Synthesis and biological evaluation of 2-aminoimidazole/carbamate hybrid anti-biofilm and anti-microbial agents

Article abstract of DOI:10.1016/j.bmcl.2010.12.057

The successful marriage of structural features from our 2-aminoimidazole and menthyl carbamate classes of anti-biofilm agents has resulted in the development of a novel hybrid scaffold of biofilm modulators. The compounds were evaluated against a panel of

Full text of DOI:10.1016/j.bmcl.2010.12.057

Bioorganic & Medicinal Chemistry Letters 21 (2011) 1257–1260
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Synthesis and biological evaluation of 2-aminoimidazole/carbamate hybrid
anti-biofilm and anti-microbial agents
⇑
Steven A. Rogers, Erick A. Lindsey, Daniel C. Whitehead, Trey Mullikin, Christian Melander
Department of Chemistry, North Carolina State University, Raleigh, NC 27695, United States
a r t i c l e i n f o
a b s t r a c t
Article history:
The successful marriage of structural features from our 2-aminoimidazole and menthyl carbamate classes
of anti-biofilm agents has resulted in the development of a novel hybrid scaffold of biofilm modulators.
The compounds were evaluated against a panel of four bacterial strains for anti-biofilm and anti-micro-
bial activity.
Received 12 November 2010
Accepted 13 December 2010
Available online 22 December 2010
Ó 2010 Elsevier Ltd. All rights reserved.
Keywords:
Biofilms
Anti-biofilm
Antibiotic
2-Aminoimidazole
Menthyl carbamate
Biofilms represent a particularly hardy phenotype of bacterial
growth.1 Owing to their encasement in a robust extracellular ma-
trix of biomolecules, bacteria in these surface-adhered communi-
ties are uniquely resilient, often displaying resistance toward
conventional antibiotics, antiseptics, and host defense mecha-
nisms.2 Indeed, more than 80% of all bacterial infections are the di-
rect result of biofilms comprising medically relevant pathogens.3
Biofilms have been implicated in persistent infections of medical
implants,4 and are responsible for the mortality and morbidity of
cystic fibrosis (CF) patients.5
observed biological activity of these compounds. These studies
have culminated in the discovery of our current lead compound
5 that is active against both Gram-positive and Gram-negative
bacteria.
Very recently we disclosed a new, structurally unrelated class of
molecules based on a menthyl carbamate scaffold (e.g., 7, Fig. 2)
that possess potent non-microbicidal biofilm inhibition activity
against various Staphylococcal strains.14 We arrived at 7 through
the systematic optimization of the SCRC3P79 (Cytophaga sp.)
Despite the involvement of bacterial biofilms in a host of med-
ical maladies, the development of small, drug-like compound clas-
ses that influence their formation and maintenance has lagged
significantly.6 Currently, relatively few scaffolds are known to pos-
sess anti-biofilm activity, and these include homoserine lactones,7
brominated furanones,8 and ursine triterpenes.9 Additionally, com-
puter aided drug design protocols10 and high throughput screening
methods11 have also led to the discovery of a few novel scaffolds
that possess anti-biofilm activity. Despite these advances, potent
biofilm modulators are still sorely underdeveloped.
Our group has developed an array of novel molecular scaffolds
that both inhibit and disperse bacterial biofilms across order, class,
and phylum via a non-microbicidal mechanism.12 Our inspiration
for the design of these molecules was to extract and systematically
optimize structural motifs embedded within the marine natural
product bromoageliferin (1, Fig. 1).13 The 2-aminoimidazole
(2-AI, highlighted in red) heterocycle has proven crucial for the
N
NH2
H2N
N
H
NH2
TAGE, 2
H2N
NH
O
N
N
N
O
N
H
H
N
Br
H2N
N
N
H
N
H
Br
H2N
Br
N
H
Dihydrosventrin, 3
Br
HN
O
H
N
N
NH
H2N
11
O
N
H
Br
Bromoageliferin, 1
RA-11, 4
N
N
5
H2N
N
N
N
H
5
⇑
Corresponding author.
E-mail address: Christian_Melander@NCSU.edu (C. Melander).
Figure 1. 2-Aminoimidazole anti-biofilm agents based on Bromoageliferin.
