Sciencemadness Discussion Board

Useful intermediates from amino-acids and related compounds

Boffis - 1-3-2014 at 17:28

This is a theoretical review of the conversion of amino acids to nitriles by oxidative decarboxylation. It is my intention over they next six months or so to actually try out some of these reactions and report accordingly. The reason for this post is to air the subject in advance to see if anyone can add constructive comments and ideas before I dive in and start mixing chemicals. I will try and find the references and link them with future edits, if not I will repost the articles.

The α-amino acids and several related peptides and their degradation products are now easily and cheaply available from health food and body building supply outlets and the purpose of this short review is to look at the possible conversion of these compounds into useful intermediates for further organic synthesis or as an end product in their own right. While there are many possible conversion routes available I have decided to consider initially just one route built around the oxidation of these compounds with trichloroisocyanuric acid (TCCA) because this compound is cheap, easily available and there are several published papers available that give experimental details for several amino acids; the following two papers have already been posted on the SM forum under other threads:

1. An insight of the Reactions of Amines with Trichloroisocyanuric Acid, Luca et al.; SynLetts; No12; pp2180-2184; 2004

2. Conversion of α-Amino acids into Nitriles by oxidative decarboxylation with Trichloroisocyanuric Acid; Hiegel et al., Synth Comms; v34: pp3449-3453; 2004

The oxidative decarboxylation and formation of a nitrile is niether new nor unique to TCCA, the paper by Hiegel et al. reviews some of the other reagents but most are either expensive, difficult for the amateur to source or require specialized condition. The beauty of TCCA is that it works in an aqueous medium, it is readily available and is cheap. The papers referenced above offer simple routes to several interesting compounds such as acetonitrile, cyanoacetic acid, malonic acid, malononitrile, succinonitrile, phenylacetonitrile (benzyl cyanide), 3-indoylacetic acid and 4-imidazolylacetonitrile for example either directly or via further conversions. The conversions of nitriles to free acids, amide, thiamides, tetrazoles, amidoximes (-C(NOH)NH2), amines etc. are well covered in the literature.

I have included in this discussion certain compounds related to the α-amino acids such as citrulline and ornithine because they poses the same reactive moiety as the protein yielding amino acids and I have also included proline, β-alanine and γ-aminobutyric acids because, while not strictly α-amino acids they are readily available and should respond to some of the reactions described in Luca et al.

The last three compounds do not contain the reactive α-amino-carboxylic acid moiety -CH(NH2)CO2H and so will react differently. β-alanine and γ-aminobutyric acids will probably react like primary amines and so yield a dichloramine initially and them a nitrile in the presence of a base give cyanoacetic acid and 3-cyanopropionic acid respectively (see Luca et al.). These could then be converted via the ester and amide to the corresponding dinitrile. The preparation of the cyanopropionic methyl ester is described in the paper by Hiegel et al. listed above and the conversion of esters to amides has been discussed many time before on this forum under other threads. The conversion of these amides into the dinitrile may be possible using the ammonium sulphamate process described by Boivin (J. L. Boivin; Canadian J Research; Vol 28 sect B; p671; 1951) which I believe has already been posted under another thread.

Proline is a unique case, being a cyclic azolidine and not having a free amino group it will not undergo the decarboxylation and nitrile formation of the other normal amino acids. It is possible it will react as a secondary amine, particularly in alkaline conditions (to prevent self protonation of the nitrogen) and therefore give a chlorimide that then loses HCl to give a dihydropyrrole-2-carboxylic acid of which there are 4 possible isomers, probably tautomeric. Three of these isomers would represent regeneration of the secondary amine and so allow further oxidative dehydrogenation to pyrrole-2-carboxylic acid, in theory! The partial loss of HCl from a secondary amine chlorimide is described by Luca et al.

Luca et al. investigated the reaction of TCCA with various primary and secondary amines as well as amino acids and their paper provides more insight into the mechanism of the reaction and the isolation of intermediates. They also report that the outcome of the reaction with amino acids depends on whether the free acid or the ester is used.

