Structure activity relationships of human galactokinase inhibitors

Article abstract of DOI:10.1016/j.bmcl.2014.11.061

Classic Galactosemia is a rare inborn error of metabolism that is caused by deficiency of galactose-1-phosphate uridyltransferase (GALT), an enzyme within the Leloir pathway that is responsible for the conversion of galactose-1-phosphate (gal-1-p) and UDP-glucose to glucose-1-phosphate and UDP-galactose. This deficiency results in elevated intracellular concentrations of its substrate, gal-1-p, and this increased concentration is believed to be the major pathogenic mechanism in Classic Galactosemia. Galactokinase (GALK) is an upstream enzyme of GALT in the Leloir pathway and is responsible for conversion of galactose and ATP to gal-1-p and ADP. Therefore, it was hypothesized that the identification of a small-molecule inhibitor of human GALK would act to prevent the accumulation of gal-1-p and offer a novel entry therapy for this disorder. Herein we describe a quantitative high-throughput screening campaign that identified a single chemotype that was optimized and validated as a GALK inhibitor.

Full text of DOI:10.1016/j.bmcl.2014.11.061

Bioorganic & Medicinal Chemistry Letters xxx (2014) xxx–xxx
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Structure activity relationships of human galactokinase inhibitors
Li Liu a, Manshu Tang b, Martin J. Walsh a, , Kyle R. Brimacombe a, Rajan Pragani a, Cordelle Tanega a,
Jason M. Rohde a, Heather L. Baker a, Elizabeth Fernandez a, Burchelle Blackman a,à, James M. Bougie a,
William H. Leister a, Douglas S. Auld a,§, Min Shen a, Kent Lai b, Matthew B. Boxer a,
⇑
a National Center for Advancing Translational Sciences, National Institutes of Health, 9800 Medical Center Drive, Rockville, MD 20850, USA
b Division of Medical Genetics, Department of Pediatrics, University of Utah, Rm 2C412 SOM, 50 N. Medical Drive, Salt Lake City, UT 84132, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Classic Galactosemia is a rare inborn error of metabolism that is caused by deficiency of galactose-
1-phosphate uridyltransferase (GALT), an enzyme within the Leloir pathway that is responsible for the
conversion of galactose-1-phosphate (gal-1-p) and UDP-glucose to glucose-1-phosphate and UDP-
galactose. This deficiency results in elevated intracellular concentrations of its substrate, gal-1-p, and this
increased concentration is believed to be the major pathogenic mechanism in Classic Galactosemia.
Galactokinase (GALK) is an upstream enzyme of GALT in the Leloir pathway and is responsible for
conversion of galactose and ATP to gal-1-p and ADP. Therefore, it was hypothesized that the identification
of a small-molecule inhibitor of human GALK would act to prevent the accumulation of gal-1-p and offer
a novel entry therapy for this disorder. Herein we describe a quantitative high-throughput screening
campaign that identified a single chemotype that was optimized and validated as a GALK inhibitor.
Published by Elsevier Ltd.
