Structure Prediction with FAMS for Proteins Screened Critically

to Autoimmune Diseases based upon Bioinformatics

Shigeharu Ishida

, Hideaki Umeyama

, Mitsuo Iwadate

and Y-h. Taguchi

Department of Physics, Chuo University, Tokyo 112-8551, Japan

Department of Biological Sciences, Chuo University, Tokyo 112-8551, Japan

Keywords:

FAMS, Autoimmune Diseases, Structure Prediction, Drug Discovery.

Abstract:

Drug discovery for autoimmune diseases is recently recognized to be an important task. In this study, we try to

perform structure prediction of proteins whose gene promoter regions were previously reported to be specif-

ically methelysed or demethylased commonly for three autoimmune diseases, systemic lupus erythematosus,

rheumatoid arthritis, and dermatomyositis. FAMS were employed for this purpose and we can predict three

dimensional structure with signiﬁcantly small enough P-values. Most of them are suggested to be self im-

munology related proteins and will be important drug target candidates. We also found some proteins which

form complex with each other. The possibility of a new drug target, i.e., suppression of protein complex

formation is suggested.

1 INTRODUCTION

Autoimmune diseases are recently recognized as seri-

ous symptom. For example, systemic lupus erythe-

matosus (SLE), which is known to be one of sys-

temic autoimmune diseases, most often harms the

heart, joints, skin, lungs, blood vessels, liver, kidneys,

and nervous system. The cause of this disease is un-

known. The lack of the knowledge about basic mech-

anism of the disease prevents us from generating ef-

fective drugs to cure this disease. SLE is the secondly

frequent connectivetissue disease, while the most fre-

quent one is Rheumatoid Arthritis (RA), which is also

known to be one of autoimmune diseases. Although

there are some proposals about the cause of RA, it has

not yet been fully understood. In RA, the arthritis of

joints known as synovitis is inﬂammation of the syn-

ovial membrane that lines joints and tendon sheaths.

Joints become swollen, tender and warm, and stiff-

ness limits their movement. Another example of au-

toimmune disease is dermatomyositis (DM), which is

also a connective-tissue disease related to polymyosi-

tis that is characterized by inﬂammation of the mus-

cles and the skin. Its cause is unknown, too.

In spite of the lack of basic understanding of dis-

eases’ causes, there is a general belief; there should be

a common cause of autoimmune diseases. Following

this line, in accordance with the recent development

of genome science, several conjectures are proposed.

For example, O’Hanlon et al recently showed that

there are common pathways which contribute to

multiple systematic autoimmune diseases (O’Hanlon

et al., 2011b), based upon gene expression analy-

sis. More recently, they have conﬁrmed their ﬁndings

using proteomic analysis (O’Hanlon et al., 2011a).

However, Zhou et al (Zhou et al., 2005) found that

unaffected monozygotic (MZ) twins share ﬁbroblast

gene expression with systemic sclerosis (SSc) pa-

tients (counter parts). SSc is also believed to be re-

lated to autoimmune diseases. On the other hand,

Gervin et al (Gervin et al., 2012) recently found

that combined analysis between gene expression and

methylation enables them to detect slight difference

between affected and unaffected twins. Their ﬁnd-

ings are not contradict to the study by Javierre et

al (Javierre et al., 2010) who could not ﬁnd any

shared methylation patterns among multiple autoim-

mune diseases. Thus, at the moment, it is a little bit

confusing what kind of aspects can be shared with

multiple autoimmune diseases.

A few years ago, we reanalyzed (Taguchi, 2010)

Javierre et al’s data (Javierre et al., 2010) using prin-

cipal component analysis (PCA) and found that some

genes’ methylation are commonly and signiﬁcantly

different from healthy controls. In this paper, we try

to validate our ﬁndings using Full Automatic Model-

ing System (FAMS), which is protein structure pre-

diction program that perform comparative modeling

261

Ishida S., Umeyama H., Iwadate M. and Taguchi Y..

Structure Prediction with FAMS for Proteins Screened Critically to Autoimmune Diseases based upon Bioinformatics.

DOI: 10.5220/0004188802610267

In Proceedings of the International Conference on Bioinformatics Models, Methods and Algorithms (BIOINFORMATICS-2013), pages 261-267

ISBN: 978-989-8565-35-8

 2013 SCITEPRESS (Science and Technology Publications, Lda.)

with database search and simulated annealing. Using

FAMS(Umeyama and Iwadate, 2002), we can predict

functionality of genes by comparing them with the

proteins whose function and structure are known. We

also validate if these genes can form complex and ﬁnd

many candidates to form protein complex. The possi-

bility that they can be a drug target will be discussed.

