Synthesis development of an aminomethylcycline antibiotic via an electronically tuned acyliminium Friedel-Crafts reaction

Article abstract of DOI:10.1016/j.tetlet.2008.08.011

With the goal of improving the synthetic efficiency, the development of a convergent synthesis of a minocycline derivative PTK0796 via an intermolecular acyliminium Friedel-Crafts reaction (Tscherniac-Einhorn reaction) is described. The entire C9 neopentylaminomethyl side chain was installed in one step using an electronically optimized chloromethylacyliminium precursor in 83% yield. Deprotection and re-equilibration to the C4 α-epimer in the presence of CaCl₂ and ethanolamine or NaOH afforded the target aminomethylcycline antibiotic. The corresponding crystalline tosylate salt was found to exhibit improved solid state stability.

Full text of DOI:10.1016/j.tetlet.2008.08.011

Tetrahedron Letters 49 (2008) 6095–6100
Contents lists available at ScienceDirect
Tetrahedron Letters
journal homepage: www.elsevier.com/locate/tetlet
Synthesis development of an aminomethylcycline antibiotic via
an electronically tuned acyliminium Friedel–Crafts reaction
*
John Y. L. Chung , Frederick W. Hartner, Raymond J. Cvetovich
Process Research, Merck Research Laboratories, PO Box 2000, Rahway, NJ 07065, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
With the goal of improving the synthetic efficiency, the development of a convergent synthesis of a min-
ocycline derivative PTK0796 via an intermolecular acyliminium Friedel–Crafts reaction (Tscherniac–Ein-
horn reaction) is described. The entire C9 neopentylaminomethyl side chain was installed in one step
using an electronically optimized chloromethylacyliminium precursor in 83% yield. Deprotection and
Received 12 July 2008
Revised 30 July 2008
Accepted 4 August 2008
Available online 8 August 2008
re-equilibration to the C4 a-epimer in the presence of CaCl2 and ethanolamine or NaOH afforded the tar-
get aminomethylcycline antibiotic. The corresponding crystalline tosylate salt was found to exhibit
improved solid state stability.
Keywords:
Minocycline
Aminomethylcycline
Acyliminium
Ó 2008 Elsevier Ltd. All rights reserved.
Friedel–Crafts
Back-epimerization
The tetracyclines comprise a family of antimicrobials which
have been in clinical use for over 50 years and include tetracycline,
doxycycline, and minocycline.1 As a class, they have a broad spec-
trum of antimicrobial activity, including gram-positive, gram-neg-
ative, anaerobic, and atypical bacteria. Recently, their clinical use
has declined, primarily as a result of the increased prevalence of
tetracycline resistance and availability of effective alternative ther-
apies. Due to the rising incidence of multi-drug resistant bacteria,
there is a growing need for the development of new antibacterials
that are effective against these organisms.2 Despite intense effort,
very few of these new antibiotics have been approved by the
FDA in the past decade, and even fewer show oral activity.3
organisms with either ribosomal protection and/or efflux mecha-
nism of resistance. In vitro studies have shown it to have activity
against the vast majority of gram-positive pathogens, including
multi-drug resistant Staphylococcus aureus and vancomycin-resis-
tant Enterococci.5 It has also shown improved activity over minocy-
cline against a wide spectrum of gram-negative and anaerobic
clinical bacterial isolates. Both oral and intravenous formulations
were developed for clinical testing.6
The current synthesis of PTK0796, depicted in Scheme 1, has
been used for multi-kilogram preparations. In step 1, minocycline
hydrochloride was reacted with nearly three equivalents of hydro-
xymethylphthalimide in triflic acid. Compound 1 possesses several
reactive function groups, the C1 primary amide being more reactive
than C9 or the C10 phenol. Due to this fact, the reaction afforded a
mixture of bis- and tris-adducts 3 and 4. In step 2, the phthalimides
were deprotected with large excess of methylamine to afford a mix-
ture of mono- and bis-aminomethyl 5 and 6, along with solid waste
phthalamide 7. Compound 5 was found to be unstable and required
storage at sub-ambient temperatures. In step 3, de-aminomethyla-
tion of the primary amide and a concomitant reductive amination
of C9 methylamine with pivaldehyde afforded crude PTK0796. After
reverse phase chromatographic purification, pH adjustment and
precipitation afforded PTK0796 as an amorphous, unstable solid.
The overall crude yield of PTK0796 was about 35%, and 50% was
recovered after the purification.
N
N
H
H
OH
NH2
H
N
OH
OH
OH
O
O
O
PTK0796
PTK0796, discovered by Paratek, is a novel semi-synthetic anti-
biotic in the tetracycline class known as the aminomethylcyclines
containing a neopentylaminomethyl group attached at the C9 po-
sition.4 PTK0796 represents a significant advance within this class
of antibiotics due to its activity against tetracycline resistant
For a long-term manufacturing route, the following issues
needed to be addressed: (1) the phthalimide protection and de-
protection, (2) instability of aminomethyl intermediate 5, (3) chro-
matographic purification, and (4) the amorphous nature of the final
* Corresponding author. Tel.: +1 732 594 3760.
E-mail address: john_chung@merck.com (J. Y. L. Chung).