0960-894X/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2010.12.057

1258
S. A. Rogers et al. / Bioorg. Med. Chem. Lett. 21 (2011) 1257–1260
H
H
N
N
O
O
88 analogues
NaH, DMF
5
N
O
O
7
(eq. 1)
N
O
SCRC3P79 metabolite, 6
I
6
N
O
(-)-menthyl carbamate anti-biofilm agent, 7
84%
11
Figure 2. (À)-Menthyl carbamate anti-biofilm agent.
1.
N
N3
n
H2N
N
N
12a-c, n = 1,2,3
Boc
N
N
11
8a-c (eq. 2)
n
CuSO4, Na Ascorbate,
EtOH, DCM, H2O
2. TFA/DCM, then HCl/MeOH
56-95%
N
NH2
HCl
N
N
O
H
N
O
1. H2, 10% Pd/C, MeOH
12a-c
9a-c (eq. 3)
8a-c, n = 1,2,3
2. (-)-menthyl chloroformate, TEA, DCM
3. TFA/DCM, then HCl/MeOH
71-81%
O
N
N
O
O
n
H2N
HCl
H
N
H
1. (-)-menthyl chloroformate, TEA, DCM
N3
n NH2
10a-b
(eq. 4)
9a-c, n = 1,2,3
2.
N
13a-b, n = 1,2
H2N
14
N
O
Boc
N
CuSO4, Na Ascorbate, tBuOH, DCM, H2O
N
N
N
H
5 N
n
H2N
HCl
N
H
3. TFA/DCM, then HCl/MeOH
41-42%
O
10a-b, n = 2,3
1.
Figure 3. Hybrid 2-AI/menthyl carbamate anti-biofilm targets.
N
H
15
O
6
N
5 N
N
bacterial metabolite 6.13 Unlike our 2-aminoimidazole scaffold, the
menthyl carbamate series lacked the ability to disperse pre-formed
biofilms and demonstrated poor anti-biofilm properties against
non-Staphylococcal strains. Nonetheless, these shortcomings were
somewhat offset by the trivial synthesis of compounds such as 7.
We next sought to investigate hybrid scaffolds that included
structural motifs from both classes of molecules. Namely, we pro-
posed a series of targets that blended the 2-AI head group from our
bromoageliferin analogues (2–5) with the menthyl carbamate
(highlighted in blue) moiety in our most recent class of biofilm
inhibitors (e.g., 7). Presented herein is an account of the successful
marriage of these two classes of biofilm inhibitors and the evalua-
tion of these hybrid structures as anti-biofilm and anti-microbial
agents.
Initially we targeted three separate scaffolds that combined
structural elements of the 2-AI and menthyl carbamate motifs
(Fig. 3). Compounds 8a–c represent a series of three molecules that
comprise the entire menthyl carbamate lead compound 7 directly
tethered to the 2-aminoimidazole head group with an intervening
triazole similar to our current lead compound 5. Compounds 9a–c
represent menthyl carbamate analogues of 2-AI amide derivatives
dihydrosventrin (3) and RA-11 (4). Finally, 10a–b were designed to
closely resemble our lead compound 5 by replacing the aryl olefin
in 5 with a menthyl carbamate linkage.
Scheme 1 details the preparation of compounds 8a–10b (see
Supplementary data for full details). Compound 714 was alkylated
with 8-iodo-1-octyne, providing carbamate derivative 11 (Eq. 1).
Azides 12a–c12 were each independently reacted with alkyne 11
in the presence of copper sulfate and sodium ascorbate,15 followed
by Boc deprotection and HCl exchange to return the target hybrid
structures 8a–c (Eq. 2). Alternatively, 2-AI carbamate analogues
9a–c were prepared from 2-AI azides 12a–c via an efficient three
step sequence including azide reduction, carbamate formation,
and Boc deprotection/HCl exchange (Eq. 3).12
12a
H2N
HCl
6 H
(eq. 5)
N
N
CuSO4, Na Ascorbate,
EtOH, DCM, H2O
2. TFA/DCM,
N
H
16
then HCl/MeOH
76%
Scheme 1. Preparation of 2-AI/menthyl carbamate hybrids.
