Báo cáo khoa học: Peptides from purified soybean b-conglycinin inhibit fatty acid synthase by interaction with the thioesterase catalytic domain pot

Schuler2and Elvira Gonzalez de Mejia1 1 Department of Food Science and Human Nutrition, University of Illinois at Urbana-Champaign, IL, USA 2 Department of Cell and Developmental Biology

acid synthase by interaction with the thioesterase

catalytic domain

Cristina Martinez-Villaluenga1, Sanjeewa G Rupasinghe2, Mary A Schuler2and Elvira Gonzalez de Mejia1

1 Department of Food Science and Human Nutrition, University of Illinois at Urbana-Champaign, IL, USA

2 Department of Cell and Developmental Biology, University of Illinois at Urbana-Champaign, IL, USA

Keywords

b-conglycinin-derived peptides; fatty acid

synthase; inhibitors; soybean; thioesterase

Correspondence

E Gonzalez de Mejia, 1201 West Gregory

Drive, 228 ERML, MC-051, Urbana,

IL 61801, USA

Fax: +1 217 265 0925

Tel: +1 217 244 3196

E-mail: edemejia@illinois.edu

(Received 19 August 2009, revised 7

January 2010, accepted 8 January

2010)

doi:10.1111/j.1742-4658.2010.07577.x

Fatty acid synthase (FAS) is uniquely expressed at high levels in cancer cells and adipose tissue The objectives of this study were to identify, purify and validate soy FAS inhibitory peptides and to predict their binding modes Soy peptides were isolated from hydrolysates of purified b-conglyci-nin by co-immunoprecipitation and identified using LC-MS⁄ MS Three peptides, KNPQLR, EITPEKNPQLR and RKQEEDEDEEQQRE, inhib-ited FAS The biological activity of these peptides was confirmed by their inhibitory activity against purified chicken FAS (IC50= 79, 27 and 16 lm, respectively) and a high correlation (r =)0.7) with lipid accumulation in 3T3-L1 adipocytes The FAS inhibitory potency of soy peptides also corre-lated with their molecular mass, pI value and the number of negatively charged and hydrophilic residues Molecular modeling predicted that the large FAS inhibitory peptides (EITPEKNPQLR and RKQEEDE-DEEQQRE) bond to the thioesterase domain of human FAS with lower interaction energies ()442 and )353 kcalÆmol)1, respectively) than classical thioesterase inhibitors (Orlistat, )91 kcalÆmol)1 and C75, )51 kcalÆmol)1) Docking studies suggested that soy peptides blocked the active site through interactions within the catalytic triad, the interface cavity and the hydro-phobic groove in the human FAS thioesterase domain FAS thioesterase inhibitory activities displayed by the synthetic soy peptides EITPEKNPQLR and RKQEEDEDEEQQRE (IC50= 10.1 ± 1.6 and 10.7 ± 4.4 lm, respectively) were higher than C75 (58.7 lm) but lower than Orlistat (0.9 lm) This is the first study to identify FAS inhibitory peptides from purified b-conglycinin hydrolysates and predict their binding modes at the molecular level, leading to their possible use as nutraceuticals

Structured digital abstract

l MINT-7544766 , MINT-7546418 , MINT-7546830 : Beta-conglycinin (uniprotkb: P25974 ) binds ( MI:0407 ) to Alpha subunit of BC (uniprotkb: P13916 ) by pull down ( MI:0096 )

l MINT-7547140 , MINT-7547249 : Beta-conglycinin (uniprotkb: P25974 ) binds ( MI:0407 ) to Beta subunit of BC (uniprotkb: P25974 ) by pull down ( MI:0096 )

Abbreviations

ACP, acyl carrier protein; CIP, co-immunoprecipitation; DMEM, Dulbecco’s modified Eagle’s medium; ER, b-enoyl reductase; FAS, fatty acid synthase; PDB, Protein Data Bank; SBC, soybean b-conglycinin; TBST, Tris-buffered saline containing 0.1% Tween 20; TE, thioesterase.