0040-4039/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2008.08.011

6096
J. Y. L. Chung et al. / Tetrahedron Letters 49 (2008) 6095–6100
HCl
H
N
O
N
N
N
25 equiv
CH3NH2
H
H
H
4
7
O
O
Triflic acid
O
OH
OH
NH2
H
10
1
EtOH
N
N
N
9
OH
HO
OH
HO
O
O
O
O
O
O
OH
O
O
O
N
O
R
OH
NH
HN
Minocycline HCl (1)
2
3/4 = 60 : 40
3, R = H; 4, R = (CH2)Phthalimide
O
7
N
H
N
H
1. Reverse-phase
Chromatography
OH
PTK0796
Crude PTK0796
CHO
H
H2N
N
Pd/C, H2
TEA, MeOH
~15-18% overall yield
amorphous solid
Thermal - and air-unstable
35% overall yield
as a ~60wt% solid
2. pH adjustment,
CH2Cl2 extractions,
& precipitation w/
heptane/MTBE
R
OH
HO
OH
O
O
O
5, R = CH2NH2
6, R = H
Scheme 1. Current synthesis of PTK0769.
product. We envisioned that a direct method of installing the en-
tire C9 side chain that should address the first two issues. For the
later two issues, we adopted a salt screening strategy, hoping to
find a crystalline salt of PTK0796 that would simplify purification
and improve stability.
O
O
(CH2O)n
N
R
O
R'
NH2
hexanes
neat or in
CH2Cl2
70 oC, 3 h
N
N
Three strategies were explored for installing the entire C9 side
chain (Scheme 2): (1) Mannich reaction of minocycline with imin-
ium ion 8; (2) Tscherniac–Einhorn reaction of minocycline with
acyliminium 9; and (3) Suzuki–Miyaura cross-coupling of 9-bro-
mo-minocycline (11) with trifluoroborate 10.
In the Mannich route, the reactions of minocycline with several
iminium ions and precursors of 8 were investigated under a variety
of reaction conditions.7 The outcomes were either no reaction,
hydroxymethylation of the primary amide, C4 epimerization, or
decomposition. In all cases, no or only trace desired product was
detected. Similarly, attempted Suzuki cross-coupling of 10 with
either 9-bromo-minocycline (11) or bromobenzene afforded no de-
sired products.8 On the other hand, the use of acyliminium ion 9
derived from neopentylamine offered promising results.
12
90-95%
a, R,R'=CF3
O
R
b, R,R'=CCl3
c, R,R'=CHCl2
d, R,R'=CH2Cl
e, R,R'=CH3
O
R
N
O
R'
N
9
O
13
f, R=H, R'=tBu
Order of Reactivity: R = CF3 ~ CCl3 > CHCl2 ~ CH2Cl > CH3 > H
Scheme 3. Preparation of acyliminium precursors 13.
ties of crude compounds 13a–f were confirmed by NMR, and each
was used directly in the next step without purification.
While solvents for the Tscherniac–Einhorn reaction were not
extensively screened, triflic acid proved to be a superior solvent
relative to methanesulfonic acid or sulfuric acid in yield and purity.
Triflic acid solutions of minocycline and PTK0796 are stable to air
oxidation, C4 epimerization, and other modes of degradation. More
importantly, triflic acid also protonates11 acyliminium ions to gen-
erate a super-electrophilic species A that promotes efficient reac-
tion with an electronic deficient, positively charged D-ring of
minocycline (B).
Acyliminium precursors 13 were conveniently prepared in two
steps starting from neopentylamine and paraformaldehyde to af-
ford triazane 12,9 followed by treatment with anhydrides (Scheme
3).10 The reactivity of the later reaction increases with increasing
electron-withdrawing ability of the anhydride R group. The identi-
Mannich
R
N
N
N
H
H
N
OH
NH2
NH
OH
8
+
O
R
+
N
R
Acyliminium
OH
OH
N
OH
O
O
O
O
R
OH
OH
PTK-
0796
Minocycline
A
B
N
The formation of acyliminium precursor 13a from 12 with tri-
fluoroacetic anhydride was complete within 10 min at 22 °C based
on NMR spectroscopic analysis. It was used directly in the reaction
with minocycline HCl in triflic acid at 22 °C, and quickly afforded a
mixture of hydroxymethylated amide 14 and desired 15 (Scheme
4) based on LC/MS and NMR analysis. Further aging at 22 °C led
to disappearance of 15 and the formation of bis-hydroxymethyl
adduct 16. Upon warming to 35 °C, 16 was transformed to tris-
hydroxymethylated 17 and the cyclic ether 18. A similar outcome
9
N
N
H
H
Suzuki-Miyaura
OH
NH2
H
N
BF3K
+
Br
OH
OH
OH
O
O
O
10
11
Scheme 2. Synthetic approaches to PTK0796.