Table 1
Biofilm inhibition (IC50 values) against various bacterial strains
Compound
MRSAa
PA14a
A. baumannii (ATCC# 19606)a
8a
8b
8c
9a
9b
9c
10a
10b
16
—
—
—
—
—
—
>200
>200
>200
18.0 5.1
18.0 5.3
—
58.7 1.5
40.3 5.2
>200
19.2 2.0
18.4 0.95
16.7 1.5
19.2 2.0
—
16.7 1.2
94.9 0.2
—
29.9 5.8
20.5 4.9
>200
>200
a
IC50 values are in lM.
The 2-AI triazole analogues 10a–b were generated by initially
reacting amino azides 13a–b with (À)-menthyl chloroformate in
the presence of triethylamine. The azido carbamates were then
coupled with known 2-AI alkyne 1412 via the Cu(I) mediated click
reaction.15 Antibiofilm agents 10a–b were isolated after Boc re-
moval and HCl exchange (Eq. 4).
Additionally, we elected to prepare the acetamido analogue of
2-AI/menthyl carbamate hybrid 8a as a control compound featur-
ing the 2-AI head group capped with an acetamido group in lieu
of the (À)-menthyl carbamate moiety (Eq. 5). The 2-AI azide 12a
was reacted with acetamido-alkyne 15 under copper mediated
click conditions15 to generate the corresponding 2-AI 1,4-disubsti-

S. A. Rogers et al. / Bioorg. Med. Chem. Lett. 21 (2011) 1257–1260
1259
tuted 1,2,3-triazole, which was then deprotected to provide acet-
amido analogue 16 after HCl salt exchange.
Table 2
MIC values against various bacterial strains
Compound MRSAa S. aureus
PA14a A. baumannii
MDRABa
After their synthesis, compounds 8a–10b and 16 were initially
screened for their ability to inhibit biofilms of the medically rele-
vant bacteria methicillin resistant Staphylococcus aureus (MRSA),
S. aureus (ATCC# 29213), Pseudomonas aeruginosa (PA14), and Aci-
netobacter baumannii (ATCC# 19606). We subjected each com-
pound to a dose–response study in an attempt to determine the
IC50 value for biofilm inhibition (i.e., the concentration necessary
to inhibit 50% of biofilm formation) as judged by a crystal violet re-
porter assay.16 The data for these experiments is collected in Table
1. During the course of these experiments, most of the compounds
displayed a precipitous drop in biofilm inhibition activity over a
narrow concentration range. For example, the compounds would
(ATCC# 29213)a
(ATCC# 19606)a
8a
8b
8c
8
4
4
4
4
4
>128
>128
>128
32
16
32
>128
32
16
>128
>128
32
16
8
128
64
>128
9a
9b
9c
32
8
8
32
8
8
64
16
64
8
10a
10b
16
64
16
>128
16
16
>128
>128
64
64
64
>128
>128
a
MIC values are in lg/mL.
exhibit >90% inhibition at 10
lM concentrations, but a dismal
<10% inhibition at 5 M. Such behavior is commonly diagnostic
l
32
lg/mL against the ATCC# 19606 strain, but was completely
of an underlying microbicidal mechanism for biofilm inhibition.
Such dose–response behavior precludes the ability to calculate a
reliable IC50 value, and compounds exhibiting this behavior are
indicated with a dash (À) in Table 1. It should be noted that all
of the compounds exhibited this type of behavior for S. aureus
(ATCC# 29213) (data not depicted in Table 1).
inactive against MDRAB (>128
l
g/mL). Analogue 8c exhibited no
toxicity against either A. baumannii strain. In the 9a–c series, a
structure activity trend emerged whereby lengthening the inter-
vening methylene tether in between the 2-AI head and (À)-men-
thyl carbamate tail resulted in increased potency. Against both A.
baumannii strains MIC values of 32, 16, and 8
for 9a, 9b, and 9c, respectively. Finally, compound 10a returned
MIC values of 64 and 128 g/mL against ATCC# 19606 and MDRAB,
respectively, while 10b gave an MIC of 64 g/mL against both A.
lg/mL were found
Some compounds, however, returned dose–response data suit-
able for the determination of IC50 values. Against MRSA, compounds
l
10a and 10b returned IC50 values of 29.9 and 20.5
lM, respectively.