Received 22 October 2014
Revised 19 November 2014
Accepted 21 November 2014
Available online xxxx
Keywords:
Galactosemia
Structure–activity relationships
Galactokinase
The galactosemias are rare inherited metabolic disorders caused
by deficiencies of the enzymes in the Leloir Pathway.1,2 As the Lel-
oir Pathway is the predominate mechanism for the metabolism of
galactose, all of the galactosemias result in aberrant levels of gal-
actose and/or its down-stream metabolites (Fig. 1). Classic Galacto-
semia (Type I) is the most common form of the galactosemias and
is also the most severe. It is a potentially lethal disorder with a high
mortality rate when left untreated. Currently, many state newborn
screening programs include testing for galactosemia, and if
detected, immediate removal of lactose and galactose from an
infant’s diet is required.3 With widespread incorporation of this
screening, the morbidity rates for those affected with Classic
Galactosemia has significantly decreased, but the life-long galact-
ose restricted diet fails to prevent developmental delay, neurolog-
ical disorders, and premature ovarian insufficiency, which occur in
many affected patients later in life.4–7 Classic Galactosemia is char-
acterized by deficient galactose-1-phosphate uridyltransferase
(GALT), which results in the buildup of its substrate galactose-
1-phosphate (gal-1-P).8 Although the exact pathogenic mechanism
of Classic Galactosemia has not been established, elevated galact-
ose-1-phosphate (gal-1-p) levels have been proposed as a major
pathogenic factor.3,9–11 Immediately upstream from GALT is galac-
tokinase (GALK), which is the enzyme responsible for converting
galactose into gal-1-p. Deficiency in the GALK enzyme results in
Type II Galactosemia and these GALK-deficient patients have much
milder and even benign phenotypes.12 This positions GALK as a
target for reduction of gal-1-p levels and a potential therapeutic
target for Classic Galactosemia.13,14
Using recombinant GALK1 we performed
high-throughput screen (qHTS) against ꢀ274,000 compounds in
1536-well plate format at 6 doses (3 nM–57
M).13,15 ATP was held
at 35 M and galactose at 100 M, both near their Km values
a quantitative
l
l
l
determined under the 1536-well assay conditions. We used the
KinaseGlo™ detection system, which measures remaining ATP
levels after conversion of galactose to gal-1p by GALK1. While the
screen performed acceptably (Z0 = 0.48 0.16, signal/background =
4.5 1.6 and CV = 17 9), there was an extremely low initial hit
rate of 0.05%. Counter-screens were run to identify potential false
positives with activity against the KinaseGlo detection reagent
itself in ATP containing buffer; and with interfering redox activity,
as measured by H2O2 production in an HRP-phenol red coupled
assay (PubChem AIDs 1379 and 2502, respectively). With the
exclusion of these two classes of potential artifacts only 65
⇑
Corresponding author. Tel.: +1 301 217 4681; fax: +1 301 217 5736.
E-mail address: boxerm@mail.nih.gov (M.B. Boxer).
Present address: Dow AgroSciences LLC, Crop-Protection Discovery, Group, Bldg.
306/E2/980, 9330 Zionsville Road, Indianapolis, IN 46268, USA.
à
Present address: Imaging Probe Development Center, National Heart, Lung, and
Blood Institute, National Institutes of Health, 9800 Medical Center Drive, Rockville,
MD 20850, USA.
§
Present address: Center for Proteomic Chemistry, Novartis Institutes for Biomed-
ical Research, Cambridge, MA 02139, USA.
http://dx.doi.org/10.1016/j.bmcl.2014.11.061
0960-894X/Published by Elsevier Ltd.
Please cite this article in press as: Liu, L.; et al. Bioorg. Med. Chem. Lett. (2014), http://dx.doi.org/10.1016/j.bmcl.2014.11.061

2
L. Liu et al. / Bioorg. Med. Chem. Lett. xxx (2014) xxx–xxx
Figure 1. The Leloir pathway for galactose metabolism, highlighting the deficient enzymes responsible for the galactosemias.
compounds remained. Within this small list of compounds, the
spiro-benzoxazole series, exemplified by (PubChem CID
1286615), emerged as a validated hit series. Upon resynthesis of
using the one-pot procedure reported by Potapov et al.
(Scheme 1),16 we confirmed the GALK1 inhibitory activity with
an IC50 of 6 M and began initial structure activity relationship
We first looked at the importance of the gem-dimethyl group
and found that replacement with a mono-phenyl substituent (2)
gave reduced activity, while complete removal of the methyl
1
1
groups gave improved activity to 1 lM (3). Contracting the fused
cyclohexenone ring to a 5-membered ring (4) reduced activity.