Fatty acid synthase (FAS, EC 3.2.1.85) is a

multicom-ponent enzyme that catalyzes the de novo biosynthesis

of long-chain fatty acids from acetyl-CoA and

malo-nyl-CoA through a NADPH-dependent cyclic reaction

[1] FAS is homodimeric and each polypeptide chain

(270 kDa) carries seven catalytic domains integrating

all the steps needed for fatty acid synthesis [2,3] The

growing fatty acid is covalently attached to an acyl

carrier protein (ACP), which transports it through the

active sites where each reaction is catalyzed Once the

fatty acid reaches 16–18 carbon atoms in length, it is

released by the thioestherase (TE) domain [1]

Human FAS is downregulated in most normal

human tissues but is highly expressed in adipose and

malignant tissues [4,5] Because of FAS overexpression

in certain chronic diseases, it has become an important

molecular target for chemoprevention and therapeutic

intervention [6,7] The discovery and development of

pharmacologic FAS inhibitors promise the prevention

of obesity, related metabolic disorders and cancer [5,8]

Inhibition of FAS in the central nervous system

mark-edly reduces food intake and body weight in animal

models [9] In particular, inhibition of FAS in the

hypothalamus and pancreatic b cells protects mice

against high fat diet-induced metabolic syndrome [10]

Pharmacological inhibition of FAS results in a 7–10%

longer survival time in mice with gastrointestinal

cancer [11]

Cerulenin and C75 represent two synthetic

com-pounds reported to inhibit FAS [12] but their use has

been limited by several drawbacks, including their

irreversible behavior, low specificity, high chemical

reactivity, interference with other cellular processes

and controversial toxic effects [13–15] Orlistat

(tetra-hydrolipstatin) is an anti-obesity drug intended to

inhibit gastric and pancreatic lipases [16] but it has

been found to inhibit FAS by interacting with its TE

domain; however, it has shown poor systemic stability

and bioavailability [17] The selective FAS inhibitor

GSK837149A discovered by Vazquez et al [18] was

shown to have very low cell permeability in cell

cul-ture Additional plant-derived compounds have also been discovered as potential FAS inhibitors [19–22] Our previous in vitro studies have shown that soybean b-conglycinin (SBC) contains active peptides that inhibit FAS [23] and fatty acid biosynthesis in adipo-cytes [24] The objectives of this study were to iden-tify SBC-derived peptides with FAS inhibitory activity using co-immunoprecipitation (CIP) and pro-teomic techniques The FAS inhibitory activity of the identified peptides was established in both biochemi-cal assays and cell-based models of 3T3-L1 adipo-cytes The relationship between the chemical characteristics of these peptides and their FAS inhibi-tory potency was defined, and their binding modes were predicted by docking simulations using the crys-tal structures of mammalian FAS (PBD ID code: 2VZ8) [25] and human FAS (PBD ID code: 2PX6) [17] This study provides valuable information for the rational design of new FAS inhibitors and the prepa-ration of natural compounds potentially preventing the development of cancer, obesity and related meta-bolic disorders

Results Identification of FAS inhibitory peptides from purified SBC hydrolysate

Previous work in our laboratory demonstrated that hydrolysates from b-conglycinin with alcalase (Bacil-lus licheniformis) exert a potent inhibitory effect (IC50= 30 lm) on FAS activity [23] In this study, amino acid sequences, identified by LC-MS⁄ MS, of FAS inhibitory peptides co-immunoprecipitated from purified SBC hydrolysate were found to be KNPQLR (758 Da), EITPEKNPQLR (1324 Da) and RKQEEDE-DEEQQRE (1847 Da) (Table 1) blast results indicated that the peptides KNPQLR and EITPEKNPQLR matched sequences present in both the a and b subunits

of SBC, whereas RKQEEDEDEEQQRE matched sequences in the a subunit

Table 1 Identification of co-immunoprecipitated soybean b-conglycinin-derived peptides by LC-MS ⁄ MS The peptide sequences listed were found with a confidence of at least 95% (P < 0.05).

Experimental mass (Da) Theoretical mass (Da) Putative sequence Protein fragment Protein source Accession no.

Confirmation of the FAS inhibitory activity of the

identified peptides

To confirm that the identified peptides (KNPQLR,

EITPEKNPQLR and RKQEEDEDEEQQRE)

inhib-ited FAS activity, their amino acid sequences were

custom synthesized and their FAS inhibitory activities tested Figure 1 shows the FAS inhibitory activity of the three synthetic peptides and the C75 positive con-trol The FAS inhibitory responses for each of these peptides were dose dependent, reaching 38.2, 76.5 and 79.4% inhibition at 50 lm for KNPQLR, EITPEKNPQLR and RKQEEDEDEEQQRE, respec-tively compared with 77.8% inhibition at 150 lm for C75 Interestingly, although the three synthetic pep-tides showed FAS inhibitory activity, their potency was different RKQEEDEDEEQQRE exerted a strong inhibitory activity ( 40% inhibition) at doses as low

