Grantville Gazette 38

The New Royal Touch: Synthesis of Anti-TB Drugs

This "mini-article" is intended to supplement Brad Banner's "The White Plague" article by providing some hopefully informed speculation as to how, and how soon, various anti-tuberculosis drugs might become available in the 1632 universe.

The chemists in Grantville have access to only a minute fraction of modern chemical literature. I have tried to reconstruct what can be figured out from "Grantville Literature" alone; sources that I deem to be so available are marked with *, and maybes with **.

Mycobacteria are nasty; a cure usually requires treating the subject for several months. If you try to eke out a limited supply by using a short drug course, you will be breeding drug-resistant strains, not curing your patients. Moreover, since the critters are quick to develop drug resistance, it's de rigeur to treat with a combination of two structurally unrelated pharmaceuticals.

My proposed availability is based on a subjective evaluation of synthetic difficulty relative to the canonical DDT (June 1633), chloramphenicol (September 1633), sulfanilamide (winter 1633-34), hexachlorobenzene (same), and coal tar dyes (1634), given the limitations of Grantville Literature. See Cooper, Industrial Alchemy, part 3 (Grantville Gazette 26).

Production of Selected Drugs

Isoniazid. This was first synthesized in 1912, by two young chemists pursing dissertation research (Ryan 359). Isoniazid is a derivative of pyridine. Pyridines make up about 0.01% of coal tar (KO 20:661) and the coal tar pyridine derivatives include the 2-, 3- and 4-methyl picolines (*M amp;B 1083). A chemist would know that pyridines are bases (*M amp;B 1087) and would therefore consider purifying them from coal tar by acid extraction, neutralization of the acid with inorganic base, and fractional distillation or crystallization. Limited characterizing data (melting point, boiling point, refractive index, optical rotation, solubility) is available (*MI, *CRC).

Isoniazid is derived from gamma-picoline (*CCD 484). 4-methyl (gamma) picoline may be oxidized with potassium permanganate to yield pyridine carboxylic acid (isonicotinic acid), which has COOH in the 4-position. (*M amp;B 1083). If you convert COOH to CONHNH₂, you get isoniazid, a.k.a. isonicotinic acid hydrazide. (*M amp;B 1084).

So how can you do that? Isoniazid was first made from the methyl ester of isonicotinic acid (*G amp;G 1186). It may also be made from (a) isonicotinic acid ethyl ester and hydrazine (NH₂NH₂), or (b) 4-cyanopyridine and hydrazine. (*MI 586).

A chemist would know that a carboxylic acid can be esterified by heating it with the corresponding alcohol (methyl, ethyl) in the presence of concentrated sulfuric acid or hydrogen chloride. (*M amp;B 601). This is a reversible reaction, so if you want a good yield, you supply the alcohol in great excess, or find a way to remove one of the products, to force it forward. Alternatively, you convert the acid to the acid chloride and then the latter to the ester.

We then need to go from the ester to the hydrazide. It turns out that just adding hydrazine to the ester works fine. Elementary organic chemistry texts mention ammonolysis, in which an ester (-CO-O-hydrocarbon) is converted to an amide (-CO-NH₂) by reaction with ammonia (NH₃) (*M amp;B 603). Some texts mention that the reaction can be with other primary or secondary amines (*Solomons 822). Anyway, hydrazonolysis is the analogous reaction with hydrazine (NH2NH2). The reaction produces water so it helps to evaporate it off.

So, where does the hydrazine come from? If you react nitrogen with hydrogen, either nothing happens, or you get ammonia. The standard Raschig process (suggested in 1893 and demonstrated in 1906, KO 13:560) involves reacting ammonia, chlorine and sodium hydroxide, distilling the hydrazine monohydrate, and then eliminating the water by azeotropic distillation with aniline (*CCD 450). The chlorine and sodium hydroxide react to form sodium hypochlorite (13:578), which could be used directly. Aniline (aminobenzene) is a coal tar dye and should be available by nitrating coal tar benzene with nitric acid and then reducing the nitrobenzene. It may take a bit of fiddling to get the process working; it requires pressure and elevated temperature (120-150oC) to get a good reaction rate, and you need to eliminate metal ion impurities that might catalyze a side-reaction. (KO 13:575). **Eagleson (504) suggests carrying the reaction out in the presence of gelatin, which inhibits that side reaction and also catalyzes the desired reaction.

An interesting twist, not in Grantville Literature, is using urea in place of ammonia; capital costs are low and the method found favor in China (KO 13:580).

We could have isoniazid within a year or so after we first isolate gamma-picoline from coal tar. (Pyridines may also be obtained by destructive distillation of bone (CCD 694) but coal tar is a richer source (KO 20:661).) Figure 1634 or a bit later.

It's true that after 1950 it became more common to produce pyridines synthetically (KO 20:642) but I don't see that as a viable option in the 1630s. While the 1906 Chichibabin pyridine synthesis is in Grantville Literature (*MI 1152), it took half a century to commercialize (KO 20:661) and the commercial embodiments involve vapor phase reactions with proprietary catalysts.