l
AgainstPA14, compounds 8a–c exhibited no anti-biofilmproperties,
baumannii strains. In the case of control compound 16, we did
not observe any anti-microbial behavior up to the highest concen-
tration tested (128 lg/mL). This result serves to highlight the
while 9a and 9b each returned an IC50 value of 18
10a and 10b gave IC50 values of 58.8 and 40.3
l
l
M. Additionally,
M, respectively
against PA14. For A. baumannii, IC50 values of 19.2, 18.4, 16.7, 19.2,
16.7, and 94.9 were obtained for compounds 8a–c, 9a, 9c, and 10a,
respectively. Control acetamido compound 16 exhibited no anti-
biofilm activity for any of the four bacterial strains at the highest
concentration tested (200 lM), thus confirming the necessity of
the carbamate moiety for anti-biofilm activity.
importance of both the 2-AI head and the (À)-menthyl carbamate
functionality in eliciting the observed anti-microbial properties.
Given their observed anti-biofilm activity via both toxic and
non-toxic means, we were eager to evaluate whether or not several
of the 2-AI/menthyl carbamate hybrids could disperse pre-formed
bacterial biofilms. This study was of particular interest given the
dichotomy between to two parent scaffolds. While our 2-AI leads
(e.g., Scheme 1) effectively disperse pre-formed biofilms across
order, class, and phylum, our menthyl carbamate lead 7 failed to
exhibit biofilm dispersal capability.
In order to assess the potential for biofilm dispersal, we chose
compounds 8a and 9c, based on the fact that they exhibited the
lowest MIC values against the S. aureus and A. baumannii strains.
Additionally, we elected to screen 8c for non-toxic dispersal activ-
ity given that it inhibited the formation of A. baumannii (ATCC#
19606) films in a non-microbicidal fashion. Pre-formed biofilms
from MRSA, S. aureus (ATCC # 29213), and A. baumannii (ATCC#
19606) were treated with lead compounds 8a, 8c, and 9c. Dose–re-
sponse curves were generated to determine the EC50 values for bio-
film dispersal (i.e., concentration required to disperse 50% of a pre-
formed biofilm, see Table 3).
We next conducted growth curves at the IC50 concentration for
each of the compounds in Table 1 (where appropriate) to assess the
viability of planktonic bacteria. Growth curve analysis against
MRSA revealed that compound 10a and 10b reduced bacterial
growth at their respective IC50 value, thus some of their anti-bio-
film activity was due to inhibition of bacterial growth. The same
was true for compounds 8a, 8b, 9a, 9c, and 10a against A. bauman-
nii (ATCC# 19606). Importantly, however, compounds 9a, 9b, and
10a–b inhibited PA14 biofilm formation via non-microbicidal
activity. Additionally, compound 8c was found to inhibit A. bau-
mannii (ATCC# 19606) biofilms in a non-toxic fashion.
Given that most of the 2-AI/carbamate hybrids exhibited
anti-biofilm activity by toxic means, we elected to evaluate their
potential as anti-microbial agents. To determine the extent of
microbicidal activity, we measured the minimum inhibitory con-
centration (MIC) of each compound against MRSA, S. aureus
(ATCC# 29213), PA14, and A. baumannii (ATCC# 19606). Addition-
ally, we evaluated the compounds against multi-drug resistant A.
baumannii (MDRAB, ATCC# BAA-1605). The data for this study is
summarized in Table 2.
While compound 8a failed to disperse pre-formed Staphylococ-
cal biofilms, it exhibited an EC50 value of 37.5
mannii (ATCC# 19606). Compound 8c gave EC50 values of 38.1
and 32.5 M against MRSA and S. aureus (ATCC# 29213) respec-
tively; however, these concentrations are microbicidal to plank-
lM against A. bau-
l
Compounds 8a–c, as well as 9b and 9c proved to be especially
active against the Staphylococcal strains, each exhibiting MIC val-
tonic bacteria (MIC = 4 lg/mL against both strains). Conversely,
ues of 8
MIC values of 32, 64, and 16
and 32, 16, and 16 g/mL against S. aureus (ATCC# 29213). In keep-
ing with the biofilm inhibition data collected in Table 1, all of the
compounds exhibited relatively high MIC values for PA14. Com-
pounds 8a–c and 10a were found to be non-toxic at the highest
l
g/mL or lower. Compounds 9a, 10a, and 10b exhibited
l
g/mL respectively against MRSA,
Table 3
l
Biofilm dispersal (EC50 values) against various bacterial strains
Compound
MRSAa
S. aureus
A. baumannii
(ATCC# 29213)a
(ATCC# 19606)a
8a
8c
9c
>200
38.1 5.7
53.5 5.1
>200
32.5 4.6
38.8 4.5
37.5 2.3
20.6 1.7
68.3 2.7
concentration tested (128
10b exhibited MIC values of 32, 64, 64, and 64
Against the two A. baumannii strains, compound 8a returned an
MIC value of 16 g/mL. Compound 8b exhibited an MIC value of
lg/mL) while compounds 9a–c and
l
g/mL, respectively.