Luca et al. describe four synthetic schemes, those relevant to amino acids are:

Scheme 1: the reaction of a β or γ amino acid or an α amino acid ester with TCCA in methylene dichloride at room or below; Luca et al. give three examples:

1. Phenylalanine methyl ester to the dichloramine (methyl 3-phenyl-1-(dichloramino)propionate)
2. Valine methyl ester to the dichloramine (methyl 3-methyl-1-(dichloramino)butyrate)
3. γ-Aminobutyric acid to the dichloramine (4-dichloraminobutyric acid)

Then in Table 2 they report the treatment of the Phenylalanine dichloramine with triethylamine to give the corresponding chlorimine. In the text they described the hydrolysis of the chlorimine with HCl in tetrahydrofuran to give ketones; in this case the product would be methyl 3-phenylpyruvate (by extension of Scheme 2 but no amino acid specific details given). Valine esters should likewise give α-ketobutyric esters.

Scheme 4: The direct treatment of free α-amino acids with TCCA in sodium hydroxide solution at 5°C to give the nitrile, five examples are given:

1. Phenylalanine to Phenylacetonitrile (Benzyl cyanide)
2. Aspartic acid to Cyanoacetic acid
3. Glutamic acid to 3-Cyanopropionic acid
4, Tyrosine n-butyl ether to 4-n-Butoxyphenylacetonitrile
5. L-isoleucine to (S)-2-methylbutyronitrile

Heigel et al. describe the direct conversion of α-amino acids to nitriles by TCCA in aqueous solution in the presence of pyridine. They give a summary of seven such conversions of α-amino acids and details of two representative preparation, L-isoleucine to (S)-2-methylbutyronitrile and Glutamic acid (Hiegel incorrectly refers to Glutaric acid but clearly means Glutamic acid as the former is nitrogen free) to cyanopropionic acid recovered as the methyl ester. The later ester is a useful intermediate for the preparation of succinonitrile.

In general the yields reported by Hiegel a significantly lower than Luca et al (60-70% as against 90%+), this coupled with the use of pyridine renders their procedure less desirable in theory but I would caution against the near quantitative yields of Luca et al. until I have tried them out. Triethylamine or similar tertiary amine can probably be substituted for pyridine.

Luca et als.' method described at the end of their paper is worth a brief discussion. Assuming the reaction to proceed as follows:

3C<sub>6</sub>H<sub>5</sub>CH<sub>2</sub>CHNH<sub>2</sub>CO<sub>2</sub>H + 6NaOH + 2C<sub>3</sub>N<sub>3</sub>O<sub>3</sub>Cl<sub>3</sub> &rarr; 3C<sub>6</sub>H<sub>5</sub>CH<sub>2</sub>CN + 3CO<sub>2</sub> + 2C<sub>3</sub>N<sub>3</sub>O<sub>3</sub>H<sub>3</sub> + 6NaCl + 6H<sub>2</sub>O

Luca's 1.2g of phenylalanine would require 0.58g of NaOH (7.25ml of 2N NaOH) and roughly 1.13g of TCCA. From the reported procedure they added 1.17g of TCCA so a slight excess but used only half the amount of sodium hydroxide solution so a significant amount of hydrochloric acid will be produced during the reaction as well as carbon dioxide. They then add much more HCl (15ml) and 2.5ml of 3N hydrochloric acid. I presume that the 15ml is concentrated i.e. 36% acid. Its purpose is not clear but it may be to dissolve the by-product NaCl and cyanuric acid prior to the ether extraction, the 2.5ml of 3N acid seems a bit irrelevant but may be to simply maintain the water content sufficient to dissolve to salt.

The NaOH added is just sufficient to convert the phenylalanine into its sodium salt. This means that the oxidation takes place essentially under acid conditions in which you would not expect the development of hypochlorite ions suggesting that the mechanisms of oxidation is via direct interaction of TCCA with the amino acid rather than via hypochlorite development.

When I try this reaction, as I intend to do, I may try a modified procedure and neutralize the excess HCl produced and then extract with a solvent such as methylene dichloride in which NaCl and cyanuric acid are practically insoluble. Luca et al. do not state how they recovered the more polar and soluble products like cyanoacetic acid. For these compounds it may be best to use methanol as the solvent as Hiegel et al does or to vacuum evaporate the neutralised reaction mixture, leach with methanol and filter off the cyanuric acid and salt and then esterify the product directly. Alternatively the cyanoacetic acid (and possibly cyanopropionic acid) may be extracted with methyl isobutyl ketone (see US pat 2338834). Some reactions should yield cyano-amides e.g. glutamine should give 3-cyanopropionamide, I can't find solubility data on this substance but it is likely to be sparingly soluble in cold water particularly if there is a lot of salt present but soluble in organic solvents such as alcohol; so a possible route may be neutralization, evaporation to dryness and extraction with alcohol.

