Boffis
International Hazard
Posts: 1836
Registered: 1-5-2011
Member Is Offline
Mood: No Mood
|
|
The cyclisation of l-lysine to pipecolic acid
This idea has cropped up occasionally in several Science Madness threads over the years (1,2) usually as a step in the preparation of piperidine from
OTC compounds, however, these threads are universally lacking in experimental detail. These threads do, however, present references of which the most
pertinent seems to be two Japanese papers (3,4). The first by Fujii & Miyoshi (3) provides a concise description of the synthesis of l-pipecolic
acid from l-lysine, unfortunately the procedure given incorporates some very amateur unfriendly procedures such as hydrolysis with sodium hydride or
metallic sodium in liquid ammonia. The second paper by Yamada et al (4) in which the preparation of pipecolic from lysine is discussed as an
intermediate step in the synthesis of various alkaloids of the nicotine-coniine type provides a more amateur friendly route. The problem with these
two papers is that they are trying to make the synthesis enantiomer selective. For most amateur chemists, interested in the resulting pipecolic acid
as a precursor to piperidine or metal ligands the chirality is not necessarily important. As I was interested in pipecolic acid for other reasons,
essentially the preparation of metal ligands and as part of my ongoing investigations in the chemistry of the amino-acids I decided to investigate
these reactions with the intention of finding a simple route from easily available l-lysine to (racemic) pipecolic acid.
Other routes to pipecolic acid have been described but as far as I can see only the paper by Ladenburg (5) provides an amateur accessible route. In
this paper the author describes the preparation of pipecolic acid (piperidine-2-carboxylic acid) from picolinic acid (pyridine-2-carbxylic acid by
reduction with sodium amalgam. Given access to mercury and picolinic acid this is probably the simplest route, however, a route via more accessible
compounds would certainly be of interest to amateurs.
The routine from l-lysine described by Yamada et al (4) is the diazotization of l-lysine in excess hydrochloric acid to give 6-amino-2-chlorohexanoic
acid (not isolated) followed by alkali induced de-hydrohalogenation and cyclisation with barium hydroxide. After the precipitation of the barium with
sulphuric acid the remaining liquor was concentrated to a syrup. This syrup presumably contains much pipecolic acid and was used as an intermediate
without further purification or isolation. In the process of preparing various alkaloids this was esterified and tosylated etc with the purpose of
maintaining chirality.
The approach of Fujii & Miyoshi (3) was slightly different in that they tosylated the 2-amino group of l-lysine to protect it and then replaced
the 6-amino group by diazotization with sodium nitrite in the presence of potassium bromide and excess hydrobromic acid to give
6-bromo-2-N-tosyl-aminohexanoic acid, evaporation of the solution gave a syrup. However, they did not attempt to isolate this compound from the syrup
but instead esterified it directly by treating it with hydrogen chloride in ethanol.
The complexity of both routes appears to be driven by the desire to achieve a high proportion of one chiral enantiomer. However, if a racemic product
is acceptable, it should be possible simplify these procedures.
The basic reactions appear simple enough, unfortunately the practical aspects proved much more difficult. Using the basic procedure provided of Yamada
et al (4) as a guide I have tried several modifications and simplification based on the diazotization of l-lysine. In the first experiments the
l-lysine as hydrochloride was diazotized with sodium nitrite at room temperature or on gentle warming until no further nitrogen evolution occurred;
very little extra acid was added, just enough to make the solution strongly acid. The idea was that on warming the nitrite ions and lysine would react
to form a transient diazonium compound that rapidly hydrolyses on warming to give 6-amino-2-hydroxyhexanoic acid, which, on treatment with strong
caustic alkali would cyclise directly to the sodium salt of pipecolic acid. The free acid would then be recovered by rendering weakly acid with
hydrochloric acid and evaporating the solution until the solution was almost saturated with respect to sodium chloride and cooled. The high
concentration of salt would then salt-out the weak pipecolic acid. Unfortunately, this didn’t happen, a viscous pale straw-brown syrup was the
result from which I could not persuade anything to crystallise i.e. by trituration with ether etc. further evaporation yielded only salt crystals. On
evaporation to dryness on a water bath a light brown fudge like mass resulted, that looked rather unpromising. However, in the light of later
experience in which this type of product was almost always obtained but has now been partly resolved into the desired material these early experiments
may be worth repeating.