J. Y. L. Chung et al. / Tetrahedron Letters 49 (2008) 6095–6100
6097
O
CF
3
N
N
N
N
N
O
CF
3
H
H
H
H
O
CF3
OH
OH
13a
O
3-10 equiv
H
rt, 18 h
H
+
minocycline HCl
N
N
OH
N
OH
TfOH, rt
OH
OH
OH
OH
OH
O
O
O
OH
O
O
O
15 30%
50%
14
N
N
N
O
N
N
N
H
H
H
H
H
H
OH
OH
OH
H
N
H
35 o
18 h
C
H
+
OH
HO
N
OH
HO
N
OH
OH
OH
OH
OH
OH
O
O
O
O
O
O
O
O
OH
O
OH
O
O
16
OH
18
minor
17
major
Scheme 4. Tscherniac–Einhorn reaction of minocycline with 13a.
was also observed with 13b. These results indicated that these
highly electronically deficient trifluoroacetamide and trichloroace-
tamide groups are good leaving groups, and are readily displaced
by water presumably via quinone methide intermediates.
13e was significantly less reactive than 13a/b, and the amide
groups at C9 in the resulting intermediates are not displaced by
water nor were hydroxymethylene byproducts at C10 AOH
observed.
Hoping to suppress the leaving-group problem, we then exam-
ined electronically neutral acyliminium precursor 13e which was
prepared from 12 with acetic anhydride at 100 °C for 1 h. In the
subsequent reaction with minocycline HCl in triflic acid at 22 °C,
the reaction slowly produced a mixture of primary amide function-
alized 19 and 20 (Scheme 5). Further aging at 35–40 °C with addi-
tional 13e afforded a mixture of desired C9 functionalized 21 and
22 in 92% combined yield. As expected with acetamide not being
a good leaving group, no bis-hydroxylated 16 was detected. Thus,
Since neither electron-deficient 13a/b nor electron-neutral 13e
provided the desired reactivity profile, so screening was continued
with electronically attenuated dichloro and monochloro acylimin-
ium precursors 13c and 13d. These compounds were prepared
from 12 and dichloroacetic anhydride or chloroacetic anhydride,
respectively, at 20–35 °C for 20 min and used directly without iso-
lation in the reaction with minocycline in triflic acid (Scheme 6). In
contrast to the above studies, each of these reactions produced a
single product, 25 or 26, respectively. Optimized yield (83% assay
N
N
N
N
H
H
H
H
O
OH
OH
H
H
O
N
OH
N
N
OH
N
O
OH
OH
OH
OH
OH
O
O
O
OH
O
H
O
O
53%
O
13e
13e
(4 equiv)
19
21
33%
(4 equiv)
minocycline HCl
+
+
35-40 o
24 h
C
TfOH
rt, 3d
N
N
N
N
H
H
H
OH
O
O
OH
O
H
H
N
N
N
N
N
OH
OH
OH
OH
O
O
O
OH
O
OH
O
O
20
50%
39%
22
Scheme 5. Tscherniac–Einhorn reaction of minocycline with 13e.
N
N
Cl
N
N
H
H
X
H
H
OH
O
OH
TfOH, rt,
30 min
35 oC
24 h
H
H
N
OH
minocycline HCl
N
N
OH
Cl
OH
OH
OH
OH
O
OH
O
O
O
Cl
Cl
OH
O
O
O
N
O
23
Cl
25 X = Cl
26 X = H
O
13c
+
or
N
N
Cl
H
H
X
O
Isolated
Purity
OH
O
Cl
O
Cl
corrected
yield of 26
67%
A% (280
nm)
Acyliminium
precursor 13d
4 equiv
N
H
C4 α : β
95:5
N
N
O
13d
90%
OH
OH
O
98:2
94:6
OH
O
O
83%
53%
92%
95%
5 equiv
10 equiv
(5 equiv)
24
Scheme 6. Tscherniac–Einhorn reaction of minocycline with 13c or 13d.

6098
J. Y. L. Chung et al. / Tetrahedron Letters 49 (2008) 6095–6100
yield on the isolated crude) and purity were achieved using 5 equiv
of 13c or 13d at 35–40 °C for 24 h. Thus, a one-step installation of
the entire C9 side chain was achieved.
With compounds 25 and 26 in hand, the removal of the
(di)chloroacetyl and hydroxymethyl groups remains to complete
the synthesis. It is known that the chloroacetyl group could be
cleaved by an intramolecular- assisted removal strategy, such
as with thiourea12 or hydrazinedithiocarbonate.13 However, all
attempts on 26 with these and other mono- and bi-functional
nucleophiles resulted in clean displacement of the chloride
(e.g., 27, 28), but no subsequent inter- or intramolecular cleavage
of the acetyl group (Scheme 7). Examination of molecular models
suggests that the congested steric nature of the region near the
amide carbonyl prohibited proper alignment of the reactive cen-
ters even in the intramolecular sense. Additional attempts to
cleave the dichloroacetyl group in 25 by treatment with tertiary
amines bases14 and other methods did not afford any desired
product either.
Due to problems with the assisted removal strategy deprotec-
tion, further studies focused on 26 under conventional hydrolysis
conditions. Interestingly, simple heating of 26 in 3 N HCl at 70 °C
afforded PTK0796, in which both the on the C2 amide hydroxy-
methyl on the C2 amide and the chloroacetyl groups were cleaved,
as well as small amounts of hydroxymethyl 29. However, the C4
N
Cl
N
?
