Palmitic Acid Analogs Exhibit Nanomolar Binding Affinity
for the HIV-1 CD4 Receptor and Nanomolar Inhibition of
Elena E. Paskaleva1, Jing Xue2, David Y-W. Lee3, Alexander Shekhtman2, Mario Canki1*
1Center for Immunology and Microbial Disease, Albany Medical College, Albany, New York, United States of America, 2Department of Chemistry, State University of New
York at Albany, Albany, New York, United States of America, 3Mailman Research Center, McLean Hospital, Harvard Medical School, Belmont, Massachusetts, United States
Background: We recently reported that palmitic acid (PA) is a novel and efficient CD4 fusion inhibitor to HIV-1 entry and
infection. In the present report, based on in silico modeling of the novel CD4 pocket that binds PA, we describe discovery of
highly potent PA analogs with increased CD4 receptor binding affinities (Kd) and gp120-to-CD4 inhibition constants (Ki). The
PA analogs were selected to satisfy Lipinski’s rule of drug-likeness, increased solubility, and to avoid potential cytotoxicity.
Principal Findings: PA analog 2-bromopalmitate (2-BP) was most efficacious with Kd,74 nM and Ki,122 nM, ascorbyl
palmitate (6-AP) exhibited slightly higher Kd,140 nM and Ki,354 nM, and sucrose palmitate (SP) was least efficacious
binding to CD4 with Kd,364 nM and inhibiting gp120-to-CD4 binding with Ki,1486 nM. Importantly, PA and its analogs
specifically bound to the CD4 receptor with the one to one stoichiometry.
Significance: Considering observed differences between Ki and Kd values indicates clear and rational direction for
improving inhibition efficacy to HIV-1 entry and infection. Taken together this report introduces a novel class of natural
small molecules fusion inhibitors with nanomolar efficacy of CD4 receptor binding and inhibition of HIV-1 entry.
Citation: Paskaleva EE, Xue J, Lee DY-W, Shekhtman A, Canki M (2010) Palmitic Acid Analogs Exhibit Nanomolar Binding Affinity for the HIV-1 CD4 Receptor and
Nanomolar Inhibition of gp120-to-CD4 Fusion. PLoS ONE 5(8): e12168. doi:10.1371/journal.pone.0012168
Editor: Linqi Zhang, Tsinghua University, China
Received May 19, 2010; Accepted July 16, 2010; Published August 13, 2010
Copyright: ? 2010 Paskaleva et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits
unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This work was supported by NIH grant number AT003371 to MC and American Diabetes Association Career Development Award number 1-06-CD-23
and 1R01GM085006 to AS. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing Interests: The authors have declared that no competing interests exist.
* E-mail: email@example.com
Recently we isolated and identified palmitic acid (PA) as a novel
natural small molecule that inhibits HIV-1 fusion and infection by
the mechanism of binding to the CD4 receptor and blocking
gp120-to-CD4 attachment [1,2,3]. We showed that PA binds to
the CD4 receptor with Kd,1.5 mM , and it blocks gp120-to-
CD4 attachment with Ki,2.53 mM (submitted for publication).
PA also inhibited R5-tropic HIV-1 infection in cervical explant
model of human vagina and in the underlying cervical submucosa
primary PBL and macrophage cells, without toxicity (submitted for
publication). Collectively, these results indicated potential for PA’s
However, the efficacy of HIV-1 inhibition by PA remains in
submicromolar range, suggesting potential for improving its
efficacy. The PA molecule binds to the CD4 receptor via its
hydrophobic methyl and methelene groups located away from the
PA carboxyl end, and the carboxyl group functions by blocking
efficient gp120-to-CD4 attachment and fusion . Considering
PA’s molecular structure and bifunctional mechanism of inhibi-
tion, we in silico modeled this structure-activity relationship (SAR),
and we searched chemical databases for PA analogs that would
satisfy Lipinski’s rule of drug-likeness . In the present study we
report nanomolar CD4 binding affinities and nanomolar blocking
efficacies of gp120-to-CD4 fusion, by three analogs of PA: 2-BP, 6-
AP, and SP.
PA and gp120 binding sites on CD4 overlap
To gain structural insights into the mechanisms of PA binding
to the CD4 receptor, we used in silico molecular docking based on
the known X-ray structure of the two N-terminal domains of CD4
(aa 26-206) (PDB code 1GC1) and a flexible PA ligand (Figure 1).