The next set of experiments used a scaled-up version of the Yamada ((4) page 627) procedure where the l-lysine was dissolved in excess hydrochloric
acid and diazotised with a near stoichiometric amount of sodium nitrite under temperature-controlled conditions to give a solution of
6-amino-2-chlorohexanoic acid. An attempt was also made to isolate the intermediate 6-amino-2-chlorohexanoic acid though without success, again the
result of an intractable syrup. As a result, this syrup was treated directly with either sodium or barium hydroxides in modest excess and refluxed for
an hour or so. When barium hydroxide was used the excess barium was removed by bubbling CO2 through then boiling to granulate the precipitate and
filtered. When sodium hydroxide was used it was neutralised with hydrochloric acid and the weakly acid solution evaporated down to a syrup from which
nothing crystallised. Further evaporation resulted in first the separation of copious salts (sodium and or barium chlorides) and ultimately the same
light brown fudge-like solid I had obtained before.
Ladenburg (5) states that the best way to recover the acid was via its hydrochloride. When the brown syrups were treated with concentrated
hydrochloric acid a white precipitate formed that is easily recrystallised from dilute hydrochloric acid to remove most of the co-precipitated salts
(particularly barium chloride). This latter problem was eventually remedied by evaporating down the syrups on a water bath to the fudge-like solid and
then leaching with warm methanol. The pure white salt residue can then be filtered off and the methanol solution evaporated to a thick syrup again.
The syrup was treated with about half its volume of concentrated (c 36%) hydrochloric acid and chilled in the fridge for a couple of hours and then
filtered through a glass sinter. The white solid was drained as dry as possible and recrystallised from as little water as possible, it is very
soluble in hot water, with the addition of a little hydrochloric acid. This is believed to be pipecolic hydrochloride but work is in progress on this.
The filtrate from the original hydrochloric acid treatment was evaporated down to about 2/3 of its original volume and more colourless crystals formed
slowly. In the hope that the syrup represents the amino acid that has polymerised rather than cyclised I tried to hydrolyse it by adding more conc.
hydrochloric acid and refluxing for an hour. The pale brown colour rapidly turned very pale yellow. After an hour refluxing the acid was largely
removed by evaporating on a water bath and the pale-yellow liquid cooled. After standing for 2 days in the fridge it appears to have finished
crystallising, depositing tiny colourless prisms and plates. These crystals are also under investigation but again appear to be pipecolic acid. A
small amount of less viscous liquid remains which seems to support the idea that the syrup contains a polymer but the last bit of polymer does not
seem to be readily hydrolysable.
When the possible pipecolic acid hydrochloride is treated with a little sodium hydroxide solution and evaporated to dryness, leached with methanol and
evaporated to a thin syrup it does eventually deposit colourless prisms or thick blades of what appear to be pipecolic. Crystallisation is very slow
but once started it can be chilled in the fridge for a week or so. When only a little liquid remains, the pale syrup can be drained from the crystals
using a small Buchner funnel and a gentle vacuum. The crystals are very soluble in water but can be recrystallised from it.
Work is still ongoing and this is only a preliminary report.
1) http://www.sciencemadness.org/talk/viewthread.php?tid=29248#...
2) http://www.sciencemadness.org/talk/viewthread.php?tid=4834#p...
3) Fujii & Miyoshi; A novel synthesis of l-pipecolic acid; Bull. Chem. Soc. Jpn. 48, 1341 (1975).
4) Aketa, Terashima & Yamada; Biomimetic conversion of l-Lysine into optically active 2-substituted piperidines, Chem. Pharm. Bull., v24(4),
621-631 (1976)
5) A. Ladenburg: About the piperidine carboxylic acids, Ber. Deu. Chem. Gesch, v24, p640-643, (1891)
|
|
|
AvBaeyer
National Hazard
Posts: 644
Registered: 25-2-2014
Location: CA
Member Is Offline
Mood: No Mood
|
|
Boffis,
Very nice! You are doing some very difficult work - hope to see more soon.
AvB
|
|
|
Boffis
International Hazard
Posts: 1836
Registered: 1-5-2011
Member Is Offline
Mood: No Mood
|
|
@AvBaeyer, thankyou for the encouragement! I have been sent by my company to the arse end of nowhere for a month but when I get back in a few weeks
I'll pick this up again. I have acquired some more lysine and will try two experiments at about 0.2 Mole scale. One via the hydroxy-aminohexanoic acid
and the other via the chloro-aminohexanoic acid to see if either route is preferable and see what the yield is like. It was impossible to estimate
yield from the experiments above because the intermediated products were sub-divided so much.
|
|
|
|
![[*]](images/xpblue/default_icon.gif){kind=icon}