as 12 lm To compare the potency of the FAS inhibi-tory response across all of the peptides and the C75 positive control, classical sigmoidal dose–response curves were plotted and used to calculate the IC50 val-ues listed in Table 2 In these, the larger peptides, RKQEEDEDEEQQRE (IC50= 16.5 lm) and EI-TPEKNPQLR (IC50= 27.4 lm), showed significantly higher (P < 0.05) potency (lower IC50 value) than the C75 positive control (IC50= 80.3 lm) The smaller KNPQLR peptide had a higher IC50 value (79.9 lm) (P < 0.05) more comparable with the positive control C75 (P > 0.05)

Structure–potency relationship of FAS inhibitory peptides

To further understand the structure–potency relation-ship of the FAS inhibitory peptides, we examined the relationships between the physicochemical and biochemical features of these peptides and their respec-tive inhibitory potency (Table 2) Posirespec-tive correlations were observed between their potency (lower IC50value,

a a

a abc

bcd

cde def

def

defg

efg fg

f

12 25 30 50 60 150

[Compound] (µ M )

120

100

80

60

40

20

0

KNPQLR

EITPEKNPQLR

RKQEEDEDEEQQRE

C75

Fig 1 Fatty acid synthase (FAS) inhibitory activity of synthetic

pep-tides and C75 Synthetic peppep-tides KNPQLR, EITPEKNPQLR and

RKQEEDEDEEQQRE inhibited FAS in a dose-dependent manner

which was similar to the positive control C75 Evaluation of FAS

activity was performed after 20-min preincubation with different

concentrations of each compound Values were expressed as

per-cent inhibition of FAS activity compared with a negative control that

included no inhibitors Each dataset corresponds to the mean of

three independent replicates with error bars indicating the standard

deviations Different letters indicate significant differences at

P < 0.05 in one-way ANOVA analysis.

Table 2 Fatty acid synthase (FAS) inhibitory potency and physicochemical and biochemical characteristics of synthetic peptides MM, pep-tide molecular mass; pI, theoretical isoelectric point of each peppep-tide; GRAVY, grand average of hydropathicity index Parameters were obtained using the Protparam tool in the ExPASY Proteomic Server.

Physicochemical properties

Biological activity

No of charged residues

Hydrophilic amino acids (%) GRAVY

Aliphatic index

IC50value of FAS inhibitory

a Data represent the mean ± SD of three independent experiments; b,c Different letters (b,c) in the column indicate statistical difference (P < 0.05, in one-way ANOVA analysis).dStatistical correlations were carried out between the indicated parameter with the FAS inhibitory potency (the lower the IC50value the higher the potency) of each peptide.

higher potency), their molecular masses (r = +0.89)

and the number of negatively charged (r = +0.64)

and hydrophilic (r = +0.69) residues By contrast, a

strong negative correlation was observed between their

potency and their pI values (r =)0.99) Together,

these results suggest that peptides with higher

inhibi-tory potency are larger and have more negatively

charged and hydrophilic residues No correlations were

found between their IC50 values and other

physico-chemical and biophysico-chemical parameters, such as the

number of positively charged residues, grand average

of hydropathicity index and aliphatic index

Identification of the potential binding site of FAS

and inhibitory peptides from SBC

Peptides EITPEKNPQLR and

RKQEEDEDEEQ-QRE were selected for use in the ligand–enzyme

docking simulations because they displayed higher

FAS inhibitory potency To identify potential binding

sites for these peptide inhibitors, the multidomain

porcine FAS crystal structure (PBD ID code: 2VZ8)

[25] was searched for cavities near the identified active

site residues in each domain Because this structure

lacked the ACP and TE domains, the human ACP

structure (PBD ID code: 2CG5) and human TE

domain structure (PBD ID code: 1XKT) lacking three

short loop regions were also included in this search

Modeling of these regions using the moe (Chemical

Computing Group, Montreal, Canada) program

allowed for cavity identification in the complete

assembled structure, excluding openings extending

into the structurally undefined interdomain regions

Of the seven crystallographically defined domains, the

TE domain had the largest cavity closest to the active site (3959 A˚3) and the b-enoyl reductase (ER) domain had the second largest cavity (3697 A˚3) (Fig 2) Docking of the EITPEKNPQLR inhibitory peptide in both sites predicted interaction energies in the ER domain higher than those in the TE domain (Table 3) Docking of the RKQEEDEDEEQQRE inhibitory peptide predicted near equivalent interac-tion energies in the ER and TE domains In both sites, the predicted energies of the protein–ligand complexes were much higher for the larger RKQEE-DEDEEQQRE peptide than for the smaller EI-TPEKNPQLR peptide These results suggest that the