We used Autodock 4.0 molecular docking program [5,6], which is
widely used to identify a ligand binding contiguous envelope of
maximum affinity for a given macromolecular structure. The
geometry of PA-CD4 with a highest score is shown in Figure 1A,
and crystal structure of gp120-CD4 (PDB code 1GC1) binding is
shown in Figure 1B. Comparison between PA-CD4 and gp120-
CD4 structures shows the overlapping binding sites for gp120 and
PA, suggesting that PA directly inhibits complex formation
between CD4 and gp120 that is necessary for HIV-1 entry. To
access the importance of hydrophilic and hydrophobic interactions
between PA and CD4, in a close-up of the PA-CD4 binding cavity
we mapped CD4 electrostatic potential onto the molecular
PLoS ONE | www.plosone.org1August 2010 | Volume 5 | Issue 8 | e12168
surface, with blue and red colors representing positively and
negatively charged surfaces, respectively (Figure 1C). PA hydro-
phobic aliphatic chain fits tightly into the CD4 binding cavity,
formed by Phe52, Ile60, Ile62, Leu63, and Leu70. The negatively
charged carboxylic group of PA is in the vicinity of the positively
charged epsilon amino group of Lys61. PA methelene groups
proximal to the PA carboxyl end do not make extensive contacts
with the CD4 binding cavity and are possibly flexible. These
results are in agreement with our previous STD NMR results,
which identified PA binding epitope on CD4 that consists of the
hydrophobic aliphatic chain located away from the PA carboxyl
PA analogs bind to CD4 receptor with nanomolar affinity
We hypothesized that based on the CD4-PA model, by adding
polar groups to the carboxyl terminal of PA as well as by creating
partially negative charge on the hydrocarbon chain close to
carboxyl end we would increase the affinity of PA for CD4. We
used the following criteria in designing PA analogs: 1) The
compounds should be non-toxic, 2) The compounds should be
readily available from the chemical catalogs, and 3) The
compounds should satisfy Lipinski’s rule of five for solubility,
lipophilicity, and hydrogen bond formation. We chose three
compounds that meet these criteria: SP and 6-AP, and 2-BP. 6-AP
is used as a substitute for vitamin C, is approved for human
Figure 1. PA-CD4-gp120 interaction model. A) Molecular docking software Autodock 4.0 was used for blind docking of flexible PA onto rigid
two N-terminal domains of CD4 (PDB code 1GC1). The resultant PA-CD4 conformations were ranked and categorized based on the value of free
energy of binding. 386 out of 1000 docking runs fell into conformations that are ranked with the highest score (216 kcal/mol). The root mean square
deviation of these conformations was 1.2 A suggesting very similar binding modes. One of the ligand bound conformations of PA-CD4 with a highest
score (217 kcal/mol) is shown in cyan (PA aliphatic chain) and red (PA carboxylic terminus). B) Crystal structure of gp120-CD4 (PDB code 1GC1). The
backbone of gp120 is shown by using ribbon model. The N-terminal D1 and D2 domains of CD4 are indicated. Comparison between PA-CD4 and
gp120-CD4 structures shows the overlapping binding sites for gp120 and PA. C) Close-up of the PA-CD4 binding cavity shown in A. PA occupies this
cavity, which is formed by Phe52, Ile60, Ile62, Leu63, and Leu70 of CD4. Electrostatic potential calculated using DelPhi software (B. Honnig’s Lab) was
mapped onto the molecular surface of CD4. Positively and negatively charged surfaces are in blue and red, respectively. Non-polar surface is in white.
We used Discover Studio software (AccelRys) to prepare this figure.
PA Nanomolar Analogs
PLoS ONE | www.plosone.org2 August 2010 | Volume 5 | Issue 8 | e12168
consumption, and is available over the counter (OTC) without
prescription in most pharmacies and health food stores. SP is a non-
toxic sugar ester of PA that is used in food and cosmetics industry,
and although it does not satisfy the Lapinskis’s rule, we used it
because of its increased solubility. 2-BP is a non-toxic compound
used to inhibit palmitoylation [7,8]. As compared to PA, these
compounds have favorable physico-chemical properties shown in
Table 1 and Table S1. Higher solubility, critical micelle
concentration (cmc), and lower lipophilicity of these PA analogs
HIV-1 in human tissue and minimize possible cytotoxic effects.