Next, we looked at the spiro ring and made cyclohexane (5),
dimethyl (6) and cyclobutane (7) analogs. While all three analogs
maintained GALK1 inhibitory activity, they were significantly
weaker than 3. Attempts to replace the benzoxazole were limited
by the chemistry in this multi-component reaction, but we were
able to access two guanidinyl benzimidazoles (8 and 9) through
chemistry outlined in Scheme 3, followed by the standard neat
melting chemistry typified in Scheme 2. Unfortunately, both
analogs were inactive. Due to the low solubility of 3 and with
l
(SAR) exploration.
While the resynthesis of 1 performed well in refluxing isopro-
panol, the chemistry did not translate well to more diverse sub-
strates. After unsuccessfully trying a few acid and base catalysts,
we settled on mixing the three components neat, sealing in a
microwave vial, and submerging into a 120 °C oil bath (Scheme 2).
This protocol allowed for rapid access to a number of analogs for
initial SAR (Table 1).
O
O
O
N
O
iPrOH
NH
N
N
+
+
NH2
reflux
HN
O
N
H
N
H
O
Scheme 1. Resynthesis of 1.
R1 R2
H2N
NH
NH
O
O
O
neat, sealed tube
N
O
N
N
+
+
R1
120 oC
R3
R4
R3
R4
R2
O
N
N
H
O
H
n
R1/R2 = cycloalkyl
or H/Aryl
R3/R4= H/H, H/Ph
or Me/Me
n = 0 or 1
Scheme 2. Use of neat melting procedure for broader applicability to diverse substrates.
Please cite this article in press as: Liu, L.; et al. Bioorg. Med. Chem. Lett. (2014), http://dx.doi.org/10.1016/j.bmcl.2014.11.061

L. Liu et al. / Bioorg. Med. Chem. Lett. xxx (2014) xxx–xxx
3
Table 1
Initial SAR around hit 1
a
Compounds
Structure
IC50
6.0
(
lM)
Max responseb (%)
O
N
N
1
ꢁ82
Me
Me
O
N
H
N
H
O
N
N
2
3
4
ND
ꢁ36
ꢁ86
ꢁ75
O
Ph
O
N
N
H
H
N
N
1.0
O
N
H
N
H
O
N
N
13.3
O
N
H
N
H
O
N
N
5
11.9
ꢁ99
O
N
H
N
H
O
O
Me Me
N
N
N
6
7
16.8
16.8
ꢁ74
ꢁ74
O
O
N
H
N
H
N
N
H
N
H
O
O
N
N
N
8
ND
ND
<10
<10
N
N
H
N
H
Me
N
N
9
N
N
Bn
a
IC50 values were determined utilizing the luminescence GALK-ATP-depletion assay. ND = not determined; for compounds with max
M weaker than 50%, IC50 was ND.
Max Resp. represents the % inhibition at 57 M compound.
response at 57
l
b
l
the realization that the chemistry enabled the use of non-enoliz-
able aldehydes in the 3-component reaction, we synthesized ana-
logs that replaced the alkyl-spiro substituent with aryl groups
(Table 2).
in Table 1 (ꢀ25 fold less potent). A scan of para and meta substit-
uents (11–18) resulted in very weak inhibitors showing minimal
GALK inhibition at the top dose of 57 lM. Interestingly, ortho
substituted phenyl analogs showed a preference for chloro (19),
While the phenyl analog (10) had GALK1 inhibitory activity, it
was severely diminished compared to many of the spiro analogs
bromo (21) or CF3 (22) groups, all of which gave IC50 values under
8 lM. The corresponding methoxy analog (20) showed weak inhi-
bition. Combining two ortho chloro substituents (23), resulted in
reduced potency. Pyridine and pyrazole analogs not bearing a
2-substituent had minimal activity (24–26).