ER domain is not flexible enough to accommodate relatively large inhibitors and that the TE domain is

a better target for these types of peptide inhibitors

To confirm the binding of these peptide inhibitors in the TE domain, a biochemical enzyme inhibition assay was performed using a recombinant human FAS TE The TE inhibitory activity displayed by peptides

compared with C75 and Orlistat (Table 3) Soy pep-tides were more potent (10 lm) than C75 (58.7 lm), however, their potency was  10-fold lower (P < 0.05) than Orlistat (0.9 lm) Similar to C75, soy peptides blocked > 50% of the TE activity; however, this inhibition was lower (P < 0.05) than Orlistat (77.3%) at 100 lm

Binding and interaction modes of FAS inhibitory peptides

To evaluate in more detail the binding modes of these peptide inhibitors in the TE domain, they were

KS

DH

KR ER

ACP

TE

MAT

Fig 2 Identification of active-site cavities in fatty acid synthase (FAS) In this representa-tion, the multidomain FAS is compiled in MOE from the swine FAS crystal structure (PDB ID code: 2VZ8) [25], the human ACP structure (PBD ID code: 2CG5) [41] and the human thioesterase (TE) domain structure (PBD ID code: 1XKT) [1] The protein back-bone is represented as an orange line, active-site cavities are shown as blue spheres and catalytic residues are shown in space-filling format MAT, malonyl-CoA transacylase domain; KS, b-ketoacyl synthase domain; KR, b-ketoacyl reductase domain; DH, dehydratase domain; ER, b-enoyl reductase domain; ACP, acyl-carrier protein domain; TE, thioestherase domain.

docked individually within the predicted binding site

using the DOCK function within moe and compared

with the docking modes predicted for C75 and

Orli-stat [14,26] The predicted lowest energy conforma-tions of these peptides, C75 and Orlistat inhibitors, in the human FAS TE domain model are shown in

His 2481

Ser 2308 Asp 2338

His 2481

Ser 2308 Asp 2338

His 2481

Ser 2308 Asp 2338

His 2481

Ser 2308 Asp 2338

His 2481

Ser 2308 Asp 2338

His 2481

Ser 2308 Asp 2338

His 2481

Ser 2308 Asp 2338 His 2481

Ser 2308

Subdomain B

N

C

Loop II

Loop III

Loop I

A

C

B

D

Fig 3 Predicted overall fold of the

thioes-therase (TE) domain with inhibitors bound.

The lowest-energy binding mode for

EITPEKNPQLR (A), RKQEEDEDEEQQRE

(B), C75 (C) and Orlistat (D) rendered in

ball-and-stick format in the human TE domain

model (backbone in tube format) is shown

with catalytic triad residues Ser2308,

His2481 and Asp 2338 in space-filling

for-mat Details regarding docking simulations

are summarized in Table 3.

Table 3 Molecular docking within the fatty acid synthase (FAS) potential binding site and human FAS thioesterase (TE) inhibitory potency (IC50) of soybean b-conglycinin-derived peptides, C75 and Orlistat.

b-Enoyl reductase

Thioesterase

Distance to the catalytic triad (A ˚ )

Human FAS TE inhibitory potency (IC 50 , lM)b 10.05 ± 1.60c 10.71 ± 4.36c 58.71 ± 6.74e 0.93 ± 0.13d

a

C75 and Orlistat were docked only in the active site of TE domain because they are known to target the TE domain [Cheng et al [31]].

b Compounds were tested at a concentration of 100 lM Values indicate the mean ± SD of at least two independent experiments c,d,e Dif-ferent letters (c, d, e) in the same row indicate significant difference at P < 0.05 in one-way ANOVA analysis.