Ethambutol. A symmetric aliphatic compound with a ten atom main chain and two types of functional groups. Heat ethylene dichloride with (+)-2-aminobutanol, or reduce 2-aminobutanol and glyoxal with sodium borohydride. (*MI 424).

Ethylene and chlorine react to form ethylene dichloride. 2-aminobutanol is obtainable by reduction or catalytic hydrogenation of 2-nitro-butanol (Merck Index), and nitrobutanol by nitration of butanol with nitric acid. Note the "(+)"; there are three isomers, and what you get is a mixture. The (+) isomer is the most active, and all are equally toxic (**Eagleson 383). **Remington (1663) says to resolve it "via its tartrate" and then condense with 1,2-dichloroethane (ethylene dichloride) in a dehydrochlorinating environment. Another route, from nitropropane and formaldehyde, is given in Sriram (491).

For glyoxal, oxidize acetaldehyde with nitric acid or hydrolyze dichlorodioxane (*MI 502). For sodium borohydride (a very useful reducing agent), heat methyl borate and sodium hydride (MI 955). For trimethyl borate, react boric acid with methanol (*CCD 896), and sodium hydride, react sodium with hydrogen (*CCD 802).

With two synthetic routes available, ethambutol might be available in R amp;D quantities as early as 1634, although later is more likely.

Pyrazinimide. Heterocyclic compound made by ammonolysis of methyl pyrazinoate, derived in undisclosed manner from quinoxaline (*MI 888). Unfortunately, this isn't a coal tar compound. The Gutknecht pyrazine synthesis is one of the organic name reactions generically described in the Merck Index 8th ed, p 1173, so that offers a clue as to a synthetic strategy. Also, *Eagleson 317 says that quinoxaline is synthesized from 1,2-dicarbonyl (i.e., glyoxal) and o-phenylenediamine.

A synthesis is set forth in Vardanyan (528), which is not available in Grantville. The first step is making quinoxaline per *Eagleson and then we oxidize with potassium permanganate, remove one carboxylic acid group with a radical initiator, esterify the other, and then ammonolyze. Even if we had Vardanyan to consult, the tricky part is coming up with the radical initiator.

**Remington advises preparation by "thermal decarboxylation of 2,3-pyrazinedicarboxylic acid to form the monocarboxylic acid, which is esterified with methanol and then subjected to controlled ammonolysis." (1663) This suggests that perhaps heat will do in place of the radical initiator; Sriram (490) implies that's correct! So, if we have Eagleson and Remington, we can rough out the synthesis, and we might have some product by 1636.

Thicetazone. This is a chemical of moderate complexity, a disubstituted benzene. Treat para-acetamidobenzaldehyde (I) with thiosemicarbazide (II) in alcohol (*MI 1036).

I) Para-acetamidobenzaldehyde: There's no derivation in Grantville Literature.

You can't just combine acetamide with benzaldehyde, you will get benzylidene-diacetamide (Watts 4). I have two plans to propose.

Plan A. First, we make p-acetamidotoluene. We start with toluene (methylbenzene), a coal tar component, and react it with nitric acid in presence of sulfuric acid to make p-nitrotoluene. (*M amp;B777). We reduce this with iron and hydrochloric acid to p-aminotoluene; reduction of aromatic nitros to amines is standard. (*M amp;B 728ff). We then acetylate the amine with acetyl chloride (*M amp;B 751). This shouldn't affect the ring since iron chloride is needed for Friedel-Craft acylation (*M amp;B 342). Actually, I know that this all works (**Johnson 423) but I figured it out before I looked it up! Acetyl chloride is made by reacting acetic acid (vinegar) with any of SOCl₂ (thionyl chloride), PCl₃, and PCl₅; for availability of these chlorinating agents see Cooper, Industrial Alchemy, part 2 (Grantville Gazette 25).

Next, I take the p-acetamidotoluene and add chlorine and heat, hopefully converting the methyl (-CH₃) group to -CHCl₂, and then hydrolyze with water at 100^oC to get the formyl (-CHO) of benzaldehyde. The -CH₃ to -CHO conversion works for toluene (*M amp;B 619) and I think it would work for p-acetamidotoluene, too.

Plan B. First, we make acetanilide (phenylacetamide). This was once used as an analgesic; the body metabolizes it into acetaminophen (Tylenol^R). It's derivable as follows: (coal tar -› nitrobenzene -› aminobenzene (aniline) -› acetanilide, the last step requiring acetyl chloride (*MI 5) or acetic anhydride (*M amp;B 742). Acetanilide is already, implictly, in canon, because it's an intermediate in the standard (undergraduate lab) synthesis of the antibiotic sulfanilamide.

Now, here's the trick. We use the "Duff Reaction" (*MI 1160), in which an aromatic amine is formylated at the para position by hexamethylenetetramine (hexamine, an oligomer of formaldehyde and ammonia) in the presence of an acidic catalyst. While MI assumes that the aromatic amine is a dialkylamine, I am guardedly hopeful that acetanilide (a monoacylamine) will react properly.