While there was a preference for ortho-halogenated analogs for
reasonable potencies, these analogs all had aqueous solubilities of
H2N
NH
NH
NH2
Tf
1. DIPEA
DCM, reflux
N
N
N
N
N
R
+
BocHN
NHBoc
2. TFA, DCM
R
<1 lg/mL. In an effort to improve this property, we decided to
investigate whether this motif could be combined with heterocyclic
groups in hopes of obtaining improved compounds. To this end,
R = Me or Bn
Scheme 3. Synthesis of benzimidazole-guanidines.
we synthesized
a variety of analogs utilizing 2-halogenated
Please cite this article in press as: Liu, L.; et al. Bioorg. Med. Chem. Lett. (2014), http://dx.doi.org/10.1016/j.bmcl.2014.11.061

4
L. Liu et al. / Bioorg. Med. Chem. Lett. xxx (2014) xxx–xxx
Table 2
SAR around the aryl ring showing preference for ortho-substitution
R
O
N
N
O
N
H
N
H
a
a
Compounds
R
IC50
25
(lM)
Max responseb (%)
Compounds
R
IC50
6
(lM)
Max Responseb (%)
Cl
10
ꢁ70
19
ꢁ77
F
OMe
Br
11
12
13
ND
ND
ND
ꢁ38
ꢁ11
ꢁ33
20
21
22
ND
3
ꢁ36
ꢁ84
ꢁ79
Cl
MeO
Me
CF3
Cl
8
14
ND
ꢁ35
23
ND
ꢁ44
Cl
N
N
15
16
17
18
ND
ND
ND
18.1
ꢁ32
ꢁ45
ꢁ25
ꢁ60
24
25
26
ND
ND
ND
ꢁ15
ꢁ37
ꢁ24
Me
F
HN
N
Cl
MeO
a
IC50 values were determined utilizing the luminescence GALK-ATP-depletion assay. ND = not determined; for compounds with max response at 57
IC50 was ND.
lM weaker than 50%,
b
Max Resp. represents the % inhibition at 57 lM compound.
heterocyclic aldehyde starting materials (Table 3). 2-Chloro,
-bromo and -trifluoromethyl pyridine isomers were made (27–33),
half-life, respectively) with a concordantly high PAMPA permeabil-
ity. We also determined that 36 was stable in PBS buffer (pH = 7.4)
as well as in the presence of 5 mM glutathione, had moderate
mouse liver microsomal stability (t1/2 = 34 min) and good mouse
plasma stability (t1/2 >60 min) (Table 5). The compound was also
stable to oxidants (air and MnO2) indicating it was not redox
active. Being that 36 had the best preliminary in vitro ADME pro-
file, we decided to carry it on for further selectivity, mechanism
of action and cell-based studies.
To determine the mode of inhibition for 36, mechanistic sub-
strate competition studies were performed by independently
titrating both ATP and galactose substrates in a coupled GALK1
assay system. Analog 36 appeared to shift the Km of ATP (Fig. 3A
and B), suggesting that the compound is competitive with respect
to ATP, an observation which was corroborated by mixed-model
with most giving potency in the range of ꢀ3–10
lM, while the
2-halopyridin-3-yl analogs 28 and 29 were essentially inactive.
Next, two bromo-pyrazoles (34 and 35) were synthesized, giving
potencies of 5.7 and 4.2
lM, respectively. The corresponding
chloro-pyrazoles (36 and 37) were 4.7 and 12.8
l
M respectively,
while the regioisomeric methyl-chloropyrazole (38) had no
detectable inhibition. In an effort to both understand the effect of
the ortho-substituent and confirm the structure of the product of
the 3-component reaction, we crystallized 32 and obtained an
X-ray structure (Fig. 2). The structure was confirmed and the
presence of the ortho-trifluoromethyl group was shown to position
the phenyl ring with a torsion angle of ꢀ90° from the plane of the
adjacent ring. It is surmised that this orientation is preferred for
inhibition of GALK.