2 MATERIALS AND METHODS

2.1 Selection of Candidate Genes

Although details are reported previously(Taguchi,

2010), here we brieﬂy describe how we have selected

candidate genes. Javierre et al (Javierre et al., 2010)

measured promoter methylation patterns using mi-

croarray technology (Illumina GoldenGate Methyla-

tion Cancer Panel I) for SLE, RA, and DM. Their

expression patterns are deposited to Gene Expression

Omnibus with the accession number of GSE19033.

We downloaded series

matrix.txt from there, applied

PCA to them and picked up gene whose promot-

ers’ methylation is signiﬁcantly different from healthy

controls.

2.2 Structure Prediction of Selected

Genes

Selected genes’ amino acid sequences are down-

loaded from SWISS Prot. Then their protein struc-

tures are inferred by FAMS.

2.3 Protein Complex Formation

Prediction

We checked if a pair of model proteins used for struc-

ture prediction can form protein complex or not as

follows. First, PDB ﬁles which contains at least

one model protein as a member of protein com-

plex are downloaded. Then, which model proteins

are included into the common PDB ﬁles is investi-

gated. Thirdly, inter-atomic distances between pairs

of model proteins which belong to the same PDB ﬁle

are computed. If there are at least a pair of atoms

whose distances are less than 3.5

A, a pair of model

proteins is listed as a candidate to form protein com-

plex.

Figure 1: Comparison between reference protein

2OQ0(Green) and model protein AIM2(Cyan).

3 RESULTS

3.1 Biological Signiﬁcance Figured out

by FAMS

In Table 1, we have listed genes (i.e., reference pro-

teins) selected by PCA(Taguchi, 2010), together with

the model proteins which are inferred to have sim-

ilar structure to each of them by FAMS. First of

all, FAMS has successfully listed model proteins for

most of reference proteins with very small P-values.

Fig. 1 shows a typical example of model proteins.

It is the model protein 2OQ0

B for the reference

protein AIM2. Alignment regions are 192 amino

acid sequence from total length of 209 amino acid

of 2OQ0

B and 191 amino acid sequence from to-

tal length 343 amino acid sequence of AIM2. Se-

quence similarity between two alignment regions is

44 %. P-value attributed is 4× 10

−92

. Although this

is only one example of typical relationship between

model/reference proteins, generally we could get this

quality of structural similarities. This means struc-

tural similarity between models and references is reli-

able. In addition to this, biological features attributed

to the model proteins are often reasonable. Due to the

limitation of the space, we cannot explain all of them

one by one, we will point out some of these exam-

ples. Then, modeling yields predictions that need to

be experimentally veriﬁed.

TRIP6 is expected to have similar structure to

CRP1, which is inferred as immune response(Latonen

et al., 2010). TM7SF3 is recognized as cytochrome

c oxidase, which was reported to bind to immune

gamma-globulins (Frey et al., 1978). TIE1, PECAM1

BIOINFORMATICS2013-InternationalConferenceonBioinformaticsModels,MethodsandAlgorithms

262

Table 1: Selected genes and model protein used for structure prediction. Bold ID of PDB indicates that reference protein itself

is detected in PDB.

Reference Model

gene symbol

PDB ID P-value gene symbol

AIM2 2OQ0 B 4× 10

−92

GAMMA-INTERFERON-INDUCIBLE PROTEIN IFI-16

CARD15

3CIY B 7× 10

−64

TOLL-LIKE RECEPTOR 4, VARIABLE LYMPHOCYTE (TLR4)

CD82

2BG9 A 0.46 ACETYLCHOLINE RECEPTOR PROTEIN, ALPHA CHAIN

CSF1R

3B43 A 5× 10

−83

TITIN

CSF3

1GNC A 2× 10

−66

GRANULOCYTE COLONY-STIMULATING FACTOR

CSF3R

3DMK A 1× 10

−71

DOWN SYNDROME CELL ADHESION MOLECULE (DSCAM)

DHCR24

2Q4W A 1× 10

−115

CYTOKININ DEHYDROGENASE 7 (CKO7)

ERCC3

2W74 D 1× 10

−152

TYPE I RESTRICTION ENZYME ECOR124II R PROTEIN (HSDR)

GRB7

3HK0 B 2× 10

−73

GROWTH FACTOR RECEPTOR-BOUND PROTEIN 10 (GRB10)