O
H
N
N
OH
OH
PTK0796
26
center was epimerized to a 55:45
a/b mixture (Scheme 8). Under
H2N
or
S
H2N
the same hydrolysis conditions, dichloride 25 hydrolyzed about
two times slower. Two other aqueous acids, 3 N TfOH and 4.5 N
H2SO4, were also investigated in the hydrolysis of 26, and found
to be inferior to 3 N HCl with respect to yield and de-
hydroxymethylation.
no reaction
NH2
NH2
NH2
NH
NH
H2N
S
N
N
Applying Noseworthy’s back-epimerization protocol for tetra-
O
O
cycline,15 a 55:45 PTK0796 C4
a
/b mixture was back-epimerized
/b mixture by heating at 105 °C in aqueous n-butanol
in the presence of CaCl2 and ethanolamine. Likewise, PTK0796 that
was rapidly epimerized to a 22:78 b mixture by dissolution in
or
to a 92:8
a
N
N
OH
OH
28
a
27
acetic acid was re-equilibrated to a 92:8 mixture with similar effi-
ciency (Scheme 9). Subsequently, it was found that ethanolamine
could be replaced by NaOH, which allowed hydrolysis and epimeri-
zation to be carried out in a one-pot process. Normally, tetracy-
clines are sensitive to strong bases, but presumably, the
compound is stabilized by chelation to CaCl2, attenuating the
destructive effect of NaOH. Thus, after the completion of the HCl
hydrolysis of 26 in the presence of CaCl2, the mixture was directly
basified with NaOH and subjected to the back-epimerization con-
ditions in the same pot (Scheme 10). The overall assay yield of
crude PTK0796 for the two steps was 45% (67% for each step).
other failed attempts:
S
NH2
NH2
MeNH2 etc.,
N
NH2
-S
N
,
HS
HO
,
,
H
... w/ NaOH or TMG led to 7-membered lactam
N
O
A
O
Scheme 7. Attempted deprotection of compound 26.
α/β 55:45
N
N
N
N
X
Cl
N
4
4
OH
O
OH
3N HCl
H
N
H
OH
H
OH
N
N
OH
70 oC, 19 h
R
OH
O
OH
O
O
OH
O
OH
O
O
PTK0796: R = H
29: R = CH2OH
26: X = H
25: X = Cl
>95% conversion
50% conversion
Aq. Acid
70 oC, 24 h
PTK0796
(C4 α/β epimer)
29
+
26
(C4 α/β epimer)
ratio of PTK0796 : 29
(C4 α/β epimers)
HPLC yield of
PTK0796 epimers
Aqueous
acid
50 (55:45)
50 (55:45)
:
3 N TfOH
4.5 N H2SO4
3 N HCl
47%
7 (45:55)
4 (40:60)
11 (49:51)
93 (45:55)
96 (45:55)
89 (53:47)
:
:
:
64%
67%
49%
3 N HCl
(μwave at
140 oC/12 min)
Scheme 8. Deprotection/epimerization of 25 and 26.

J. Y. L. Chung et al. / Tetrahedron Letters 49 (2008) 6095–6100
6099
n
BuOH-H2O
CaCl2
Acknowledgments
PTK0796
C4 α : β
92 : 8
PTK0796
C4 α : β
55 : 45
3N HCl
70 ο
26
We thank Robert Reamer and Peter G. Dormer for their help
with NMR.
ethanolamine or NaOH
105 oC, 2-4 h
C
Scheme 9. Epimeriation and back-epimerization of PTK0796 C4 center.
References and notes
1. (a) Lopez-Ruz, M. A.; Lopez-Gomez, M.; Pasquau-Liano, J.; Jimenez-Alonso, J.
Antibiotics 2005, 83–142; (b) Nelson, M. L.; Projan, S. J. Front. Antimicrob. Resis.
2005, 29–38; (c) Finch, R. G. Tetracyclines. In Antibiotic and Chemotherapy, 7th
ed.; O’Grady, F., Ed.; Churchill Livingstone: Edinburgh, UK, 1997; pp 469–484.
2. (a) Shlaes, D. M. Curr. Opin. Invest. Drugs 2006, 7, 167–171; (b) McMurry, L. M.;
Levy, S. B. Tetracycline Resistance Determinants in Gram-positive Bacteria. In
Gram-Positive Pathogens, 2nd ed.; Fishetti, V. A., Ed.; American Society for
Microbiology: Washington, DC, 2006; pp 801–820.
3. Tigecycline, a novel glycylcycline antibiotic with potent antibacterial activities
against resistant pathogens, was approved by FDA in June 2005. It is not
approved for use in children. There is no oral form available. (a) Rose, W. E.;
Rybak, M. J. Pharmacotherapy 2006, 26, 1099–1110; (b) Frampton, J. E.; Curran,
M. P. Drugs 2005, 65, 2623–2635.
4. (a) Ohemeng, K.; Amoo, V.; Kim, O.; Bowser, T.; Assefa, H.; Bhatia, B.; Berniac, J.;
Chen, J.; Grier, M.; Honeyman, L.; Pan, J.; Mechiche, R. World Patent WO
2,005,009,944.; (b) Nelson, M. L.; Ohemeng, K.; Frechette, R.; Ismail, M. Y.;
McIntyre, L.; Bowser, T. World Patent WO 2,004,091,513.; (c) Nelson, M. L.;
Frechette, R.; Viski, P.; Ismail, M.; Bowser, T.; Dumornay, J.; Rennie, G.; Liu, G.;
Koza, D.; Sheahan, P.; Stapleton, K.; Hawkins, P.; Bhatia, B.; Verma, A.;
McIntyre, L.; Warchol, T. World Patent WO 2,002,004,406.
n
BuOH-H2O
CaCl2
PTK0796
C4 α : β
92 : 8
PTK0796
C4 α : β
55 : 45
3N HCl
70 ο
26
ethanolamine or NaOH
105 oC, 2-4 h
C
Scheme 10. One-pot hydrolysis of 26 and back-epimerization of PTK0796 C4
epimer.