a
EC50 values are in lM.
l

1260
S. A. Rogers et al. / Bioorg. Med. Chem. Lett. 21 (2011) 1257–1260
Table 4
V Foundation for a Predoctoral Jimmy V Scholar award (S.A.R.)
and the NCSU Office of Undergraduate Research for an undergrad-
uate research fellowship (T.M.).
Blood lysis (HD50) assay of compounds 8a–10b and 16
a
Compound
HD50
8a
93.8 3.5
8b
8c
9a
9b
9c
10a
10b
16
46.0 4.9
42.7 1.7
463.7 0.7
228.8 29.9
473.7 15.0
474.0 7.2
317.6 14.6
>800
Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.bmcl.2010.12.057.
References and notes
1. Musk, D. J.; Hergenrother, P. J. Curr. Med. Chem. 2006, 12, 2163.
2. Rassmussen, T. B.; Givskov, M. Int. J. Med. Microbiol. 2006, 296, 149.
3. Davies, D. Nat. Rev. Drug Discov. 2003, 2, 114.
a
HD50 values are in lM.
4. Donlan, R. M.; Costerton, J. W. Clin. Microbiol. Rev. 2002, 15, 167; Stewart, P. S.;
Costerton, J. W. Lancet 2001, 358, 135.
5. Costerton, J. W.; Stewart, P. S.; Greenberg, E. P. Science 1999, 284, 1318.
6. Richards, J. J.; Melander, C. ChemBioChem 2009, 10, 2287.
7. Geske, G. D.; O’Neill, J. C.; Blackwell, H. E. Chem. Soc. Rev. 2008, 37, 1432;
Manefield, M.; Kjelleberg, S.; Givskov, M. Curr. Med. Chem.: Anti-Infect. Agents
2003, 2, 213.
8. Hentzer, M.; Wu, H.; Andersen, J. B.; Riedel, J. B.; Rassmussen, J. B.; Bagge, N.;
Kumar, N.; Schembri, M. A.; Song, Z. J.; Kristoffersen, P.; Manefield, M.;
Costerton, J. W.; Molin, S.; Eberl, L.; Steinberg, P.; Kjelleberg, S.; Hoiby, N.;
Givskov, M. EMBO J. 2003, 22, 3803; Wu, H.; Song, Z.; Hentzer, M.; Andersen, J.
B.; Molin, S.; Givskov, M.; Hoiby, N. J. Antimicrob. Chemother. 2004, 53, 1054.
9. Hu, J. F.; Garo, E.; Goering, M. G.; Pasmore, M.; Yoo, H. D.; Esser, T.; Sestrich, J.;
Cremin, P. A.; Hough, G. W.; Perrone, P.; Lee, Y. S. L.; Le, N. T.; O’Neil-Johnson,
M.; Costerton, J. W.; Eldridge, G. R. J. Nat. Prod. 2006, 69, 118; Ren, D. C.; Zuo, R.
J.; Barrios, A. F.; Bedzyk, L. A.; Eldridge, G. R.; Passmore, M. E.; Wood, T. K. Appl.
Environ. Microbiol. 2005, 71, 4022.
compound 8c also showed non-toxic (MIC >128
dispersal activity against A. baumannii (ATCC# 19606) with an EC50
value of 20.6 M (13.1 g/mL). Finally, compound 9c had EC50 val-
ues of 53.5, 38.8, and 68.3 M against MRSA, S. aureus (ATCC#
29213), and A. baumannii (ATCC# 19606), respectively. In addition
to the data collected in Table 3, we also screened the entire library
against pre-formedP. aeruginosa (PA14) biofilms owing to their rel-
atively low toxicity in the MIC studies against that strain. Unfortu-
nately, none of the compounds in this study effectively dispersed
pre-formed PA14 biofilms.