HGF

2F83 A 1× 10

−111

COAGULATION FACTOR XI

HOXB2

2D5V A 9× 10

−24

HEPATOCYTE NUCLEAR FACTOR 6 (HNF-6)

IFNGR2

1FNF A 1× 10

−37

FIBRONECTIN

LCN2

1X71 A 1× 10

−51

NEUTROPHIL GELATINASE-ASSOCIATED LIPOCALIN (NGAL)

LMO2

2XJY A 2× 10

−33

RHOMBOTIN-2

LTB4R

2KS9 A 2×10

−83

SUBSTANCE-P RECEPTOR

MMP14

1SU3 B 1×10

−160

INTERSTITIAL COLLAGENASE (MMP-1)

MMP8

1SU3 B 1×10

−171

INTERSTITIAL COLLAGENASE (MMP-1)

MPL

3L5H A 4× 10

−63

INTERLEUKIN-6 RECEPTOR SUBUNIT BETA (IL6RB)

PAD14

2DEW X 0.0 PROTEIN-ARGININE DEIMINASE TYPE IV

PECAM1

3DMK A 1× 10

−104

DOWN SYNDROME CELL ADHESION MOLECULE (DSCAM)

PI3

1TWP A 2×10

−19

WHEY ACIDIC PROTEIN (WAP)

RARA

3DZY A 4× 10

−95

RETINOIC ACID RECEPTOR RXR-ALPHA

S100A2

2RGI A 4× 10

−19

PROTEIN S100-A2

SEPT9

3FTQ A 1× 10

−137

SEPTIN-2

SLC22A18

1PW4 A 1× 10

−108

GLYCEROL-3-PHOSPHATE TRANSPORTER

SPI1

1GVJ B 1×10

−21

C-ETS-1 PROTEIN (ETS1)

SPP1

1D2T A 3× 10

−14

ACID PHOSPHATASE (ACP)

STAT5A

1Y1U A 0.0 SIGNAL TRANSDUCER AND ACTIVATOR OF TRANSCRIPTION (STAT5A)

SYK

2OZO A 1× 10

−168

TYROSINE-PROTEIN KINASE ZAP-70

TIE1

3DMK A 2× 10

−84

DOWN SYNDROME CELL ADHESION MOLECULE (DSCAM)

TM7SF3

[1AR1 A] 6× 10

−88

CYTOCHROME C OXIDASE

TRIP6

1B8T A 2×10

−32

CYSTEINE-RICH PROTEIN 1 (CRP1)

VAMP8

2KOG A 1× 10

−21

VESICLE-ASSOCIATED MEMBRANE PROTEIN 2 (VAMP2)

Figure 2: Ligand binding to MMP8. Magenta and Cyan

are reference and model proteins. Stick models are ligand

molecules binding to 95% homologous proteins with refer-

ence.

Figure 3: Ligand binding to MMP14. Magenta and Cyan

are reference and model proteins. Stick models are ligand

molecules binding to 95% homologous proteins with refer-

ence.

StructurePredictionwithFAMSforProteinsScreenedCriticallytoAutoimmuneDiseasesbaseduponBioinformatics

263

Table 2: The number of common PDB ID detected homology search results between two genes are listed. The threshold of

both searches set to 1× 10

−10

, that is enough low to conserve the protein tertiary structure. Then the model proteins is likely

to bind to other model proteins.

AIM2 P624 F

CARD15

P302 R

CD82

P557 R

CSF1R E26 F

CSF3

E242 R

CSF3R P472 F

DHCR24

P652 R

ERCC3

P1210 R

GRB7

E71 R

HGF

E102 R

HOXB2

P99 F

IFNGR2

P377 R

LCN2

P86 R

LMO2

E148 F

LTB4R P163 F

MMP14 P13 F

MMP8

E89 R

MPL

P62 F

PADI4 E24 F

PECAM1

E32 R

PI3

P274 R

RARA

P1076 R

S100A2

E36 R

SEPT9 P374 F

SLC22A18

P216 R

SPI1

E205 F

SPP1 P647 F

STAT5A

P704 R

SYK P584 F

TIE1 E66 R

TM7SF3

P1068 R

TRIP6

P1090 F

VAMP8 P241 F

AIM2 P624 F 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

CARD15

P302 R 0 0 2 0 5 0 16 0 3 0 0 0 0 0 0 0 0 0 5 0 0 0 0 0 0 0 8 0 0 0 0 0

CD82

P557 R 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

CSF1R

E26 F 0 2 0 0 186 0 0 10 25 0 5 8 0 0 0 0 5 0 410 0 0 2 0 0 0 0 9 1223 1355 0 0 0

CSF3

E242 R 0 0 0 0 52 0 0 0 0 0 42 0 0 0 0 0 28 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0