The hydroxy and mercapto bearing nitriles may be difficult to prepare in reality by this technique because of the tendency of α-hydroxynitriles to split into hydrogen cyanide and an aldehyde under strongly acid conditions or polymerise to an oxazoline derivatives under basic conditions (see Arrhenius et al., JOC, 1997, v62, pp5522-5525). Hydrogen cyanide and carbon dioxides are also the theoretical products of glycine and TCCA; this may be the easiest way to OTC HCN but the separation from the CO2 without poisoning yourself could be tricky; possibly absorb into a calcium hydroxide suspension and then filter off the precipitated chalk, treat the filtrate with alkali carbonate to precipitate the calcium and then use the alkali cyanide immediately as an aqueous solution or alternatively it could be distilled off and condensed to liquid (HCN Bp 25°C CO2 Bp -57°C) (for the very brave and/or suicidally insane).

Cysteine will probably suffer oxidation to the disulphide, cystine, before decarboxylation. This reaction may still be useful as the disulphide can be reduced back to the mercapto group. Another possibility that I am looking into is the reaction of cysteine with 1,2-dibromoethane in the presence of excess NaOH to give a di-cysteine substitute dithioglycol and a precursor to sulphur bearing ligands.

Phenolic amino acids such as L-tyrosine and L-dopamine may need the OH moiety protecting to prevent oxidation of the phenolic component; indeed in Luca's paper he gives the example of tyrosine as the n-butyl ether. Presumably these are prepared by treating tyrosine etc. in two or three molar equivalences of sodium hydroxide with the appropriate alkyl halide.

In most cases it does not matter which form is used (L or D) as the resulting nitriles are not chiral except in the case of a few acids with two chiral centres such as isoleucine, this situation is covered in both of the original papers.

The following list is a list of the theoretical direct product of the reaction between an α-amino acid and TCCA in sodium hydroxide solution as described by Luca et al.

The anticipated products are:
Glycine - Hydrogen cyanide - HCN; see notes above
Alanine - Acetonitrile - CH3CN - cheaper to buy acetonitrile
Leucine - 3-methylbutyronitrile - C3H6(CH3)CN
Isoleucine - 2-methylbutyronitrile - C3H6(CH3)CN
Valine - 2-methylpropiononitrile - C2H4(CH3)CN
Lysine - 5-aminopentonitrile - H2NC4H8CN
Norvaline - Butyronitrile - C3H7CN
Arginine - 4-guanylbutyronitrile - (CN3H4)C3H6CN
Aspartic acid - cyanoacetic acid - CNCH2CO2H - potentially one of the most useful as it gives access to this compound with the use of alkali cyanides.
Glutamic acid - 3-cyanopropionic acid - CNC2H4CO2H - a useful precursor to succinonitrile & succinic acid
Asparagine - Cyanoacetamide - CNCH2CONH2 - precursor to malononitrile and fulminuric acids
Glutamine - Cyanopropionamide - CNC2H4CONH2 - better precursor for succinonitrile
Serine - Hydroxyacetonitrile (glycolonitrile) - HOCH2CN - see notes above but precursor to oxazolines and aminoacetonitrile derivatives (EDTA etc.)
Threonine - 2-hydroxypropiononitrile - HOC2H4CN - possibly precursor to 2-aminopropiononitrile
Cysteine - 2-mercaptoacetonitrile - HSCH2CN -almost certain to be oxidized to the disulphide unless the mercapto group is protected first
Methionine - Methylthioacetonitrile - CH3SCH2CN - no protection need here - possible precursor to sulphur bearing ligands?
Phenylalanine - Benzyl cyanide - C6H5CH2CN - useful precursor with a reactive methylene group
Tyrosine - 4-hydroxybenzyl cyanide - HOC6H4CH2CH - OH may require protection
Dopamine - 3,4-dihydroxybenzyl cyanide - (HO)2C6H3CH2CN - OH will almost certainly require protection
Tryptophan - 3-Indolylacetonitrile - C8H6N.CH2CN
Histidine - 4-Imidazolylacetonitrile - C3N2H3.CH2CN - possibly some interesting ligands from this compound i.e. Imidazolyltetrazolylmethane
Proline - uncertain - see notes above
Ornithine - 4-aminobutyronitrile - H2NC3H6CN - precursor to γ-aminobutyric acid
Citrulline - 4-ureamidobutyronitrile - H2NCONHC3H6CN
Theanine - N-ethyl-cyanopropionamide - C2H5N.CO.C2H4CN - hydrolysis to ethylamine and succinic acid
β-Alanine - Cyanoacetic acid - CNCH2CO2H - see comments for aspartic acid
γ-Aminobutyric acid - Cyanopropionic acid - CNC2H4CO2H - see comments for glutamic acid