The tryptophan fluorescence of the soluble extracellular portion
of CD4 (sCD4) was used to estimate binding affinity of PA analogs
forsCD4 in-vitro(Figure 2).Saturating concentrations ofPAanalogs,
2-BP, 6-AP, and SP quenched 50%, 40%, and 60% of the sCD4
tryptophan fluorescence, respectively, and resulted in a red shift of
the emission peak of 4, 3, and 5 nm, respectively. Based on the
fluorescence titration experiments, dissociation constant, Kd, for 2-
BP, 6-AP and SP was estimated to be ,7464 nM, 140636 nM,
and 364677 nM (Figure 2 A, B, and C, respectively).
Since PA analogs form micelles at micromolar concentration it
was important to characterize binding stoichiomentry of PA
analogs to CD4 (Figure 2D). We estimated binding stoichiometry
of the most potent PA analog, 2-BP, by using the fluorescent
titration experiment. The concentration of sCD4 was kept at
200 nM, which is well above the 74 nM dissociation constant of 2-
BP-CD4. In this case, titrated 2-BP was fully bound. The observed
changes in sCD4 fluorescence would level off when 2-BP-sCD4
reached proper stoichiometry. Binding of 2-BP to sCD4 resulted in
fluorescence quenching that was saturated when the concentration
of 2-BP reached 200 mM, indicating that stoichiometry of 2-BP-
CD4 is 1:1. This result proved that 2-BP does not form micelles
when it binds directly to CD4.
PA analogs inhibit in vitro gp120-to-CD4 complex
To determine PA analogs inhibition constant for blocking gp120-
to-CD4 attachment that is independent of in vivo cellular CD4
expression, we performed in vitro gp120 capture ELISA (Figure 3 A–
C). 96-well plates were coated with gp120 (IIIB) and incubated with
biotin-sCD4 in absence or presence of increasing PA analogs
concentrations, as indicated. All three PA analogs inhibited CD4
attachment to HIV-1 X4-tropic (IIIB) gp120 envelope in a dose
dependant manner. 2-BP, 6-PA, and SP blocked gp120-CD4
complex formation with Ki of ,122, 354, and 1486 nM (Figure 3
A, B, and C, respectively). These results are consistent and validate
our CD4 binding affinities results (Figure 2). In reverse experiments,
we coated plates with sCD4, incubated gp120 with increasing
concentrations of 2-BP, 6-AP, and SP, and tested for inhibition of
gp120-CD4 complex formation, which wasnot inhibited, indicating
that PA analogs do not bind to soluble gp120 (not shown). This
result is also consistent with previously published observation that
PA does not bind to gp120 .
PA analogs are not toxic to in vitro cell culture
To ascertain possible toxicity of PA analogs, we tested each
analog in cell culture system (Figure 4). 1G5 cells were treated with
increasing concentrations of 2-BP, 6-AP, or SP, from 100 to
10,000 nM, and viability was determined by MTT assay at 24 h.
(not shown) and at 72 h (Figure 4). Viability for all treatments and
at all concentrations tested, remained $92% for up to 72 h. of
follow-up. These results are consistent with predicted safety for
these PA analogs and lack of toxicity previously reported for PA in
primary PBL and macrophages .
We have demonstrated that palmitic acid is a novel class of
small molecule that binds to the CD4 receptor and blocks gp120-
to-CD4 fusion and HIV-1 infection. We determined Kdfor the
CD4 receptor to be ,1.5 mM . In primary PBL and
macrophages, PA blocked productive X4 and R5-tropic HIV-1
infection with IC50of ,0.34 and 33 mM, respectively . We also
showed that PA inhibited R5 HIV-1 infection in human ex vivo
model of human vagina, demonstrating opportunity for a
microbicide development (submitted for publication).
However, restricted sub-micromolar inhibition efficacy, limits
PA’s clinical utility. To this end we were interested to investigate
the possibility of PA to serve as a model scaffold molecule for
chemical modifications to improve its efficacy to nanomolar range.