In a final set of experiments, all of the compounds in the study
were preliminarily assessed for mammalian cytotoxicity using a
red blood cell hemolysis assay using defibrinated sheep blood.17
HD50 values (i.e., the concentration at which 50% hemolysis is ob-
served) are collected in Table 4. No hemolysis is observed at the
MIC concentrations for the lead antibiotics depicted in Table 2.
In conclusion, we have introduced a new class of hybrid 2-
aminoimidazole/menthyl carbamate anti-biofilm agents that exhi-
bit the ability to inhibit biofilm formation and to disperse mature,
pre-formed biofilms. While most of the 2-AI/carbamate hybrids
elicited their anti-biofilm activity via underlying microbicidal
means, this study also resulted in a panel of non-toxic inhibitors
for the P. aeruginosa strain PA14 (i.e., compounds 9a, 9b, 10a,
and 10b). Additionally, compound 8c both inhibited and dispersed
biofilms of Acinetobacter baumannii (ATCC# 19606) in a non-toxic
fashion.
lg/mL, vide supra)
l
l
l
10. Zeng, Z.; Qian, L.; Cao, L.; Tan, H.; Huang, Y.; Xue, X.; Shen, Y.; Zhou, S. Appl.
Microbiol. Biotechnol. 2008, 79, 119.
11. Musk, D. J.; Banko, D. A.; Hergenrother, P. A. Chem. Biol. 2005, 12, 789; Junker, L.
M.; Clardy, J. Antimicrob. Agents Chemother. 2007, 51, 3582.
12. Huigens, R. W., III; Richards, J. J.; Parise, G.; Ballard, T. E.; Zeng, W.; Deora, R.;
Melander, C. J. Am. Chem. Soc. 2007, 129, 6966; Richards, J. J.; Ballard, T. E.;
Melander, C. Org. Biomol. Chem. 2008, 6, 1356; Richards, J. J.; Ballard, T. E.;
Huigens, R. W., III; Melander, C. ChemBioChem 2008, 9, 1267; Richards, J. J.;
Huigens, R. W., III; Ballard, T. E.; Basso, A.; Cavanagh, J.; Melander, C. Chem.
Commun. 2008, 1698; Rogers, S. A.; Melander, C. Angew. Chem., Int. Ed. 2008, 47,
5229; Ballard, T. E.; Richards, J. J.; Wolfe, A. L.; Melander, C. Chem. Eur. J. 2008,
14, 10745; Ballard, T. E.; Richards, J. J.; Aquino, A.; Reed, C. S.; Melander, C. J.
Org. Chem. 2009, 74, 1755.
13. Yamada, A.; Kitamura, H.; Yamaguchi, K.; Fukuzawa, S.; Kamijima, C.; Yazawa,
K.; Kuramoto, M.; Wang, G. Y. S.; Fujitani, Y.; Uemura, D. B. Chem. Soc. Jpn. 1997,
70, 3061; Al Mourabit, A.; Poiter, P. Eur. J. Org. Chem. 2001, 2, 237.
14. Rogers, S. A.; Whitehead, D. C.; Mullikin, T.; Melander, C. Org. Biomol. Chem.
2010, 8, 3857.
15. Kolb, H. C.; Finn, M. G.; Sharpless, K. B. Angew. Chem., Int. Ed. 2001, 113, 2056;
Rodionov, V. O.; Fokin, V. V.; Finn, M. G. Angew. Chem., Int. Ed. 2005, 44, 2210;
Bock, V. D.; Hiemstra, H.; van Maarseveen, J. H. Eur. J. Org. Chem. 2006, 51.
16. O’Toole, G. A.; Kolter, R. Mol. Microbiol. 1998, 30, 295.
Acknowledgments
17. Liu, Z. G.; Brady, A.; Young, A.; Rasimich, B.; Chen, K.; Zhou, C. H.; Kallenbach, N.
R. Antimicrob. Agents Chemother. 2007, 51, 597.
The authors would like to thank the University of North
Carolina General Administration Competitive Research Fund, the