PTK0796, an amorphous solid, is unstable at temperatures
above 0 °C and when exposed to air. Our salt screening effort found
that PTK0796 could be crystallized as a bis-HCl salt, a MSA salt, and
a TsOH salt. The crystalline mono-tosylate salt was found to pos-
sess significantly improved solid-state stability at 25 °C. This salt
also improved the C4 a-epimer to some extent.
In summary, an efficient three-step two-pot convergent synthe-
sis of the aminomethylcycline antibiotic PTK0796 was developed
(Scheme 11).16 Installation of C9 side chain was accomplished in
one step via an electronically tuned mono-chloro acyliminium Fri-
edel–Crafts reaction (Tscherniac–Einhorn reaction) with minocy-
cline in 83% yield. Deprotection of the chloroacetyl and
hydroxymethyl groups in aqueous HCl, followed by a re-equilibra-
tion of the C4 center in a one-pot process afforded crude PTK0796
in 45% yield. A crystalline tosylate salt of PTK0796 was identified as
a potential salt for further purification and as a final active pharma-
ceutical ingredient due to its improved long-term stability.
5. (a) Presented at the 46th Annual Interscience Conference on Antimicrobial
Agents and Chemotherapy (ICAAC) in San Francisco, CA. These four studies
were presented in the Poster Summary Session #226 on September 30, 2006.
MK-2764/PTK0796 is potent against all S. aureus strains tested, irrespective
of antibiotic resistance. Smith, K.; Tanaka, S. K.; Appelbaum, P. C.
Antistaphylococcal Activity of MK-2764/PTK0796 Compared to Other Agents,
Poster F1-1971.; (b) MK-2764/PTK0796 exhibits potent efficacy against gram-
positive and gram-negative pathogens including MSSA, MRSA, SP, EC and KP in
animal studies. Craig, W. A.; Andes, D.; Odinecs, A. In Vivo Pharmacodynamics
of MK-2764/PTK0796 Against Various Gram-Positive and Gram-Negative
Bacteria in the Thighs of Neutropenic and Normal Mice, Poster F1-1974.; (c)
MK-2764/PTK0796 is an effective agent against the pneumococcus in this
animal model, and the pharmacodynamic parameter predictive of antibacterial
activity is the (free) AUC/MIC. Tessier, P.R.; Fan, H.W.; Tanaka, S.K.; Nicolau,
D.P. Pharmacokinetic/Pharmacodynamic Profile of MK-2764/PTK0796 against
S. pneumoniae in a Murine Pneumonia Model, Poster F1-1973.; (d) These data
confirm the activity of this novel aminomethylcycline antibiotic, MK-2764/
PTK0796, against Legionella spp, the cause of Legionella pneumonia. Dubois, J.;
Tanaka, S. K. In Vitro Activity of MK-2764/PTK0796 against Legionella sp., Poster
F1-1972.
Cl
O
Cl
N
N
N
O
4
OH
NH2
6. Nelson, M. L.; Ismail, M.; Bowser, T.; Honeyman, L.; Macone, A.; Bhatia, B.;
Verma, A.; Grier, M.; Berniac, J.; Mechiche, R.; Ohemeng, K.; Cannon, P.;
Donatelli, J.; McKenney, D.; Levy, S. B. Synthesis and biological activity of PTK
0796: A novel semisynthetic tetracycline in Phase I clinical trials. Abstracts of
Papers, 231st National Meeting of the American Chemical Society, Atlanta, GA,
United States, March 26–30, 2006, MEDI-198.
O
13d
OH
TfOH, 35 oC, 24 h
OH
O
OH
O
O
minocycline . HCl
83% LC yield
7. Iminium and precursors screened included (1) neopentylamine with
paraformaldehyde or formalin, (2) N,N-bis(methoxymethyl)-2,2-dimethyl-
propan-1-amine, (3) 1,3,5-tris(2,2-dimethylpropyl)-1,3,5-triazinane, (4)
N-(1H-1,2,3-benzotriazol-1-ylmethyl)-2,2-dimethylpropan-1-amine, and (5)
N-benzyl-N-methylenebutan-1-aminium chloride, under acidic, neutral, basic,
and Lewis acids in variety of solvents at 0–120 °C.
8. Potassium neopentylaminomethyl trifluoroborate (10) was prepared
analogously by a known procedure except that neopentylamine was used
instead of secondary amines. (a) Molander, G. A.; Ham, J. Org. Lett. 2006, 8,
2031–2034; (b) Molander, G. A.; Gormisky, P. E.; Sandrock, D. L. J. Org. Chem.
2008, 73, 2052–2057.
9. (a) Taguchi, M.; Aikawa, N.; Tsukamoto, G. Bull. Chem. Soc. Jpn. 1988, 61, 2431–
2436; (b) Anderson, J. E.; Casarini, D.; Ijeh, A. I.; Lunazzi, L.; Tocher, D. A. J. Am.