Oopsy_daisy - 2-3-2014 at 00:59

Sounds very interesting especially since many aminoacids are available as a nutritional supplement. Eager to see your results.

Chemosynthesis - 2-3-2014 at 16:33

Very interesting. I have been tempted to try synthesizing pipecolic acid from lysine through 1) n-nitrosation, 2) diazotization, and 3) (oxo)deamination. This is a biomimetic pathway, as opposed to inorganic catalysed hydrodeamination.

The reaction is relevant to histochemical staining and free amino nitrogen determinations for protein, as well as some carcinogenic etiologies. Citations are, of course, available, but I feel this is common knowledge in cytochemical staining and the medical community, respectively, as well as readily found on any internet search engine, though I am happy to provide as well. Essentially, I would be treating dissolved lysine in water with a nitrite salt through the addition of HCl with stirring, ala Yamada from Chemical and Pharmaceutical Bulletin (Tokyo) 24(4):621-623 (1976), waiting approximately thirty minutes, neutralizing the resulting pipecolic acid, and extracting into a non-polar phase. Van Slyke apparently used a mL of octyl alcohol as a detergent to prevent foaming. In theory, it'd be fun to say one also played around with the Ziegler alcohol synthesis of the latter, but that's far beyond my ambitions. I suppose it's worth noting both the acid and decarboxylated base are pharmacologically/medicinally chemically relevant, though I get enough of that sort of thing at work.

I have yet to look up the actual Yamada paper itself, but am familiar with the earlier work of Van Slyke (1913), which diazotized both amino groups of lysine (1. Percival Hartley's The Free Amino Nitrogen of the Different Proteins of Ox and Horse Serum, 1915 & 2. Barnett Sure and E.B. Hart's The Effect of Temperature on the Reaction of Lysine with Nitrous Acid (1917)).

Edit: PMID's 19109018 and 4122696 as well as page 20 of Critical Literature Review of Nitrosation/Nitration Pathways
Dr. William Mitch, Yale University (link immediately below) showing a cyclization mechanism whereby the lone pair of the alpha amino group attacks the deaminized carbocation provided for convenience.

[Edited on 3-3-2014 by Chemosynthesis]

AvBaeyer - 2-3-2014 at 19:27


You need to read the experimental section of the Yamada paper carefully regarding the synthesis and isolation of pipecolic acid. It is not nearly as straightforward as you imagine, ie., extraction of the product into a non-polar solvent.


Chemosynthesis - 2-3-2014 at 20:11

Ah, thank you. Admittedly, I had only read the abstract at the moment, but I was under the impression that most of the laborious aspects of purification arose from the need for stereochemical characterization. I was hoping to avoid chromatography with fractional crystallization and/or vacuum distillation at the most, similar to the fractional crystallization of substituted pipecolic acids from Eur. Po(vm. J. Vol. 19, No. 10/11, pp. 1055-1065, 1983.

I'll add links to the articles since I can't find the "upload" button to add them to my post.

I appreciate your bringing this to my attention.

Edit- Gave it a brief readthrough and am not yet disinclined. The synthesis seems accurately summarized, and I hadn't yet discussed purification other than the addition of a nonpolar solvent. This would imply I would store the freebased carboxylic acid in solution, which I would not intend to do.