CSF3R

P472 F 0 5 0 186 52 0 0 0 107 0 166 48 0 3 0 0 115 0 1177 0 0 0 0 0 0 0 0 0 204 0 0 0

DHCR24

P652 R 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

ERCC3

P1210 R 0 16 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

GRB7

E71 R 0 0 0 10 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 170 211 10 0 0 0

HGF

E102 R 0 2 0 25 0 107 0 0 0 0 140 2 0 0 0 0 0 0 84 2 0 0 0 0 0 0 0 0 0 0 0 0

HOXB2

P99 F 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

IFNGR2

P377 R 0 0 0 5 42 166 0 0 0 140 0 0 0 0 0 0 91 0 24 0 0 0 0 0 0 0 0 0 54 0 0 0

LCN2

P86 R 0 0 0 4 0 24 0 0 0 2 0 0 0 0 0 0 0 0 16 0 0 0 0 0 0 0 0 0 0 0 0 0

LMO2

E148 F 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0

LTB4R

P163 F 0 0 0 0 0 3 0 0 0 0 0 0 0 0 0 0 0 0 6 0 0 0 0 0 0 0 0 0 0 0 0 0

MMP14

P13 F 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 125 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

MMP8

E89 R 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 125 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

MPL

P62 F 0 0 0 5 28 115 0 0 0 0 0 91 0 0 0 0 0 0 19 0 3 0 0 0 0 0 0 0 34 0 0 0

PADI4

E24 F 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

PECAM1

E32 R 0 5 0 410 1 1177 0 0 0 84 0 24 32 0 6 0 0 19 0 0 0 0 0 0 0 0 0 1 129 0 0 0

PI3

P274 R 0 0 0 0 0 0 0 0 0 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

RARA

P1076 R 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0

S100A2

E36 R 0 0 0 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 1 1 0 0 0

SEPT9

P374 F 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0

SLC22A18

P216 R 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

SPI1

E205 F 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

SPP1

P647 F 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

STAT5A

P704 R 0 8 0 9 0 0 0 0 170 0 0 0 0 0 0 0 0 0 0 0 0 0 2 1 0 0 0 164 9 0 0 0

SYK

P584 F 0 0 0 1223 0 0 0 0 211 0 0 0 0 0 0 0 0 0 0 1 0 0 1 0 0 0 0 164 1200 0 0 0

TIE1

E66 R 0 0 0 1355 0 204 0 0 10 0 0 54 0 0 0 0 0 34 0 129 0 0 1 0 0 0 0 9 1200 0 0 0

TM7SF3

P1068 R 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

TRIP6

P1090 F 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

VAMP8

P241 F 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

and CSF3R are recognized as DOWN SYNDROME

CELL ADHESION MOLECULE (DSCAM), which

is known to be immunoglobulin (Ig)-superfamily re-

ceptor in insect(Watson et al., 2005). SYK is rec-

ognized as TYROSINE-PROTEIN KINASE ZAP-70

and both SYK and ZAP-70 are reported to display dis-

tinct requirements for Src family kinases in immune

response receptor signal transduction(Zoller et al.,

1997). STAT5A itself is found in PDB, which is re-

ported to play critical role for cytokine responses and

normal immune function(Lin et al., 2012). SPI1 is

recognized as ETS1, which is known to be expres-

sive in SLE and play some function in immune sys-

tem(Pan et al., 2011). S100-A2 itself is in PDB and is

reported to be antibodies and inhibitors directed to-

ward receptor for advanced glycation end products

(RAGE) ligands(Heijmans et al., 2012). RARA is

structurally similar to RXR-α, which is reported to

be involved in inﬂammatory responses(Selvaraj et al.,

2010). PI3 is as WAP, which is reported to play a

role in innate immune(Bingle and Vyakarnam, 2008).