curvefits of the data
(a = 2.035e17) and visualized using
Being that a number of 2-halo substituted heterocycles (27, 31,
34, 35 and 36) had similar potencies, we looked at a preliminary
in vitro ADME profile (Table 4). While all analogs had good PAMPA
permeability (>750 ꢂ 10ꢁ6 cm/s), they differed in kinetic aqueous
solubility and rat liver microsomal stability (RLMS). The
ortho-chloro pyridyl analog (27) had moderate solubility and poor
RLMS, while the corresponding ortho-bromo analog (31) had low
solubility and RLMS. The two bromo-pyrazole analogs (34 and
35) differed greatly in solubility, while the chloro-pyrazole analog
Lineweaver–Burk transformation of the kinetic data. Galactose,
conversely, appeared to show changes in both Km and Vmax in the
presence of 36, suggestive of uncompetitive inhibition with respect
to galactose (Fig. 3C and D). Mixed-model curve fitting yielded
alpha values consistent with uncompetitive activity (a = 0.0007),
suggesting that 36 may bind to the GALK1-galactose complex with
greater affinity than the free enzyme itself. Taken together, these
observations confirm that 36 directly competes with ATP at the
binding site, and suggest that galactose may be required to bind
first, prior to 36 binding and inhibition. Next, 36 was first profiled
(36) gave the best solubility and RLMS (32 lg/mL and 18 min
Please cite this article in press as: Liu, L.; et al. Bioorg. Med. Chem. Lett. (2014), http://dx.doi.org/10.1016/j.bmcl.2014.11.061

L. Liu et al. / Bioorg. Med. Chem. Lett. xxx (2014) xxx–xxx
5
Table 3
SAR of various ortho-substituted aryl groups
R
O
N
N
O
N
H
N
H
a
a
Compounds
R
IC50
(lM)
Max responseb (%)
Compounds
R
IC50
6.4
(lM)
Max responseb (%)
Cl
CF3
Br
N
27
9.9
ꢁ82
33
ꢁ61
N
N
N
Cl
28
29
30
31
ND
ND
4.9
3.1
ꢁ47
ꢁ39
ꢁ64
ꢁ92
34
35
36
37
5.7
4.2
4.7
ꢁ85
ꢁ86
ꢁ86
ꢁ70
HN
N
Br
Br
Br
MeN
HN
N
Cl
N
N
Br
Cl
N
N
12.8
ND
MeN
N
CF3
Cl
N
32
7.6
ꢁ64
38
ꢁ14
N
Me
a
IC50 values were determined utilizing the luminescence GALK-ATP-depletion assay. ND = not determined; for compounds with max response at 57
IC50 was ND.
lM weaker than 50%,
b
Max Resp. represents the % inhibition at 57 lM compound.
Table 4
Early ADME on select GALK1 inhibitors
R
O
N
O
N
N
H
N
H
Compounds
R
RLMSa,b t1/2 PAMPAb
(min)
Aq solubilityb,c
g/mL)
(ꢂ10ꢁ6 cm/s)
(l
Cl
Br
Br
N
N
27
31
7
897
9
7
1515
<1
<1
Figure 2. X-ray structure of 32.
34
35
36
14
798
1351
983
HN
N
against human GALK2 (isozyme with 30% sequence identity to
human GALK1 and 45% homology) and a large portion of the
kinome using DiscoveRx’sÒ Kinome Scan. Compound 36 showed
Br
9
25
32
MeN
N
no inhibition up to 40 lM for GALK2 (Fig. 4) and an extremely
Cl
clean profile against the kinome (see Supplemental). This is likely
due to the significant difference of the ATP binding sites for protein
kinases compared to this GHMP small molecule kinase. The anal-
ogy between these two types of kinases is very low since the
GALK1 active site lacks the typical components that almost all pro-
tein kinases share, such as the hinge region, glycine-rich loop and
activation loop.
18
HN
N
a
b
c
RLMS = rat liver microsome stability.
Performed in-house at NCATS.
Kinetic aq solubility in PBS 7.4 buffer from a 10 mM DMSO compound stock
solution.