We constructed in silico model of PA-CD4 binding and gp120
interaction (Figure 1). This model was supported by our STD-
NMR and gp120 capture ELISA results, which collectively
showed that PA binds to the CD4 receptor via hydrophobic
methyl and methelene groups located away from the PA carboxyl
end, and the carboxyl group functions by blocking efficient gp120-
to-CD4 attachment and fusion . Tight fit of PA into the CD4
cavity formed by Phe52, Ile60, Ile62, Leu63, and Leu70 possibly
precludes extensive modifications of the PA aliphatic chain. At the
same time, modifying the PA carboxyl end and methylene groups
close to the carboxyl end may lead to increase in the PA-CD4
affinity, and adding bulky groups to the carboxyl end may also
increase the ability of PA to block CD4-gp120 interaction
(Figure 1). Based on this model and utilizing Lipinski’s rule of
drug-likeness we searched available chemical databases and
investigated three different PA analogs (Table 1), for their CD4
binding affinities and gp120-to-CD4 inhibition constants that are
independent of in vivo CD4 receptor expression. All three
compound showed nanomolar CD4 binding affinity and nano-
molar gp120-CD4 blocking efficacy (Figure 2 and 3), and were not
toxic to in vitro cell culture, up to 10,000 nM concentrations for
72 hours (Figure 4). Because PA analogs form micelles it was
important to characterize binding stoichiomentry of PA analogs to
CD4 (Figure 2D). Our results showed that the most efficacious 2-
BP analog, bound to CD4 with a 1:1 binding stoichiomentry. This
result signifies that PA analogs inhibit HIV-1 fusion by binding to
the CD4 receptor as a single molecule, and not in the form of
Table 1. Physico-chemical properties of PA analogs.
PAPalmitic acid28 7.24
SP Sucrose palmitate100 4.56 30
*Solubility was calculated directly from chemical structure by using ACD/
PhysChem program (ACD/Lab, Inc) .
**Computed partition coefficients (CLogP) were calculated directly from
chemical structure by using ACD/PhysChem program (ACD/Lab, Inc) .
***Critical micelle concentration (cmc) was determined experimentally by
analyzing changes in 1D proton NMR spectra of the compounds due to
PA Nanomolar Analogs
PLoS ONE | www.plosone.org3 August 2010 | Volume 5 | Issue 8 | e12168
micelles that exist at higher micromolar concentrations. Collec-
tively, our results demonstrate that PA may serve as a model
scaffold molecule to further modify its structure and improve its
efficacy. Considering we only tested commercially available
analogs, we predict that further chemical modifications based on
PA’s known SAR, will lead to additional increase in inhibition
efficacy. However, considering successes of HAART treatment in
AIDS patients, it is not unreasonable to envision utility of PA
based fusion inhibitor to be combined together with antiretroviral
drug targeting a different stage of the virus life cycle, which may
result in an effective and potent inhibitor of HIV-1 infection.
Materials and Methods
All reagents and solvents were obtained from commercial
suppliers and used without further purification. sCD4 (Progenics
Pharmaceuticals, Inc) was 95% pure, according to manufacture’s
specification. The specified chemical purity of 2-BP (2-Bromohex-
adecanoicacid, Sigma-Aldrich), 6-AP (6-O-Palmitoyl-L- ascorbic
acid, Sigma-Aldrich) and SP (sucrose palmitate, Stearinerie
Dubois, France) were better than 98%. 1G5  cells were
obtained from HIV AIDS Research and Reference Reagent
Program, Division of AIDS, NIAID, NIH, and were cultured and
maintained as specified by the reagent protocol.