Chem. Soc. 1995, 117, 3054–3056.
10. Giumanini, A. G.; Verardo, G.; Cauci, S. J. Prakt. Chem. 1987, 329, 417–432.
11. Olah, G. A.; Wang, Q.; Sandford, G.; Oxyzoglou, A. B.; Prakash, G. K. Synthesis
1993, 1077–1079.
Cl
N
O
N
4
OH
3N HCl
H
OH
O
N
N
70 oC, 20 h
OH
N
O
OH
O
OH
67% LC yield
26
N
n
BuOH-H2O
4
OH
NH2
CaCl2
H
OH
N
ethanolamine
or NaOH
12. Masaki, M.; Kitahara, T.; Kurita, H.; Ohta, M. J. Am. Chem. Soc. 1968, 90, 4508–
4509.
OH
O
OH
O
O
105 oC, 3 h
4
13. van Boeckel, C. A. A.; Beetz, T. Tetrahedron Lett. 1983, 24, 3775–3778.
14. Rajeswaran, W. G.; Cohen, L. A. Tetrahedron Lett. 1997, 38, 7813–7814.
15. (a) Noseworthy, M. M. U.S. Patent 3,009,956; Chem. Abstr., 1962, 56, 6101.; (b)
Korst, J. J.; Johnston, J. D.; Butler, K.; Bianco, E. J.; Conover, L. H.; Woodward, R.
B. J. Am. Chem. Soc. 1968, 90, 439–457.
16. Preparation of 1,3,5-tris(2,2-dimethylpropyl)-1,3,5-triazinane (12). Neopen-
tylamine (12.2 g, 136.1 mmol) was added to a mixture of paraformaldehyde
(4.3 g, 136.1 mmol) in hexanes (85 mL) and heated at reflux (68–71 °C) for
2–3 h. Anhydrous Na2SO4 (3 g) was added and stirred at 22 °C for 1 h. The
mixture was filtered, washed with hexanes (85 mL), and evaporated to dryness
to afford 12.78 waxy white solid (94.7%). NMR indicated that the resulting
product was ꢀ95 wt % pure. 1H NMR (CDCl3, 400 MHz) d 3.33 (s, 2H), 2.13
C4 α/β 55:45
67% isolated yield
Crude
PTK0796
1. chrom.
2. TsOH
PTK0796 tosylate
C4 α/β
92:8
37% overall yield
Scheme 11. Three-step two-pot synthesis of PTK0796 via Tscherniac–Einhorn
reaction.

6100
J. Y. L. Chung et al. / Tetrahedron Letters 49 (2008) 6095–6100
(s, 2H), 0.88 (s, 9H); 13C NMR (CDCl3, 100.5 MHz)
27.8.
d 78.6, 64.2, 33.6,
organic layers were extracted with 15 °C 3 N HCl (160 mL). The layers were sep-
arated and the organic phase was re-extracted with 3 N HCl (1 ꢁ 20 mL,
1 ꢁ 10 mL). The combined aqueous layers were assayed to contain 73–83% yield
of product 26. At this point, product 26 could be isolated by pH adjustment to
6.5, and extraction with dichloromethane, or taken directly into the next step as
described below.
Compound 26. NMR in DMSO-d6 indicated that the chloroacetamide exists as
two rotomers: C9fH2 as AB quartets at d 4.40 and 4.35; C2cH2 as a broad singlet
at d 4.90; LC/MS m/z (M+H)+ = 663; HRMS m/z calcd for C32H44ClN4O9 (M+H)+
663.2797, found, 663.2794.
General procedure for the preparation of 13. 3.2 equiv of the anhydride and
1 equiv of trizainane 12 were stirred neat or in dichloromethane at the
indicated temperature and time. The resulting 13 was used directly in the
Friedel–Craft reaction.
Compound 13a: R,R0 = CF3. Trifluoroacetic anhydride (3.2 equiv), 1,3,5-tris(2,2-
dimethylpropyl)-1,3,5-triazinane (1 equiv), and dichloromethane (1.6 M) were
combined and stirred at 25 °C for 20 min. A small aliquot was removed for
NMR analysis: 1H NMR (CDCl3, 400 MHz) d 5.70 (s, 2H), 3.42 (s, 2H), 1.00 (s,
9H).
Compound 13b: R,R0 = CCl3. Trichloroacetic anhydride (3.2 equiv), 1,3,5-
tris(2,2-dimethylpropyl)-1,3,5-triazinane (1 equiv), and dichloromethane
(1.6 M) were combined and stirred at 25 °C for 20 min. A small aliquot was
removed for NMR analysis: 1H NMR (CDCl3, 400 MHz) d 6.00 (s, 2H), 3.45 (s,
2H), 1.01 (s, 9H).
Cl
N
N
4
H
H
7
9e
N
3
2
9f
O
OH
8
9
6
5
9d
26
4a
5a
12a
H
9c
9b
1
11
N
OH
11a
12OH
10
9a
2c
Compound 13c: R,R0 = CHCl2. Dichloroacetic anhydride (3.2 equiv), 1,3,5-
tris(2,2-dimethylpropyl)-1,3,5-triazinane (1 equiv), and dichloromethane
(1.6 M) were combined and stirred at 40 °C for 30 min. A small aliquot was
removed for NMR analysis: 1H NMR (CDCl3, 400 MHz) d 6.49 (s, 1H), 5.98 (s,
1H), 5.67 (s, 2H), 3.40 (s, 2H), 0.96 (s, 9H).