Fig 3A–D; interaction energies and the distances

between inhibitor atoms and the catalytic triad are

presented in Table 3 In this docking mode, the

EI-TPEKNPQLR peptide is predicted to be positioned

at a distance of 4.93, 2.39 and 2.98 A˚ from Ser2308,

His2481 and Asp2338, respectively, in the TE domain

(Fig 3A), with a low interaction energy ()353.0

kcalÆmol)1) (Table 3) By comparison, the larger

RKQEEDEDEEQQRE peptide docked in the TE

domain (Fig 3B) at distances of 2.40, 2.16 and

3.98 A˚ from the Ser2308, His2481 and Asp2338,

respectively, with the lowest interaction energy

()442.3 kcalÆmol)1), suggesting that this peptide is a

better inhibitor than EITPEKNPQLR; this is in

agreement with IC50 values listed in Table 2 In their

docking modes, C75 and Orlistat were positioned at

greater distances from both the Asp2338 and Ser2308

residue in the TE domain than was the

RKQEEDE-DEEQQRE peptide and were predicted to interact

more weakly ()51.2 and )90.4 kcalÆmol)1,

respec-tively) with the TE domain than was the smaller

SBC-derived peptide (Table 3) In addition,

correla-tion analyses between the inhibitory potency of

pep-tides EITPEKNPQLR, RKQEEDEDEEQQRE and

C75 and their interaction energies with the TE

domain showed a strong correlation (r = 0.99)

Close-up views of the binding modes of SBC-derived

peptides and Orlistat with the TE active site (Fig 4)

suggest that the palmitic core of Orlistat is bound

almost exclusively to a hydrophobic groove generated

by subdomain B, and its peptidyl moiety is bound in

the interface cavity, whereas the hexanoil tail digs into

the short chain pocket where the catalytic triad exists

These views also suggest that the larger

EI-TPEKNPQLR and RKQEEDEDEEQQRE peptides bind throughout the long hydrophobic groove of the

TE domain in a orientation similar to that of Orlistat with their amino acid side chains also extending into the interface cavity and short chain pocket The poten-tial interaction modes of these peptides with the TE domain suggest that EITPEKNPQLR (Fig 5A) and RKQEEDEDEEQQRE (Fig 5B) bind mainly via hydrophilic interactions (hydrogen-bonding and elec-trostatic interactions) with active site residues By con-trast, only the hydrophilic peptidyl group of Orlistat participates in hydrogen bonding with catalytic triad residues Tyr2307, His2481 and Arg2482 located in the interface cavity

FAS inhibitory activity of synthetic peptides in 3T3-L1 adipocytes

In a cell-based model, FAS inhibition was measured

by monitoring the inhibition of lipid accumulation in 3T3-L1 adipocytes compared with the C75 positive control compound As shown in Fig 6, synthetic pep-tides displayed dose-dependent inhibition of lipid drop-let accumulation in adipocytes The highest inhibition percentages for KNPQLR, EITPEKNPQLR and RKQEEDEDEEQQRE were observed after cell treat-ment at 100 lm (30.1, 29.6 and 34.2%, respectively)

In these assays, KNPQLR and EITPEKNPQLR pep-tides showed similar (P > 0.05) inhibitory potency; however, significantly lower (P < 0.05) than C75 at

50 lm (38.8%) and 100 lm (46.3%) By contrast, the RKQEEDEDEEQQRE peptide showed an inhibitory activity similar to C75 at all concentrations tested with the exception of 50 lm, which displayed only 27.3%

Fig 4 Molecular surface representation of the thioestherase (TE) domain with inhibi-tory peptides EITPEKNPQLR (orange ball-and-stick format), RKQEEDEDEEQQRE (blue ball-and-stick format) and Orlistat (red ball-and-stick format) The potential surface

is colored to reflect hydrogen bonding (pink), strong hydrophilic (green) and mild hydrophilic (blue) regions.

inhibition (P > 0.05) Inhibition of lipid accumulation

by these peptides correlated with FAS inhibition

(r = 0.70) (Fig 6) even though the magnitudes of

inhi-bition in these peptides in the cell-based model were lower than the magnitude of FAS inhibition measured

in biochemical assays; this is probably because of the

A

B

C

Fig 5 Detailed 2D interactions between

inhibitors and the thioestherase (TE) domain.