In silico modeling
We used a standard protocol for Autodock–based blind docking
approach described and successfully implemented by Hetenyi et al
. AutoDock 4.0 uses a force-field based empirical free energy
scoring function . The Lamarckian Genetic Algorithm (LGA)
 was used as a search engine. The active site was defined using
AutoGrid. The grid size was set to 30 A˚630 A˚630 A˚points with
grid spacing of 0.3 A˚centered on the CD4 D1 domain center of
Figure 2. In vitro experiments of PA analogs binding to sCD4. Binding isotherm of the sCD4 tryptophan fluorescence at the wavelength of
350 nm with the increasing concentration of A) 2-BP, B) 6-AP, and C) SP; D) stoichiometric binding of 2-BP to sCD4. Tryptophan fluorescence was
measured using an excitation wavelength of 280 nm. In all cases, we observed tryptophan fluorescence quenching of sCD4 by the increasing
concentration of PA analogs. During stoichiometric binding experiment the concentration of sCD4 was kept at 200 nM, which is well above the
dissociation constant of 2-BP-sCD4, 74 nM. Curve fitting (OriginLab) was performed to find the best values for Kdusing a single site binding isotherm
PA Nanomolar Analogs
PLoS ONE | www.plosone.org4 August 2010 | Volume 5 | Issue 8 | e12168
mass. Step sizes of 1.0 A˚for translation and 50 degree for rotation
were chosen, a maximum number of energy evaluations was set to
4 million. PA was prepared as a flexible ligand possessing 14
torsional rotations and CD4 was treated as a rigid target. The
resultant PA-CD4 conformations were ranked and categorized
based on the value of free energy of binding. The AutoDock free
energy of binding ranged from 217 kcal/mol to 25 kcal/mol.
386 out of 1000 docking runs fell into conformations that are
ranked with the highest score (216 kcal/mol). The root mean
square deviation of these conformations was 1.2 A˚suggesting very
similar binding mode. We used Discover Studio software
(AccelRys) to prepare the figure.
Measurements were performed on a Fluorolog-3 fluorescence
spectrophotometer (HORIBA Jobin Yvon) at 25uC in a 1-ml
stirred cuvette. For fluorescence titration experiments, 100 nM of
sCD4 dissolved in 10 mM phosphate buffer [pH 7.4] and
250 mM NaCl was used, and 100 mM solution of PA analogs,
2-BP, 6-AP, and SP dissolved in dimethylsulfoxide (DMSO) was
added in 10 nM steps. Titrations in the absence of sCD4 and in
the absence PA analogs were performed as reference. Tryptophan
Figure 4. Viability of PA analogs in tissue culture. 1G5 cells were
treated with increasing concentrations of 2-BP, 6-AP, and SP, as
indicated, and cell viability was determined 72 h. after treatment by
MTT assay. % Viable cells were calculated from untreated mock control,
which was taken as 100% viable. Bars indicate 6 SD, representative of 2
Figure 3. Inhibition of gp120-CD4 complex formation. Inhibition
of gp120-CD4 complex formation was investigated by gp120 capture
ELISA. Envelope gp120 (IIIB) protein was captured on 96 well plates,
washed, and incubated in the presence of CD4-biotin alone or in the
presence of increasing concentrations of A) 2-BP, B) 6-AP, or C) SP, as
indicated. Strepavidin-HRP was added, and then developed by addition
of o-Phenylenediamine dihydrochloride (OPD) substrate. Colorimetric
reaction was stopped by adding 1N HCl, and read at 490 nm. Percent of
gp120-CD4 binding was calculated from gp120-CD4 complex formation
in the absence of any inhibitor. pKi(-log Ki, M) was calculated and
plotted in Prizm (GraphPad Software), and inhibition constant, Ki, was
calculated by using the equation Ki=IC50/(1+[CD4]/Kd) , based on
IC50concentration of bound CD4, [CD4] =50 nM, and CD4 binding
affinity for gp120, Kd=5 nM. Representative of three experiments, all
data are mean 6 SD.
PA Nanomolar Analogs
PLoS ONE | www.plosone.org5 August 2010 | Volume 5 | Issue 8 | e12168
fluorescence was measured using an excitation wavelength of Download full-text
280 nm. The fluorescence emission signal was subtracted from the
signal of the reference titrations, and the differences adjusted by
the dilution factor were plotted against the final concentration of
added PA analogs. Curve fitting (OriginLab) was performed to
find the best values for Kdusing a single site binding isotherm
gp120-CD4 capture ELISA
Inhibition of gp120-CD4 complex formation was investigated
by CD4-to-gp120 capture ELISA, in accordance with manufac-
turer’s instructions (ImmunoDiagnostics, Inc., MA), and as
previously described . Briefly, envelope gp120 (IIIB) protein
(ImmunoDiagnostics, Inc., MA) was captured on 96-well plates,
washed, and incubated in the presence of 50 nM biotin-
conjugated sCD4 (Immunodiagnostics) alone or in the presence
of serial dilutions of PA analogs, as indicated. Strepavidin-HRP
was added, and then developed by addition of OPD substrate.