In particular, there is this line "One report also concerns with the preparation of D-10c [pipecolic acid] from L-5c [lysine] using nitrosyl bromide as a deaminating agent and barium hydroxide as a base for cyclization. In these attempts, since formed cyclic a-imino acids (D10) were purified by recrystallization, precise estimation of the optical activity, which survived during the successive deamination and cyclization and deamination, could not be achieved." Page 623 of Yamada.

Imino acid is used archaically here, and their reference is given as U. Scheidt and H.G. Hoss, Hoppe Seyler's Zeit. Phys. Chem. , 308, 179 (1957)

I did not mean to imply I was looking to avoid any type of workup, and had hoped my use of "eventually" before the reaction synopsis would convey that. I'm still under the impression a fairly straightforward recrystallization may be achievable.

As long as I made sure that during my addition of base, the pH was above the pI of lysine (which I would use in XS), I should have minimal lysine contamination in my nonpolar phase. Even if non-negligible quantities of volatile n-nitrosopipecolic acid formed, the option of esterification upon subsequent purification would prevent it from forming a carboxylate. I'm largely discounting polymers due to the stability of ring formation and lack of mention in the lit I'm reading, but I understand thermodynamics and kinetics are very different.

So, what am I missing? Thanks.

Oh, aside from definitely having misread my copy of Purification of Laboratory Chemicals. Should have looked at that asterisk before benzene. Maybe I'd re-acidify after cyclization, getting the pH near whatever my calculated pHp of pipecolic acid is, which should be relatively near the pKa at any given concentration for total insolubility. That looks like the ballpark of ~2.5, and the pI of proline, the most similar amino acid to pipecolic is ~6, which are far enough under the ~9.75 pI of lysine to separate the two. The n-nitroso derivative, if formed, would be removed with a nonpolar wash, and is volatile anyway. I suppose the easiest way to start in this instance would be to remove lysine first by adjusting the pH to largely precipitate it. The ethyl pipecolic acid ester freebase is a liquid, and separable from lysine and any precipitatable salts. If the decarboxylated base were desired, I would probably not bother and reflux in acid, continuously distilling it off.

[Edited on 3-3-2014 by Chemosynthesis]

Cyphonical - 26-11-2015 at 13:28

With regards to the first post, there seems to be a few errors. TCCA decarboxylates Amino acids to a nitrile with one fewer carbon atom therefore the claim that Glutamic acid decarboxylates to 3-Cyanopropionic acid seems off, we're going from a five carbon molecule to a three carbon molecule would we not be looking at 4-Cyanobutanoic Acid as the primary product as the result of a reaction between TCCA and Glutamic Acid? Am I missing something?

Boffis - 26-11-2015 at 15:47

@Cyphonical Yes I thick you are missing something. Glutamic acid contains 5 carbon atoms in a chain, decarboxylation therefore reduces this to 4. The "cyano" radical contains one carbon so if you use the term "cyano" the residual molecule contains one less carbon atom i.e. 3 in this case therefore the parent compound is the C3 carboxylic acid i.e. propanoic acid. The "nitrile" radical term contains only one nitrogen no carbon so does not "shorten" the chain for naming purposes; so.

Consider the following examples:
methyl cyanide = acetonitrile
propyl cyanide = butyronitrile
cyanoacetic acid has 3 carbon atoms and is the half nitrile of malonic acid (C3H4O4) so...
cyanopropanoic acid has 4 carbon atoms i.e. one less than glutamic acid.

DrMethyl - 2-12-2015 at 11:47

Very nice OTC method thanks for your work !

Do you have the condition of the reaction please ? Have you tried yet ?

clearly_not_atara - 2-12-2015 at 14:01

Methionine - Methylthioacetonitrile - CH3SCH2CN - no protection need here - possible precursor to sulphur bearing ligands?

I've spent a while thinking about ways to use the sulfur atom on methionine as an odorless thioether. Unfortunately, this simple idea won't work.

If NCS will oxidize a thioether it's almost certain that TCCA will do so as well. The expected product is either methylthioacetonitrile sulfoxide or the sulfone. Another possibility is to decarboxylate methionine and use succinic anhydride to form N-(3-thiabutyl)-succinimide.

Or do the Strecker degradation with a selective reagent like alloxan or dehydroascorbic acid to form 3-thiabutanal.