One of the major phenotypes of patients with Classic Galacto-
semia is very high gal-1-p levels in their red blood cells, fibroblasts
and other tissue. With access to primary patient fibroblasts from
the University of Utah, we looked for the ability of 36 to lower
gal-1-p levels in these patient cells (Fig. 5). Glucose challenge
resulted in a ratio of gal-1-p/protein of 12 ng/lg, while galactose
challenge resulted in a ratio of 33 ng/
lg. Upon pretreatment with
compound followed by galactose challenge, 9, our inactive control,
Please cite this article in press as: Liu, L.; et al. Bioorg. Med. Chem. Lett. (2014), http://dx.doi.org/10.1016/j.bmcl.2014.11.061

6
L. Liu et al. / Bioorg. Med. Chem. Lett. xxx (2014) xxx–xxx
Table 5
Additional early ADME for 36
GALK1
GALK2
0
NH
N
Cl
O
-50
N
N
O
N
N
H
H
36
-100
In vitro ADME
-8
-7
-6
-5
-4
PBS buffer (pH = 7.4) stability (% remaining after 48 h)
5 mM aqueous glutathione stability (% remaining after 48 h)
Mouse microsomal stability (t1/2
Mouse plasma stability (t1/2
>95%
>95%
34 min
>60 min
log [36], M
)
Figure 4. Dose–response curve for 36 assayed against GALK1 and GALK2.
)
compounds described herein should represent useful tools for
studying galactose metabolism in general as well as understanding
the role of GALK1 in Classic Galactosemia.
gave no reduction in gal-1-p, while 36 gave a dose–response reduc-
tion and reduced levels below the glucose challenge level in this
primary patient fibroblasts (Fig. 5A). Both 9 and 36 had no signifi-
cant effect on viability of the patient fibroblasts as determined by
total protein content (Fig. 5B).
Acknowledgements
In summary, following a qHTS campaign for inhibitors of
GALK1, we optimized a spiro-benzoxazole-containing series to
low micromolar potency with reasonable early ADME properties.
SAR revealed that the spiro group, or a 2-substituted aryl group
was required to maintain the activity. The use of 2-substituted
heteroaryl groups was key to obtaining compounds with decent
solubility, allowing for MOA studies that found this series to be
ATP competitive. Analog 36 was assessed for its ability to lower
We thank Danielle van Leer, Misha Itkin, Crystal McKnight,
Christopher LeClair and Paul Shinn for assistance with compound
management. We also thank Arnold L. Rheingold and Curtis Moore
at UCSD for X-ray determination of 32. This research was sup-
ported by the Molecular Libraries Initiative of the National Insti-
tutes of Health Roadmap for Medical Research Grant
U54MH084681 and the Intramural Research Program of the
National Center for Advancing Translational Sciences at the
National Institutes of Health. Research Grant support to K.L.
include R03MH085689 and R01HD074844-2.
gal-1-p in primary patient fibroblasts and gave
a
nice
dose-response with no detrimental effect on cell viability. The
Figure 3. (A) Initial velocity plot of competition relationship between 36 and ATP; (B) Lineweaver–Burk transformation of 36: ATP initial rates; (C) Initial velocity plot of
competition relationship between 36 and galactose; (D) Lineweaver–Burk transformation of 36: galactose initial rates.
Please cite this article in press as: Liu, L.; et al. Bioorg. Med. Chem. Lett. (2014), http://dx.doi.org/10.1016/j.bmcl.2014.11.061

L. Liu et al. / Bioorg. Med. Chem. Lett. xxx (2014) xxx–xxx
7
Figure 5. (A) Gal-1-p levels measured in primary galactosemic patient fibroblasts after either glucose, galactose and/or galactose and compound treatment regimes. (B) Total
protein content after designated treatment in primary galactosemic patient fibroblasts.