PADI4 is itself in PDB and is reported to be important

in RA(Abd-Allah et al., 2012). MPL’s structure is in-

ferred to be similar to IL6RB. IL6R is reported to be a

key mediator of RA(Cronstein, 2007). LCN2, which

is also called as NGAL, is in PDB. NGAL is tried to

be used as a marker of inﬂammatory status for allow-

ing an early diagnosis of inﬂammatory disease such

as autoimmune disease in DS patients(Dogliotti et al.,

2010). One of HOXB2’s model proteins is HNF-

6, which is known to cause immunologically distinct

feature(Samadani and Costa, 1996). AIM2 is struc-

BIOINFORMATICS2013-InternationalConferenceonBioinformaticsModels,MethodsandAlgorithms

264

Figure 4: Protein complex formation candidates between

PECAM1 (Cyan) and CSF1R (Magenta) detected in P-

value 1e-53 and 4e-68, based upon protein structure of PDB

2ZJS. Green region are excluded from matching between

model and reference proteins.

turally similar to IFI16. AIM2 and IFI16 are reported

to play critucal role in immunology (Jin et al., 2012).

CARD15 is inferred to be similar to TLR4 which play

a role in cell antiviral response together with TLR3:

TICAM1-speciﬁc signaling pathways(Meylan et al.,

2004). CD82 is known to be ACETYLCHOLINE

RECEPTOR PROTEIN which often play a critical

role in immune system(Quek et al., 2012)

. CSF1R

is assigned to be TITAN, which is known to be in-

volved in to immune response(Skeie et al., 1998).

SPP1 is recognized as ACID PHOSPHATASE, which

is known to be related to be autoimmune prostati-

tis(Fong et al., 1997). LMO2, which is also known to

be RHOMBOTIN-2, is known to be related to ZFAT

(a zinc-ﬁnger gene in autoimmune thyroid disease

susceptibility region / an immune-related transcrip-

tional regulator containing 18 C2H2-type zinc-ﬁnger

domains and one AT-hook)(Tsunoda et al., 2010).

DHCR24 is regarded as CYTOKININ DEHYDRO-

GENASE. Cytokine has, not to mention, been used

to refer to the immunomodulating agents. SEPT9 is

homologous to SEPTIN-2, which is reported to be up-

regulated in cytoskeletal and immune function-related

proteome proﬁles (Gabr et al., 2007). IFNGR2 is re-

garded as FIBRONECTIN, which play a role in im-

mune responses in organ transplant recipients(Coito

et al., 2000). CSF3 itself is in PDB, which is knownto

have relationship with immune system (Sarkar et al.,

2012). GRB7 is also recognized as GRB10, which

play an important role in immune system, although it

is in cancer(O-Sullivan et al., 2008). HGF is related to

COAGULATION FACTOR XI, which is known to be

related to immunology(Bouma et al., 1983). LTB4R

Although P-value attributed to CD82 is not small

enough, reliability of this assignment turns out to be rea-

sonable after some more details consideration (not shown

here).

is recognized as SUBSTANCE-P RECEPTOR, which

is known to have immune response to respiratory syn-

cytial virus infection (Tripp et al., 2002).

These are only a part of immune system related

features which are attributed to each gene by FAMS.

Although more examples can easily be listed, we omit

the rest of them because of length limitation. Anyway,

it is clear that FAMS based feature attribution works

very well for genes selected by PCA(Taguchi, 2010).

3.2 Possibility of Drug Discovery

Although it is interesting enough to ﬁnd that FAMS

can be used for the validation of genes selected by

other bioinformatic method, it will be better if we can

make use of FAMS for the drug discovery.

3.2.1 Ligand Binding to ”Pocket”

The most popular method to ﬁnd drug is to ﬁnd a

small molecule to bind a ”pocket” of each protein. If

FAMS can ﬁnd or suggest such a candidate for each

of genes in Table 1, it will be very useful.

For example, there are two proteins, MMP8 and

MMP14, in Table 1. They are known to coregulate

target genes(Silva et al., 2012). Both of them are

recognized as members of matrix metalloproteinase

(MMP) family, which is inﬂammation related protein

family. For MMP8, using 1XUC

A, which is MMP-

13, as a template, FAMS successfully showed that

there are many ligands likely to bind MMP8 (Fig.

2). Similarly, for MMP14, using 1BQO

B, which is

MMP-3, as a template, FAMS successfully showed

that there are many ligands likely to bind MMP14, too

(Fig. 3). Although it is not a ﬁnding of a new drug,

this shows the potential for proteins listed in Table 1

which can be new drug targets. Further researches

following this line will be waited.

3.2.2 Termination of Protein Complex

Formation

Other and new possibility of drug target is interrup-

tion in protein complex formation. Many proteins

cannot work as a single substance but can work only

with forming protein complex with other proteins.

Thus, if we interrupt the protein complex formation,

we can also interrupt the function of protein complex.

