complished with Synergid in 95% yield.17 From an EH&S
point of view the Antibioticos process appears to be superior
to the Lilly one: The E-factors (considering the solvent
recovery) of the Antibioticos and Misner processes are 85
and around 200, respectively;18 the Antibioticos process has
more isolation steps, but the intermediate was always
recovered wet from water, thus decreasing the possible safety
and environmental drawbacks; the process avoids the use of
toxic gas, chloroform, and silica gel chromatography.
Table 1. Demethylation of quaternary saL 12: the thiol
comparison
12 nucleophile8 8 + 3
NMP, 75°C
nucleophile
CH3SNa
C6H5SNa
HO-CH2CH2-SNa
t
8/3
15
>20
12
97/3
90/10
98/2
Experimental Section
omethoxide reaction (entry 1), sodium thiophenol was less
selective, being the 8/3 ratio only 90/10 (entry 2). On the
contrary, the sodium salt of mercaptoethanol gave the best
results in terms of selectivity and reaction time (entry 3).
Additional advantages of mercaptoethanol versus thiometha-
nol are the consistently lower vapour pressure (0.13 versus
172 kPa at 20 °C) and the generation of a side product,
namely 2-(methylthio)ethanol, with a very high boiling point
(169-171 °C). The combination of these two factors
decrease the environmental impact of this step. Wet 8 was
dried by azeotropic distillation of water with CH2Cl2, and
after solvent exchange (CH2Cl2 f pyridine) methanesulpho-
nylchoride was added at 0 °C. After complete conversion,
mesylate 9 was precipitated by water addition and isolated
in 97% yield and >99% purity; dimer 15 was almost
undetectable. Wet 9 was dried by azeotropic distillation of
water with CH2Cl2 and, after solvent exchange (CH2Cl2 f
NMP), the mixture was cooled to 0-5 °C and 2 equiv of
commercially available sodium thiomethoxide was added.
Pergolide base 10 was isolated by water insolubilisation and
drying in 96% yield and >99.7% purity by HPLC. The final
salt formation was carried out in methanol/2-propanol to
control the selective formation of form I and pergolide
mesylate 1 was isolated in 92% yield and >99.8% purity.
The lab process (20-50 g) delivers the product with an
outstanding 81% yield, the industrial process at 10 kg scale
reaches 75-77% overall yield from 4. Among the solvents
used, dichloromethane, pyridine, methanol, and 2-propanol
were recycled. On the contrary, the recycling of NMP proved
to be difficult and not cost-effective, being in mixture with
water and contaminated by sulphur derivatives. This waste
was treated with bleach and burned. The use of dichlo-
romethane was compatible, the amount being very limited
and the solvent recycled; other solvents such as ethyl acetate
can be used as alternatives.
D-6-Propyl-8â-hydroxymethylergoline, 8. Into a stain-
less steel 800-L vessel containing NMP (40 L) under
nitrogen, were sequentially added 9.3 kg of 9,10-dihydroly-
sergic acid 4 (34.4 mol), 8.67 kg of sodium bicarbonate
(103.0 mol), and 30.4 kg of n-propyl iodide (178.9 mol).
The mixture was heated to 80 °C. After complete conversion
of 4 into 11, monitored by HPLC, the solution was cooled
to 45 °C. Five kilograms (45.1 mol) of CaCl2 and 6.5 kg
(172 mol) of NaBH4 were added and monitored by reverse
phase HPLC (the same method for all the analyses: Hypersil-
C18 4.6 mm × 250 mm column, 50/50 pH 7 phosphate
buffer/acetonitrile, 1.5 mL/min, 280 nm). After complete
conversion of 11 into 12, 20.6 kg of NaOH (514 mol) and
40.4 kg of HO-CH2CH2-SH (516 mol) were sequentially
added, and then the mixture was heated at 75 °C. After 12
h the conversion of 12 into 8 was completed, and 600 L of
water was added. The mixture was cooled to 5 °C and filtered
in a centrifuge and washed extensively with water until the
test for chloride was negative, affording 17.5 kg of wet 8.
A small portion of 8 was dried under vacuum at 60 °C
for 20 h for analytical evaluation. 1H NMR (DMSO-d6, 300
MHz) δ 0.85 (t, J ) 7.4 Hz, 3H, CH3-3′), 0.92 (t, J ) 12.3
Hz, 1H, CH-9ax), 1.3-1.6 (m, 2H, CH2-2′), 1.7-2.8 (m,
6H), 3.0-3.5 (m, 6H), 4.50 (t, J ) 5.2 Hz, 1H, OH), 6.75
(d, J ) 7.1 Hz, 1H, CH-12), 6.95 (s, 1H, CH-2), 6.98 (dd,
J ) 7.1, 7.7 Hz, 1H, CH-13), 7.07 (d, J ) 7.7 Hz, 1H, CH-
14), 10.57 (s, 1H, NH-1). IR (KBr, cm-1) 3648, 3352, 3095,
2958, 1606, 749. Mp ) 171.9 °C. Anal. Calcd for
C18H24N2O: C, 76.02; H, 8.51; N, 9.85; O, 5.63. Found: C,
76.25; H, 8.63; N, 9.75. Purity by HPLC >99%.
D-6-Propyl-8â-mesyloxymethylergoline 9. To a stainless
steel 800-L vessel containing CH2Cl2 (200L) under nitrogen,
was added from the previouse step wet D-6-propyl-8â-
hydroxymethylergoline, 8. The mixture was heated to reflux
to azeotropically remove water. When the moisture in the
suspension was lower than 0.1%, pyridine (150 L) was added
and the dichloromethane distilled under vacuum (GC in
process control). The solution was cooled to -5 °C, and 5.9
kg (51.5 mol) of methansulphonyl chloride was added in 1
h. After complete conversion, monitored by HPLC, water
(250 L) was added, and the mixture was filtered in a
centrifuge and washed with water (100 L), affording 26.8
kg of wet D-6-propyl-8â-mesyloxymethylergoline 9.
Conclusion
All the process design targets have been achieved. The
overall yield of the Antibioticos process, 75-77% (10 kg
scale) is consistently higher with respect to the Lilly one. In
fact, the yield of the Misner process considering 4 as the
starting material is around 67%, because the reduction of
dihydrolysergic acid 4 to dihydrolysergol 5 is easily ac-
(16) The use of the Na2S in the reaction of 8 was reported by Anastasia, L.;
Cighetti, G.; Allevi, P. J. Chem. Soc., Perkin Trans. I 2001, 2398. A 94/4
selectivity was achieved under different reaction conditions, sulpholane, K3-
PO4 (cat.) at 120 °C for 5 h.
(17) The Synergid process: Roletto, J.; Olmo, S. unreported results. Sandoz
researchers claimed but did not describe the reduction of lysergic acid with
LiAlH4.
(18) Sheldon, R.A. Pure Appl. Chem. 2000, 72, 1233.
Vol. 10, No. 2, 2006 / Organic Process Research & Development
•
201