphatic carbon with a good leaving group alkylates things, including DNA. For example, diethyl sulfate is very toxic. Cl, Br, and I on aliphatic carbon, especially Br or I and/or activated carbon, can alkylate. For example, benzyl chloride, which is activated, is quite toxic. Many halocarbons can react with amines in cells and should be assumed toxic for this reason, but in most cases most of them will be metabolized as below. Methyl bromide is used for soil fumigation because it randomly methylates things in cells, killing them. Aromatic compounds with deactivating and leaving groups are also dangerous, such as parachloronitrobenzene (an intermediate in kevlar production) or bisdichlorophenylsulfone (used in polysulfone plastic). Deactivated heterocycles with leaving groups, such as atrazine, can also react with things, though atrazine is probably more problematic as an endocrine disruptor because the amino groups decrease the effective deactivation for the chlorine. Anything dangerously reactive that's in a plant for a while would probably react before getting eaten, though of course some chemicals can be metabolized to something more reactive only by enzymes in animals.
On the other hand, esters between most classes of safe chemicals are usually safe if the esterified acid is weak enough that your compound won't alkylate stuff. Ethyl acetate is fine, but diethyl oxalate is a dangerous alkylating agent.
Strong acylating agents (such as acetyl chloride and methyl isocyanate) are also dangerous. Note that chlorocarbons can be metabolized to more reactive chemicals. Chloroform can be metabolized to phosgene, which is quite toxic. So logically, LNT applies to chloroform exposure. (On the other hand, phosgene is not a persistent pollutant in the environment, because it reacts quickly with water.) Aliphatic chlorocarbon metabolization often involves replacement of a halogen and hydrogen with an oxygen. So, relatively low carcinogenicity of dichloromethane compared to other chloromethanes is because it becomes formyl chloride which decomposes to (relatively!) less toxic carbon monoxide and hydrogen chloride, while chloroform becomes phosgene and methyl chloride becomes formaldehyde. Correspondingly, brominated vegetable oil (added to some sodas but banned in some places) could be oxidized to a diketone, which is somewhat toxic - and if it isn't metabolized it could alkylate something.
In cells, alcohols are oxidized to aldehydes and ketones. Aldehydes are oxidized to acids. Reactive aldehydes can bond to amino acids and nucleotides. Acetone is somewhat reactive, ethanol becomes acetaldehyde which is more reactive than acetone and part of the reason for the effects of alcohol, and methanol becomes formaldehyde which is quite reactive and dangerous. On the other hand, acetaldehyde can be further oxidized (to acetate) while acetone cannot. Glyoxal and (because of paal knorr) succinaldehyde are also toxic; this is mostly why ethylene glycol and butanediol are to some extent dangerous.
Strong michael acceptors can alkylate DNA. So, while fumaric acid is safe because fumarate is safe, dimethyl fumarate is not safe, and acrolein is also carcinogenic, but there are some mechanisms to detoxify it, so acrylonitrile is worse. Yeah, acrylonitrile is pretty bad for you.
Cl, Br, and I on aromatic carbon tends to mean interference with hormone systems, and so compounds with that should be assumed toxic. Especially planar aromatic systems. Various other things can do this too, but generally not as much. Problems from dioxins are noticeable at part per trillion concentrations! That means you should avoid things (such as 2,4d) that even look like them, and things that can make even small amounts of them if you burn them. Cl on benzene can be converted to a hydroxyl group or hydrogen by enzymes, but not quickly.
This is the main problem with PCBs and to a lesser extent bisphenol A. Polybrominated diphenyl ethers, a common flame retardant additive in plastic, are worse than BPA and should not be used.
Anilines and phenols are often oxidation sensitive, with the oxidized radicals binding to stuff. That's bad. A phenol injection will kill people fast; IIRC the acute toxicity of phenol is mostly via para reaction. Electron withdrawing groups can prevent this oxidation to a radical, so aminobenzoic acids and hydroxyacetophenones are relatively safe. But aniline derivatives are generally bad. For example, diphenylamine, which has been added to apples, is not something you want to eat.
Anything that can produce a reactive radical by electron transfer to or from something in a cell is toxic. For example, nitrobenzenes should never be used in perfumes, obviously. They're a bad choice for industrial solvent, too.
Hydrazine is a bad sign. Examples of compounds with hydrazine include biurea (produced from azodicarbonamide, a flour additive) and Alar, previously used on apples but now banned. Toxicity of hydrazine is mainly via reactions that amines do, but amides can undergo enzymatic hydrolysis. (The good news is that biurea generally doesn't hydrolyze if you eat it, but I still sort of dislike it on principle.)
Unsaturated compounds - even benzene! - can be oxidized to epoxides by a p450 complex. This complex oxidizes hydrophobic compounds (because the active site is in a hydrophobic pocket) to form epoxides and hydroxyl groups. p450 reactions can also create hydroxylamines. Epoxides can then react with amine groups on amino acids and nucleotides. This is why benzene is toxic. How toxic something is by this mechanism depends largely on how easily it's oxidized to an epoxide, how reactive the epoxide is, and how easily the epoxide can then be hydrated by epoxide hydratase. Styrene is somewhat bad, but it's not nearly as bad as aflatoxin because it's more easily hydrated by epoxide hydratase. The second time benzopyrene is epoxified, epoxide hydratase doesn't like it; that's what makes benzopyrene so bad. Anthracene, on the other hand, is much less dangerous.