Glutamine - Cyanopropionamide - CNC2H4CONH2 - better precursor for succinonitrile

C'mon, you of all people should know that TCCA will cause a Hofmann rearrangment in reaction with glutamine. The intermediate isocyanate will react with the NH2 to give propylene urea, or possibly a pyrimidin-2-one. Asparagine likewise gives ethylene urea or 2-imidazolone.

3-cyanopropionic acid

Under the wrong conditions I'm sort of afraid this compound will eliminate to HCN and acrylic acid. Better to err on the side of caution.

[Edited on 2-12-2015 by clearly_not_atara]

Boffis - 5-12-2015 at 07:57

@clearly-not-atara; thank you for your comments, they certainly raise some interesting and valid points. Sorry about the delay in responding but your comments warranted some research on my part. My initial response was going to be “fair cop Gov, I surrender” but then … wait… what’s this I see …. Yes a possible escape route….

My original idea was driven by a theoretical extension of the cited papers, to be honest I would be surprised if they all worked as shown.

In the case of methionine I had rather assumed that the methylthio ether group would be fairly inert and I had never heard of the Corey-Kim reaction. However, after looking into this reaction I am not so sure that it would occur as you think. On the wiki page you linked to, close to the bottom of the page there is a remark to the affect that substitutes for the rather obnoxious dimethyl sulphide were being sort and then gives an example of a long chain aliphatic thioether. If methionine could be subject to this reaction I think it would have been tried already as it is probably the most readily available and benign thioether. Furthermore, looking at the mechanism of the reaction, would you not expect the presence of the strongly electron withdrawing cyano group to interfere with this reaction by reducing the polarity of the electron pair on the sulphur (assuming the Wiki reported mechanism is right)? I am not a theoretical chemist (I’m a mining engineer) so I might be wrong but I’d be interested in your thoughts!

As to the Hofmann degradation of the amide group in the cases of the glutamine and asparagine, yes I should have spotted that one but didn’t. The outcome of this reaction therefore will depend on how fast the two competing reactions occur. It may be that condition will dictate the outcome, it will be interesting to experiment with this one. I have all of the reagents just need time now!

However, as I discussed above, the general procedure given by Luca contains insufficient alkali to neutralise all of the HCl produced in the reaction so this reaction take place under essentially acid conditions, it appears that the purpose of the alkali is simply to get the amino acid into solution. I was under the impression that alkaline conditions were required to favour Hofmann degradation. In Hiegel’s paper they use only weakly basic condition (effectively pyridine buffered) and the reaction is rapid, admittedly with a slightly different substrate, would you expect Hoffmann degradation competition under these condition either. Clearly experiments are required! However, even a hydro 2- imidazolone or a 2-pyrimidone-4-carboxylic acid would be interesting products.

I mean to try the glutamic and aspartic acid degradation but rather than esterify the resulting cyano-acids I am going to try a different work up. I intend to remove the cyanuric acid by filtration and then distil off the organic phase, take up the residue in water and extract the cyan acids with a ketone as per the industrial procedure. I don’t think that HCN production is an excessive risk give that the acids are stable enough to esterify by the Fischer method.

What do you think? Barking up the wrong tree or just barking mad?

clearly_not_atara - 5-12-2015 at 17:10

If methionine could be subject to this reaction I think it would have been tried already as it is probably the most readily available and benign thioether.

I suspect that if the sulfur atom of methionine were chlorinated, it would react with the nitrogen atom to form 1-methylisothiazolidinium-3-carboxylic acid (a 5-membered ring), possibly followed by another rearrangement such as methyl transfer from the sulfur to nitrogen to form uncharged 2-methylisothiazolidine-3-carboxylic acid. On the other hand, methionine derivatives, while not too difficult to prepare, are probably not as easy as

octanol + H2S >> octylthiol
octylthiol + Me2SO4 >> methylthiooctane

or a similar sequence, as the ingredients here are petroleum-derived. Chemical manufacturing is somewhat different from lab-scale chemistry. Even the methionine-derived thioethers may be stinky. Cysteamine certainly is.

I was under the impression that alkaline conditions were required to favour Hofmann degradation.

The first step is a geminal dehydohalogenation of an N-monochloroamide, so maybe. But maybe the purpose of alkali is to hydrolyse the isocyanate? Not sure. In any case the amide itself doesn't change conformation across the normal range of pH so I expect it to react.

[Edited on 6-12-2015 by clearly_not_atara]