OH
O
OH
O
O
Conversion of 26 to PTK0769 tosylate. The aqueous solution from above was
heated at 70 °C for 23 h and the resulting solution had a 49% assayed yield of
Compound 13d: R,R0 = CH2Cl. Chloroacetic anhydride (3.2 equiv), 1,3,5-tris(2,2-
dimethylpropyl)-1,3,5-triazinane (1 equiv), and dichloromethane (1.6 M) were
combined and stirred at 40 °C for 30 min. A small aliquot was removed for
NMR analysis: 1H NMR (CDCl3, 400 MHz) d 5.56 (s, 2H), 4.22 (s, 2H), 4.07 (s,
2H), 3.32 (s, 2H), 0.92 (s, 9H).
PTK0796 as a ꢀ1:1 mixture of C4
a/b epimers and small amounts of the corre-
sponding N-hydroxymethyl analogues. The solution was evaporated to yield
9.7 g of brown oil. n-Butanol (50 mL, degassed by N2 bubbling for 5 h, which re-
duced dissolved oxygen from 8 ppm to 0.07 ppm), degassed water (2.5 mL),
CaCl2 (3.92 g), and ethanolamine (4.0 mL) were added. The resulting brown
solution with some insoluble CaCl2 was heated to 108 °C for 2 h then cooled
to rt. Assay yield was 38% as a 92:8 C4 alpha/beta mixture. To the solution
was added Darco KB activated carbon (3 g, 100 mesh) and stirred for 30 min.
The solution was filtered through Celite and washed with nBuOH (70 mL).
The filtrate was diluted with water (60 mL) and adjusted to pH 7.5 with etha-
nolamine (5 drops). The layers were separated. The organic phase was washed
with water (60 mL), then concentrated to dryness. The resulting gummy mate-
rial was triturated with ethanol and evaporated to dryness to afford 6 g of
brown powder with 32 wt % purity (1.92 g or 34% overall yield). The material
was purified to >90 wt % purity using reversed phase HPLC method described
in International Publication Nos. WO 2005/009944, WO 2004/091513, and
WO 2002/004,406. Further purity upgrade to >98% and improved stability at
room temperature were achieved via the tosylate salt. The crystalline TsOH salt
was prepared by adding a solution of TsOH hydrate (97.0 g) in IPA (400 mL) to a
slurry of amorphous PTK0796 free base (289 g) in IPA (2 L) under nitrogen. The
water content was adjusted to 0.6 g/L with the addition of water (9 ml) and the
slurry was stirred at 20–25 °C for 18 h to produce a thick crystalline slurry. The
slurry was filtered and washed with IPA (2 ꢁ 500 ml). Excess IPA was removed
from the crystalline cake by blowing dry nitrogen through the cake for 24 h.
With the solids containing 3 wt % IPA, the cake was further dried by blowing
humidified nitrogen through the cake at a relative humidity of 70–75% for 24
hr. The cake retained 0.9 wt % IPA that was not further reduced by this method.
Excess water was then removed from the cake by blowing dry nitrogen through
the cake for 24 h to afford 337 g of PTK0796 TsOH salt with >98% purity.
Compound 13e: R,R0 = CH3. Acetic anhydride (3.2 equiv) and 1,3,5-tris(2,2-
dimethylpropyl)-1,3,5-triazinane (1 equiv) were combined and heated at
100 °C for 1 h. A small aliquot was removed for NMR analysis: 1H NMR
(CDCl3, 400 MHz) d 5.39 (s, 2H), 3.29 (s, 2H), 2.20 (s, 3H), 2.09 (s, 3H), 0.91 (s,
9H).
Compound 13f: R = t-Bu, R0 = H. Formyl pivalic anhydride (3.2 equiv) and 1,3,5-
tris(2,2-dimethylpropyl)-1,3,5-triazinane (1 equiv) were combined and heated
at 22 °C for 30 min. A small aliquot was removed for NMR analysis. The
formamide derivative exists as a 2.5:1 mixture of rotamers: 1H NMR (CDCl3,
400 MHz) d 8.42 (s, 0.7H), 8.15 (s, 0.3H), 5.41 (s, 0.6H), 5.36 (s, 1.4H), 3.20 (s,
1.4H), 3.16 (s, 0.6H), 1.26 (s, 2.6H), 1.24 (s, 6.4H), 0.97 (s, 9H).