Calculated using the MOE program following

the method of Clark & Labute [44], residues

in the TE domain that contribute to the

bind-ing of EITPEKNPQLR (A),

RKQEEDE-DEEQQRE (B) and Orlistat (C) are shown

with green circles indicating residues with

no polar or charged side chains and light

mauve circles indicating polar side chains

that are either acidic (red ring) or basic (blue

ring) Arrows indicate hydrogen bonds to

side chain (green) and backbone (blue)

residues.

low permeability of cell to these longer peptides No

effect on cell viability of 3T3-L1 adipocytes was

observed with any of the treatments used in this study,

indicating no cellular toxicity (data not shown)

Discussion

FAS is an important target for prevention and

thera-peutic interventions because multiple lines of evidence

have shown high levels of FAS expression in cancer,

obesity and metabolic disorders [27] The discovery

and development of agents that block FAS activity

highlight the potential for the prevention and

treat-ment of those chronic diseases Our previous work

demonstrated that SBC contains FAS inhibitory

pep-tides that may be released by enzymatic hydrolysis

with alcalase [23] This study has identified the FAS

inhibitory peptides in the SBC hydrolysate using CIP,

taking advantage of the specific affinity between FAS

and its inhibitory peptides This CIP approach has

identified for the first time three peptide fragments

from the a and b subunits of SBC (KNPQLR,

potential inhibitors of FAS activity and their activities

were confirmed using their custom synthesized

pep-tides Our results have indicated that the inhibitory

potency of these peptides (16.5–79.9 lm) is within the

range found for purified SBC hydrolysates (IC50= 30

lm) and soybean hydrolysates (50.4–175.1 lm) [23] In

comparison with other natural inhibitors, the inhibi-tory potency of these peptides is within the range found for flavonoids from green tea (2.3–111.7 lm) [28,29] and tannins from Geum japonicum var chinense (0.2–41.4 lm) [22], which has been evaluated in pre-clinical studies [22,30]

Molecular modeling has identified the TE domain as the potential binding site for the FAS inhibitory pep-tides from SBC Molecular docking has shown that soy peptides displayed a different inhibitory mecha-nism than C75 Soy peptides are selective inhibitors of the FAS TE domain, whereas C75 has been shown to interact at several sites in FAS [14] The predicted binding energy of C75 in the FAS b-ketoacyl synthase domain was )53.9 kcalÆmol)1, similar to that observed

in the TE domain ()51.2 kcalÆmol)1) These results indicate that C75 is not a selective inhibitor for a particular FAS domain, in agreement with previous findings [14] We also confirmed that the synthetic pep-tides EITPEKNPQLR and RKQEEDEDEEQQRE inhibited 4-methylumbelliferone heptanoate hydrolysis

by TE in in vitro experiments Therefore, these peptides are antagonists of TE under near physiologic conditions, meaning that they bind to the unoccupied enzyme and reduce substrate turnover The TE domain coordinates the terminal step of fatty acid synthesis by hydrolyzing palmitate from the 4¢-phosphopanteine arm of the ACP domain [1] Its active site is comprised

of a hydrophobic groove with a distal pocket at the interface of subdomains A and B and a hydrophilic catalytic triad (Ser2308, His2481 and Asp2338) at the proximal end of the groove [17] Palmitate, the main biological product of FAS, binds in the hydrophobic groove, its hydrophilic carboxyl group interacting with the catalytic triad, and its hydrophobic, hydrocarbon chain extending away from the triad [31] From the binding modes that we have predicted, inhibitory peptides appear to block the catalytic activity of TE through hydrophilic interactions with enzyme residues located in the catalytic triad, the hydrophobic groove and the interface cavity The biochemical parameters

of these peptides suggest that the numbers of nega-tively charged and hydrophylic residues are important predictors of their potency, in agreement with the fact that hydrophilic interactions are important to block the catalytic activity of the TE domain [31] The high number of charged and hydrophilic groups in these inhibitors provides for strong hydrogen bonding and electrostatic interactions with the catalytic residues of the TE domain

Analysis of the TE domain in a variety of species indicates that catalytic triad residues are completely conserved from insects to mammals and all other

[compound] (µM ) 0

10

20

30

40

50

60

KNPQLR

EITPEKNPQLR

RKQEEDEDEEQQRE

C75

i

ghi

hi

fgh

ghi

fghi fgh

def

fg ef

cde

ab

bc bcd abc

a

Fig 6 Inhibition of lipid accumulation in 3T3-L1 adipocytes by

thetic peptides 3T3-L1 adipocyte cells were treated with the

syn-thetic KNPQLR, EITPEKNPQLR and RKQEEDEDEEQQRE peptides

at concentrations ranging from 0 to 100 lM on days 3, 5 and 7, and

lipid accumulation was measured on day 10 using the Oil Red O

assay as outlined in Experimental procedures Each dataset

corre-sponds to the average of three independent replicates with error

bars indicating the standard deviation Different letters indicate

sig-nificant differences at P < 0.05 in ANOVA analysis.