If I remember correctly, .18% of solid polystyrene and .02% of polystyrene foam is typically monomer; typical leaching from solid polystyrene leaves 450ppb styrene in milk; water absorbs about 1/10 as much, but other liquids can absorb more; typical leaching to hot coffee in a polystyrene cup for 15 minutes is 40ppb; typical taste threshold of styrene is 23ppb in pure water, 500ppb in milk. Solid polystyrene is often used to contain milk, yogurt, etc, for extended periods of time. Well, that much styrene certainly won't kill you, and styrene isn't as bad as a lot of other commonly used things, but my point is that LNT should be applied in many cases, effects are cumulative between chemicals, and that low levels of some chemical are still worth consideration.
p450 tends to oxidize aliphatic hydrocarbons next to the end. This is why hexane is relatively toxic for a hydrocarbon: it can be converted to a 2,5 diketone.
Various things can bind to metal ions in coenzymes. H2S, CO, and cyanide can definitely kill you! Basically, anything that can act as a ligand for a transition metal under mild conditions could be dangerous.
Some things, such as carbazole, can intercalate between DNA nucleotides, which causes mutations. That's bad. Carbazole is found in coal tar, which has been used for wood treatment. Planar molecules with positive charge (DNA being negatively charged) are relatively likely to be intercalating agents. They don't have to have positive charge on the whole molecule, either. The nitrogen in carbazole has some positive charge. Likewise with dibenzofuran. In acridine, the electronegative nitrogen leaves some positive charge on the other side.
Modified versions of key molecules, such as DNA bases, can be bad. For example, the chemotherapy drug cytarabine substitutes for a DNA base.
Some molecules inhibit enzymes. (For example, fluoroacetic acid is a metabolic poison, because it binds tightly to an enzyme that normally handles acetate, which is an example of how things can replace similar molecules in important processes, interfering with those processes.) Enzyme inhibition tends to be more of an acute problem than a chronic problem, especially compared to DNA alkylation, and acute toxicity is much easier to test.
Diethyl ether ("ether") is an example of an enzyme inhibitor. It can be broken down (like other ethers) to ethanol and acetaldehyde, and also (unsurprisingly) inhibits enzymes that would convert ethanol to acetaldehyde, as well as binding to similar enzymes as ethanol. So, ether produces similar effects to being drunk, and also increases the primary effects of alcohol, which are largely GABA receptor related, but not the effects of metabolites. (Logically, it shouldn't have any long term effects worse than ethanol, and a smaller amount of ethanol with a little bit of diethyl ether could produce similar effects to a larger amount of alcohol, with somewhat less of some of the effects people don't like. Science! I could actually see Japan or China legalizing substitution of diethyl ether for ethanol, what with the clash between their love of alcohol and Asian flush. Well, that was the composition of "Hoffmann's Drops" and ether was at one point somewhat popular in the west, but it lost popularity after 1920 because of temperance/prohibition movements.)
Still, inhibiting some enzymes can cause longer term problems. For example, thalidomide causes birth defects by blocking a enzyme that controls limb formation. Another example is trans fatty acids, which are problematic because some enzymes are specific to cis fatty acids, meaning trans ones disrupt fatty acid metabolism, and so if you eat a lot of them, disrupted fatty acid metabolism causes problems. Incidentally, the FDA reached an agreement with processed food producers that products must have “0 grams of trans fats” listed, but the amount of trans fat can be rounded down if it’s less than 1 gram per serving. At least people are finally starting to avoid hydrogenated oils, what with there being no actual point to using them instead of mostly saturated vegetable oils.
Some metal ions catalyze bad reactions. For example, lead. Chromium is toxic, but usually doesn't enter cells unless it's hexavalent.
Organometallic compounds should be assumed to be dangerous, and ones containing normally toxic metals are worse. Examples:
TEL - lead, gasoline additive now largely banned
MMT - manganese, gasoline additive
ferrocene - iron, gasoline additive
dibutyltin - PVC antioxidant; previously, organolead compounds were used
thiomersal - mercury, preservative used in vaccines
merbromin - mercury, topical antiseptic now banned in the US
roxarsone - arsenic, feed additive for chickens
Some ions can substitute for other ions, thus disrupting enzyme activity. For example, Cl -> F and Na -> Li. This basically acts as temporary enzyme inhibition.
Precipitation of things can cause problems. Fluoride can kill by precipitating with calcium as CaF2. Oxalic acid can precipitate as calcium oxalate, causing kidney stones. Ethylene glycol and glycolic acid can be oxidized to oxalic acid. Problems from melamine + cyanuric acid precipitation were in the news a while back. But this is a pretty rare issue, and apart from precipitation, the positive and negative ions of a salt can be considered separately. For example, sodium is safe, and glutamate (a natural amino acid) is safe, so monosodium glutamate is safe.
In some cases, toxicity of primary amines is related to toxicity of corresponding ketones/aldehydes, because enzymes can convert between the two.
Some reactions are less dangerous than others. In terms of safety,
sulfonation > nitration
ester formation > amide formation, ether formation
hydration > hydrochlorination
hydroxylation > amination
amine methylation > amine ethylation
For further reading, Hunter's Diseases of Occupations is decent.
back to index