Conversion of minocycline HCl to compound 25. Dichloroacetic anhydride (7.78 g,
31.8 mmol) was slowly added to solution of triazinane (3.0 g, 10.0 mmol) and
dichloromethane (2 mL) over 20 min with cooling in an ice bath. The solution
was stirred at 22 °C for 1 h. In a separate flask, minocycline hydrochloride (3 g,
6.0 mmol) was slowly added to triflic acid (30 mL) with cooling while
maintaining at <35 °C. The acyliminium precursor solution was slowly added
to minocycline HCl/TfOH solution keeping at <35 °C. The mixture was stirred at
24 °C for 20 h and at 35–40 °C for 4 h, then slowly quenched into 10–15 °C
water (120 mL) over 20 min keeping at <25 °C. To the mixture was added
dichloromethane (60 mL) and the pH was adjusted to 6.5 with 29% NH4OH
(about 28 mL). The layers were separated and the aqueous phase was extracted
with dichloromethane (3 ꢁ 20 mL). The combined organic layers were
extracted with 10 °C 3 N HCl (60 mL). The layers were separated, and the
organic was extracted again with 3 N HCl (2 ꢁ 20 mL). The combined aqueous
phase was maintained at 10 °C and adjusted to pH 6.5 with 29% NH4OH (about
24 mL). It was warmed to 22 °C, and extracted with dichloromethane
(1 ꢁ 100 mL, 3 ꢁ 20 mL). The combined organic was filtered through Na2SO4
and the filtrate evaporated to dryness to afford 3.67 g of 25 as an orange solid.
The material was estimated to be 77 wt % pure using pure PTK0764 as an HPLC
standard. The corrected yield was 76% based on 89% active minocycline HCl.
Compound 25. NMR in DMF-d9, indicated that the dichloroacetamide exists as
two rotomers: C9fH as singlets at d 7.29 and 7.28; C2cH2 as a triplet at d 6.10;
LC/MS m/z (M+H)+ = 697; HRMS m/z calcd for C32H43Cl2N4O9 (M+H)+ 697.2407,
found, 697.2408.
PTK0796 tosylate: 1H NMR (D2O, 500 MHz) d 7.63 (d, J = 8.1, C2 H, C6 H), 7.52 (s,
0
0
0
0
C8 H), 7.28 (d, J = 8.1, C3 H, C5 H), 4.30 (s, C9aH2), 3.80 (d, J = 1.1, C4H), 3.11 (dd,
J = 15.6, 4.2, C6H), 2.95 (s, 4-N(CH3)2), 2.89 (m, C5a H), 2.84 (s, C9bH2), 2.72 (dt,
0
J = 12.7, 2.2, C4aH), 2.62 (s, 7-N(CH3)2), 2.34 (s, C4a H3), 2.24–2.18 (m, C5H,
C6H), 1.66 (td, J = 13.5, 11.5, C5H), 1.01 (s, C9dH9); PTK0796 C4 epimer has res-
onances at d 4.58 (d, J = 3.6), 2.99 (s), 2.59 (s); 13C NMR (D2O, 125.7 MHz)
d193.77 (C11), 191.90 (C1), 184.86 (C3), 174.23 (C12), 171.59 (C2a), 156.94
0
0
(C10), 142.84 (C4 ), 142.31 (C7), 140.30 (C1 ), 139.17 (C6a), 130.18 (C8), 129.98
0
0
0
0
(C2 , C6 ), 125.96 (C3 , C5 ), 117.59 (C10a), 117.26 (C9), 109.20 (C11a), 102.82
(C2), 75.97 (C12a), 71.69 (C4), 58.59 (C9b), 47.84 (C9a), 45.08 (7-N(CH3)2), 41.59
(br s, 4-N(CH3)2), 34.70 (C5), 34.08 (C4a), 32.48 (C5a), 30.73 (C6), 30.33 (C9c),
N
N
4
Cl
H
H
7
Cl
9d
9e
O
3
2
OH
8
9f
9c
9b
5
6
H
26.83 (C9d), 21.15 (C4a ); LC/MS, m/z (M+H)+ = 557.
4a
25
5a
0
12a
9
11
N
1
N
OH
11a
12OH
10
9a
2c
N
N
4
H
H
OH
O
OH
O
O
7
3
2
OH
NH2
9d
9c
8
9
6
5
4a
5a
H
Conversion of minocycline HCl to compound 26. Chloroacetic anhydride (9.0 g,
53.01 mmol), triazinane (5.0 g, 16.67 mmol), and dichloromethane (10 mL)
were combined with cooling in an ice bath. The solution was heated at 35 °C
for 20 min, then cooled to 20 °C. In a separate flask, minocycline hydrochloride
(5 g, 10 mmol) was slowly added to triflic acid (50 mL) with cooling while main-
taining at <35 °C. The acyliminium precursor solution was slowly added to min-
ocycline HCl/TfOH solution keeping at <40 °C. The mixture was heated at 35–
40 °C for 24 h, then quenched into 10–15 °C water (200 mL) over 20 min at
<25 °C. To this mixture was added dichloromethane (150 mL) and adjusted to
pH 6.5 with 29% NH4OH (about 50 mL). The layers were separated, and the
aqueous phase was extracted with dichloromethane (3 ꢁ 25 mL). The combined
12a
PTK0796.TsOH
.
1
11
N
TsOH
11a
10
12
9b
9a
OH
OH
O
OH
O
O
Alternate back-epimerization procedure using NaOH. Friedel–Crafts product 26
(ꢀ70 mg, ꢀ0.1 mmol) and CaCl2 (40 mg) were dissolved in 3 N HCl (1 mL).
The solution was heated to 70 °C for 14 h (or until starting material disappears).
5 N NaOH (0.76 mL, 3.8 mmol) and 1 mL nBuOH (two layers) were added. The
solution was heated at 105–110 °C for 4–5 h, then worked up as described
above.