Colorimetric reaction was stopped by adding 1N HCl, and read at
490 nm. Percent of gp120-CD4 binding was calculated from
gp120-CD4 complex formation in the absence of any inhibitor.
In vitro cell culture toxicity of PA analogs
For determination of viability, 1G5 cells were seeded in 12 well
plates at 16106cells/well, treated with increasing concentrations
of PA analogs, from 100 to 10,000 nM (0.1 to 10 mM), as
indicated. After 24 and 72 hours, cells were collected and analyzed
for viability by a CellTiter 96 Non-Radioactive Cell Proliferation
Assay [(3-(4,5-Dimethyl-2-thiazolyl)-2,5 dephenyltetrazolium, Pro-
mega,] (MTT) assay, as specified by the manufacturer, and as
previously described .
Found at: doi:10.1371/journal.pone.0012168.s001 (1.10 MB
Conceived and designed the experiments: DYWL AS MC. Performed the
experiments: EEP JX. Analyzed the data: EEP JX AS MC. Wrote the
paper: AS MC.
1. Lee DY, Lin X, Paskaleva EE, Liu Y, Puttamadappa SS, et al. (2009) Palmitic
Acid Is a Novel CD4 Fusion Inhibitor That Blocks HIV Entry and Infection.
AIDS Res Hum Retroviruses 12: 1231–1241.
2. Paskaleva EE, Lin X, Duus K, McSharry JJ, Veille JC, et al. (2008) Sargassum
fusiforme fraction is a potent and specific inhibitor of HIV-1 fusion and reverse
transcriptase. Virol J 5: 8.
3. Paskaleva EE, Lin X, Li W, Cotter R, Klein MT, et al. (2006) Inhibition of
highly productive HIV-1 infection in T cells, primary human macrophages,
microglia, and astrocytes by Sargassum fusiforme. AIDS Res Ther 3: 15.
4. Lipinski CA, Lombardo F, Dominy BW, Feeney PJ (2001) Experimental and
computational approaches to estimate solubility and permeability in drug
discovery and development settings. Adv Drug Deliv Rev 46: 3–26.
5. Goodsell DS, Morris GM, Olson AJ (1996) Automated docking of flexible
ligands: applications of AutoDock. J Mol Recognit 9: 1–5.
6. Morris GM, Huey R, Olson AJ (2008) Using AutoDock for ligand-receptor
docking. Curr Protoc Bioinformatics Chapter 8: Unit 8 14.
7. Resh MD (1999) Fatty acylation of proteins: new insights into membrane
targeting of myristoylated and palmitoylated proteins. Biochim Biophys Acta
8. Resh MD (2006) Palmitoylation of ligands, receptors, and intracellular signaling
molecules. Sci STKE 2006: re14.
9. Aguilar-Cordova E, Chinen J, Donehower L, Lewis DE, Belmont JW (1994) A
sensative reporter cell line for HIV-1 tat activity, HIV-1 Inhibitors, and T cell
activation effects. Aids research and human retroviruses 10: 295–301.
10. Hetenyi C, van der Spoel D (2006) Blind docking of drug-sized compounds to
proteins with up to a thousand residues. FEBS Lett 580: 1447–1450.
11. Goodsell DS (2009) Computational docking of biomolecular complexes with
AutoDock. CSH Protoc 2009: pdb prot5200.
12. Eftink MR (1997) Fluorescence methods for studying equilibrium macromole-
cule-ligand interactions. Methods Enzymol 278: 221–257.
13. Bhal SK, Kassam K, Peirson IG, Pearl GM (2007) The Rule of Five revisited:
applying log D in place of log P in drug-likeness filters. Mol Pharm 4: 556–560.
14. Dong S, Xu G, Hoffmann H (2008) Aggregation behavior of fluorocarbon and
hydrocarbon cationic surfactant mixtures: a study of 1H NMR and 19F NMR.
J Phys Chem B 112: 9371–9378.
15. Cheng Y, Prusoff WH (1973) Relationship between the inhibition constant (K1)
and the concentration of inhibitor which causes 50 per cent inhibition (I50) of an
enzymatic reaction. Biochem Pharmacol 22: 3099–3108.
PA Nanomolar Analogs
PLoS ONE | www.plosone.org6 August 2010 | Volume 5 | Issue 8 | e12168