residues are conserved from birds to mammals, with

the exception at Phe2370 which is changed to Ala2370

in chickens [1] This suggests that our current

predic-tions on the binding mode of SBC-derived peptides in

the human FAS TE domain can validly explain

inhibi-tion of catalytic activity in the chicken FAS

Some evidence has shown that dietary soy protein

may promote satiety and weight loss [32,33] and

protect against certain types of cancer [34] The

obesity-preventive effects of soybean protein have

been associated with its ability to decrease lipid

syn-thesis, adipogenesis and thermogenesis by regulating

gene expression [32] Dietary intake of soy protein

has also been reported to reduce tumor incidence in a

rat model of chemically induced colon cancer by

attenuating FAS expression [34] Our results provide

additional insight into the preventive mechanisms of

soy components in showing that SBC peptide

frag-ments inhibit FAS activity in adipose cells in the

same way as C75 Schmid et al [35] reported that

FAS inhibition by C75 prevented adipogenesis in a

cell-based model FAS inhibitory activity of SBC

pep-tide fragments may potentially be found in cancer

cells, liver or hypothalamus, as shown previously for

C75 Inactivation of hypothalamic FAS by C75 is

linked to satiety and dramatic weight loss [15]

because accumulation of the substrate malonyl CoA

through the inhibition of FAS appears to inhibit the

expression of neuropeptide Y which promotes

inges-tion [9] Moreover, the FAS inhibitory activity of

C75 induced apoptosis and prevented the growth of

multiple tumor xenografts in vivo [36,37] Our findings

clarify the mechanism linking FAS inhibition with the

anti-obesity effects of soy protein-derived peptides In

conclusion, the soy peptides EITPEKNPQLR and

RKQEEDEDEEQQRE inhibited the TE domain and

de novo fatty acid synthesis in adipocytes The

bind-ing mode of these peptides in the large

palmitate-binding pocket is of particular interest and will guide

future research These FAS inhibitory peptides can

serve as lead compounds to design peptoid analogs

(oligomers of N-subtituted glycine) with equivalent

biological activity, enhanced systemic stability and

bioavailability than standard peptides [38] The

rele-vance of the identification of these SBC-derived

pep-tides is noticeable because of the novelty of their

biological activity and chemical nature Molecular

docking has allowed us to predict binding modes for

SBC-derived peptides (EITPEKNPQLR and

RKQEE-DEDEEQQRE) in the TE domain Based on our

data, it is likely that the consumption of soy high in

b-conglycinin represents a preventive alternative to

improve health and wellness

Experimental procedures Materials

b-Conglycinin was purified from soybean defatted flour as described in Wang et al [39] FAS inhibitory peptides were produced from SBC hydrolysis with alcalase from Bacil-lus licheniformis, as detailed in Martinez-Villaluenga et al [24] The identified FAS inhibitory peptides (> 95% purity) were custom synthesized by GenScript (Piscataway, NJ, USA) FAS was isolated from chicken liver and purified (70% purity) as described by Tian et al [40] Human recombinant FAS TE (residues 2010–2509) was kindly pro-vided by J.W Smith (Burnham Institute for Medical Research, CA, USA) 3T3-L1 (also designated ATCC CCL-92.1) preadipocytes from Swiss albino mouse and Dulbecco’s modified Eagle’s medium (DMEM) were pur-chased from the American Type Culture Collection (Rock-ville, MD, USA) Calf bovine serum, fetal bovine serum and Dulbecco’s phosphate buffer saline were from Invitro-gen (Rockville, MD, USA) Alcalase from B licheniformis (EC 3.4.21.62) and C75 were purchased from

non-specific goat IgG and goat polyclonal IgG against a peptide mapping at the C-terminus of FAS were from Santa Cruz Biotechnology (Santa Cruz, CA, USA) Unless otherwise stated, all chemical reagents were from Sigma-Aldrich

FAS activity assay

FAS activity was assayed by a spectrophotometric method using a Synergy 2 Microplate Reader System equipped with temperature controller (Biotek Instruments, Winooksi, VA,

mea-suring the decrease in absorbance at 340 nm in a 96-well clear-bottomed polysterene plate (Corning, NY, USA) Reactions were performed in a final volume of 150 lL con-taining 3 lm acetyl-CoA, 10 lm malonyl-CoA and 35 lm NADPH and 0.3 lm FAS in 0.1 m potassium phosphate buffer Initial rates were calculated for the slope of the pro-gress curves during the first 5 min