Does hexamine sound familiar? Dr. Phil made it back in 1633. (Offord and Boatright, "Dr. Phil's Amazing Essence of Fire Tablets," Grantville Gazette 7.) The preferred acid used to be boric or acetic acid (Ahluwalia 315); it's now trifluoracetic acid if yields of the para product are low. (**March 490).

II) Thiosemicarbazide. Made from potassium thiocyanate and hydrazine (*CCD 868). The "thiocyanate" is made by reacting sulfur with alkali cyanide (*C amp;W 301).

I think that we are looking at first availability in 1634-5.

Para-aminosalicylic acid. Heat meta-aminophenol with ammonium carbonate or potassium bicarbonate under pressure (*MI 62, *CCD 48).

The former may be obtained by reduction of meta-nitrophenol (*MI 89); I would reduce with iron and dilute hydrochloric acid (*M amp;B 725). To make m-nitrophenol, it's standard to boil diazotized meta-nitroaniline with sulfuric acid and water (*CCD 624; *MI 741, **Eagleson 700); treating phenol with nitric acid won't work as you get the para- and ortho isomers. Derivatizing nitrobenzene at the meta position is difficult because nitrobenzene is about 100,000 times less reactive than benzene.

You can make meta-nitroaniline from meta-nitrobenzoic acid (*MI 736). Or from aniline, by acetylation, nitration, and then removal of the acetyl group by hydrolysis. (*CCD 619). Note that the reagents aren't spelled out. Diazotization (*M amp;B 773) is standard in dyemaking and Stoner has certainly introduced it by 1634.

An alternative route to meta-aminophenol is "by reaction of alkali hydroxides with 3-aminobenzene sulfonic acid or from resorcinol and ammonia in the presence of catalysts". (**Eagelson 62). If we have Eagleson to consult, then we might have PASA by 1635.

Streptomycin. This was isolated in 1943 from a strain of Streptomyces griseus, a microbial fungus found is soil. Re-isolating it in the new universe is essentially a matter of chance; and the more samples, from diverse sources, are screened for the presence of antibiotic-producing fungi, the more likely it is that we will find one that produces streptomycin.

To give you an idea of what the odds are like, in 1946 Waksman noted that "the production of streptomycin . . . is characteristic of only a few strains of S. griseus," and that in a recent screen of 40 griseus cultures, none produced streptomycin and only one produced an interesting antibiotic.

While the chemical structure is known (Merck Index), devising a synthesis will likely be extremely difficult. It's an aminoglycoside, which is a chemical class several orders of magnitude more complex than anything reported to have been synthesized in canon. It has three rings, each heavily substituted. The first total synthesis of streptomycin was achieved in 1974 by Umezawa. It is unlikely that the chemical synthesis is described in Grantville Literature.

Conclusion

In medieval and even premodern times, it was believed that the "royal touch" could cure the skin disease scrofula, a swelling of the lymph nodes caused by tuberculosis. On one Easter Sunday, Louis XIV touched 1600 sufferers (White). But the chemistry of Grantville offers a surer solution to the "white plague."

Bibliography:

Grantville Literature:

(The specified edition is merely the one I consulted.)

[MI] Merck Index (8th ed. 1968).

[M amp;B] Morrison amp; Boyd, Organic Chemistry (2d ed. 1966).

[CCD] Hawley, Condensed Chemical Dictionary (8th ed. 1971).

[G amp;G] Goodman amp; Gilman, The Pharmacological Basis of Therapeutics (8th ed. 1993).

Solomons, Organic Chemistry (6th ed. 1996).

[C amp;W] Cotton amp; Wilkinson, Advanced Inorganic Chemistry (1972).

Maybe:

Eagleson, Concise Encyclopedia of Chemistry (1994).

Remington, The Science and Practice of Pharmacy (21st ed., 2005)(need to check pre-RoF edition!)

Johnson, Invitation to Organic Chemistry (1999)

March, Advanced Organic Chemistry (3d. ed. 1985).

Other:

Waksman, Isolation of streptomycin-producing strains of Streptomyces griseus", J. bacteriol., 52:393-7 (1946)

http://www.ncbi.nlm.nih.gov/pmc/articles/PMC518198/pdf/jbacter00673-0139.pdf

Umezawa, Total Synthesis of Streptomycin, J. Antibiotics, 27: 997-9 (1974)

http://www.journalarchive.jst.go.jp/jnlpdf.php?cdjournal=antibiotics1968 amp;cdvol=27 amp;noissue=12 amp;startpage=997 amp;lang=en amp;from=jnlabstract

Ryan, Tuberculosis: the greatest story never told (1992)

Vardanyan, Synthesis of Essential Drugs (2006).

Sriram, Medicinal Chemistry (2010).

Ahluwalia, Organic Reaction Mechanisms (2005).

Watt's Dictionary of Chemistry, Vol. 1 (1888).

White, A History of the Warfare of Science with Theology in Christendom (1896), chapter XIII

http://cscs.umich.edu/~crshalizi/White/medicine/fetich.html

[KO] Kirk-Othmer Encyclopedia of Chemical Technology, 4th ed., Vol. 13 (1995) and 20 (1996).

****