D’Anna, K.; Gentile, J.; Guo, C. Y.; Hecht, L.; Jackson, R.; Jansma, B. M.; Li, Y.; Lip,
V.; Miller, D. T.; Murray, M.; Power, L.; Quinn, N.; Rohr, F.; Shen, Y.; Skinder-
Meredith, A.; Timmers, I.; Tunick, R.; Wessel, A.; Wu, B. L.; Levy, H.; Elsas, L.;
Supplementary data
Supplementary data (synthetic procedures, selected analytical
data, in vitro and cell-based experimental) associated with this
article can be found, in the online version, at http://dx.doi.org/
10.1016/j.bmcl.2014.11.061.
Berry, G. T. J. Inherit. Metab. Dis. 2012, 35, 279.
8. Donnell, G. N.; Bergren, W. R.; Perry, G.; Koch, R. Pediatrics 1963, 31, 802.
9. Webb, A. L.; Singh, R. H.; Kennedy, M. J.; Elsas, L. J. Pediatr. Res. 2003, 53, 396.
10. Guerrero, N. V.; Singh, R. H.; Manatunga, A.; Berry, G. T.; Steiner, R. D.; Elsas, L.
J. J. Pediatr. 2000, 137, 833.
11. Robertson, A.; Singh, R. H.; Guerrero, N. V.; Hundley, M.; Elsas, L. J. Genet. Med.
2000, 2, 142.
12. Bosch, A. M.; Bakker, H. D.; van Gennip, A. H.; van Kempen, J. V.; Wanders, R. J.
A.; Wijburg, F. A. J. Inherit. Metab. Dis. 2002, 25, 629.
13. Wierenga, K. J.; Lai, K.; Buchwald, P.; Tang, M. J. Biomol. Screening 2008, 13,
415–423.
14. Odejinmi, S. I.; Rascon, R. G.; Tang, M.; Vankayalapati, H.; Lai, K. ACS Med. Chem.
Lett. 2011, 2, 667.
15. Boxer, M. B.; Shen, M.; Tanega, C.; Tang, M.; Lai, K.; Auld, D. S. Probe Reports
from the NIH Molecular Libraries Program [Internet]. Bethesda (MD):
National Center for Biotechnology Information (US); 2010 Mar 18. updated
2011 Mar 3.
References and notes
1. Leloir, L. F. Arch. Biochem. Biophys. 1951, 33, 186.
2. Berry, G. T.; Walter, J. H. Disorders of Galactose Metabolism. In Inborn Metabolic
Diseases-Diagnosis and Treatment; Fernandes, J., Saudubray, M., van den Berghe,
G., Walter, J. H., Eds., 5th ed.; Springer: Verlag, Heidelberg, 2011.
3. Berry, G. T. Galactosemia. In: NCBI GeneReviewsTM. University of Washington,
Seattle. http://www.ncbi.nlm.nih.gov/books/NBK1518/. Last Update April 3, 2014.
4. Gitzelmann, R. Eur. J. Pediatr. 1995, 154, S45.
5. Waggoner, D. D.; Buist, N. R.; Donnell, G. N. J. Inherit. Metab. Dis. 1990, 13, 802.
6. Schweitzer, S.; Shin, Y.; Jakobs, C.; Brodehl, J. Eur. J. Pediatr. 1993, 152, 36.
7. Waisbren, S. E.; Potter, N. L.; Gordon, C. M.; Green, R. C.; Greenstein, P.;
Gubbels, C. S.; Rubio-Gozalbo, E.; Schomer, D.; Welt, C.; Anastasoaie, V.;
16. Potapov, A. Y.; Shikhaliev, Kh. S.; Krylsky, D. V.; Krisin, M. Yu. Chem. Heterocycl.
Compd. 2006, 42, 1338.
Please cite this article in press as: Liu, L.; et al. Bioorg. Med. Chem. Lett. (2014), http://dx.doi.org/10.1016/j.bmcl.2014.11.061