FAS inhibition studies

Synthetic peptides and the C75 positive control compound were used for FAS inhibition studies with stock solutions

of the synthetic peptides and C75 dissolved in deionized water and dimetylsulfoxide, respectively, and serial dilutions made in 0.1 m potassium phosphate buffer (pH 7.0) Inhibi-tion studies were performed by measuring the residual FAS activity after enzyme preincubation with inhibitors for

curves in which the range of concentrations was distributed

using nonlinear regression sigmoidal curve fit functions in

GraphPad prism 4.00 (Graphpad Software Inc., San Diego,

CA, USA)

Inhibition of FAS TE enzymatic activity was performed

using a fluorescence method described by Richardson &

Smith [41] Peptides were added to yield a final

concentra-tion of 100 lm; in this assay the ability of the recombinant

TE to cleave 4-methylumbelliferone heptanoate and

hydro-lyzed it to the fluorescent 4-methylumbelliferone was

Co-immunoprecipitation

To purify FAS inhibitory peptides a CIP approach was

performed Briefly, 200 lL of b-conglycinin hydrolysate

pH 7.0) were added with 2 lL of goat IgG and 25 lL of

pre-clearing, samples were centrifuged at 1000 g for 5 min at

2 lm FAS in 0.1 m potassium phosphate buffer (pH 7.0)

The negative control consisted of 80 lL of 0.1 m

potas-sium phosphate buffer (pH 7.0) added to 120 lL FAS

The blank consisted of 200 lL of 0.1 m potassium

phos-phate buffer (pH 7.0) These samples were incubated for

10 lL goat polyclonal antibody (FAS IgG) for 60 min at

incubation with the antibody, samples were centrifuged at

with radioimmunoprecipitation buffer The sediment was

resuspended in HPLC-grade water and boiled for 3 min to

release proteins from the beads Then, 20 lL acetonitrile

containing 0.8 lL formic acid were added to extract the

peptides and proteins, the beads were removed by

inhibitory peptides

Western blot analysis

To confirm the CIP of FAS, western blot analysis was

carried out using goat polyclonal antibody (FAS IgG)

Pro-teins released from the beads were resuspended in Laemmli

loading buffer (BioRad, Hercules, CA, USA) containing

5% 2-mercaptoethanol Samples (20 lg soluble protein)

through a mini-electrophoresis kit at 200 V constant for

40 min Further, proteins were transferred to

poly(vinyli-dene diflouride) membrane (BioRad) in blotting buffer

(25 mm Tris, 192 mm glycine pH 8.3, 0.1% SDS) using

After the transfer, membrane was blocked with 5% non-fat dry milk in Tris-buffered saline containing 0.1% Tween 20 (TBST) for 1 h, followed by an overnight incubation with

membrane was washed with TBST four times and was incu-bated with bovine anti-(goat IgG) horseradish peroxidase conjugates (1 : 1000) for 1 h at room temperature The membrane was washed again in TBST for four times and signals were visualized using chemiluminescence reagent (GE Healthcare, Chalfont St Giles, UK) and a Kodak Image Station 440 CF (Eastman Kodak Co., New Haven,

CT, USA)

LC-MS⁄ MS

Milford, MA, USA) using 0.1% aqueous formic acid as solvent A and acetonitrile with 0.1% formic acid as sol-vent B A linear gradient from 1 to 60% B was run for

60 min and back to 1% B for 10 min with the flow rate

in a Q-Tof API-US nanoAcquity LC (Waters) mass spec-trometer equipped with an electron spray ion source The Q-Tof instrument was operated in positive ion mode

for de novo peptide sequencing and database searching Peptide identification was carried out by searching against the NCBI or SWISS-PROT database [taxonomy = viridi-plantae (green plants)] Only peptides identified with a con-fidence of at least 95% were considered to be correct calls (P < 0.05)

Cell culture and treatments

medium containing 1% sodium pyruvate, 1%

2) After reaching 100% confluence, the cells were

bovine serum, 0.5 mm isobutylmethylxanthine, 1 lm dexa-methasone and 1.7 lm insulin (days 3 and 4) Cells were

1.7 lm insulin for another 2 days (days 5 and 6),

an additional 4 days (days 7–10), at which time > 90%

of cells were mature adipocytes with fat droplets Cells were treated on days 3, 5 and 7 of the differentiation process with synthetic peptides dissolved in Dulbecco’s phosphate buffer saline at a concentration ranging from

atmo-sphere for 48 h