In Table 2, we have listed protein complex candidates

inferred by FAMS. Since FAMS uses a representative

protein within each cluster having more than 95 % se-

quence similarity as a model protein, there are some-

times more than a thousand model proteins which can

bind to other proteins. We can immediately recognize

that the list includes many reasonable outcomes. For

StructurePredictionwithFAMSforProteinsScreenedCriticallytoAutoimmuneDiseasesbaseduponBioinformatics

265

example, there are 52 model proteins listed between

CSF3 and CSF3R. By name, it is rather obvious that

they are possibly ligand and its receptor. On the other

hand, there are 186 model proteins between CSF3R

and CSF1R. This represents the possibility that each

monomercan form functional protein which can func-

tion together, possibly as a receptor. In addition to

this, both CSF3R and CSF1R most frequently have

non-zero model proteins to bind to each of other refer-

ence proteins. It is reasonable since many can bind to

them as ligand or can form a receptor together. Close

look at this table will give us fruitful information re-

sources to ﬁnd drug target by the termination of the

formation of protein complex.

In addition to these known and expected protein

complex formation, there are many new ﬁndings of

protein complex formation candidates. Fig. 4 shows

one of such possible candidates. In Table 2, there

are 410 possible candidate pairs between CSF1R and

PECAM1. Among these, there is one pair having 61

atom pairs contacting with each other. This means,

there is a structure on PDB (2ZJS) which includes

monomers whose protein structures are expected to be

similar to CSF1R and PECAM1, respectively. 2ZJS

is SecYE translocon, which are expected to func-

tion as a protein-conducting channel(Tsukazaki et al.,

2008). Although this protein complex was found in

Thermus thermophilus, since this kind of proteins are

expected to be highly conserved, it is highly possi-

ble that CSF1R and PECAM1 form protein complex

which is secreted across or integrated into membranes

and play critical role in autoimmune diseases. Thus

if we can ﬁnd the drug which terminates the protein

complex formation between CSF1R and PECAM1, it

may cure autoimmune diseases.

Predicted protein-protein complex candidates de-

tected are reported in Table 2, but detailed discussion

is deferred due to space constrains. This will be re-

ported in some other opportunity.

4 CONCLUSIONS

In this study, we have demonstrated that how well

FAMS can predict protein structures of candidate

genes which may play critical roles in autoimmune

diseases. Based upon inferred structure, we could an-

notate protein functions, could infer possible ligand

pockets which can bind to proteins, and could ﬁnd

possible pairs of proteins which can form proten com-

plex, which can be possible candidates of the drug tar-

get. It is conﬁrmed that FAMS can work with other

bioinformatic programs.

REFERENCES

Abd-Allah, S. H., el Shal, A. S., Shalaby, S. M., Pasha,

H. F., Abou el Saoud, A. M., el Najjar, A. R., and

el Shahawy, E. E. (2012). PADI4 polymorphisms

and related haplotype in rheumatoid arthritis patients.

Joint Bone Spine, 79(2):124–128.

Bingle, C. D. and Vyakarnam, A. (2008). Novel innate

immune functions of the whey acidic protein family.

Trends Immunol., 29(9):444–453.

Bouma, B. N., Vlooswijk, R. A., and Grifﬁn, J. H. (1983).

Immunologic studies of human coagulation factor XI

and its complex with high molecular weight kinino-

gen. Blood, 62(5):1123–1131.

Coito, A. J., de Sousa, M., and Kupiec-Weglinski, J. W.

(2000). Fibronectin in immune responses in organ

transplant recipients. Dev. Immunol., 7(2-4):239–248.

Cronstein, B. N. (2007). Interleukin-6–a key mediator of

systemic and local symptoms in rheumatoid arthritis.

Bull NYU Hosp Jt Dis, 65 Suppl 1:S11–15.

Dogliotti, G., Galliera, E., Licastro, F., Porcellini, E., and

Corsi, M. M. (2010). Serum neutrophil gelatinase-

B associated lipocalin (NGAL) levels in Down’s syn-

drome patients. Immun Ageing, 7 Suppl 1:S7.

Fong, L., Ruegg, C. L., Brockstedt, D., Engleman, E. G.,

and Laus, R. (1997). Induction of tissue-speciﬁc au-

toimmune prostatitis with prostatic acid phosphatase

immunization: implications for immunotherapy of

prostate cancer. J. Immunol., 159(7):3113–3117.

Frey, T. G., Chan, S. H., and Schatz, G. (1978). Struc-

ture and orientation of cytochrome c oxidase in crys-

talline membranes. Studies by electron microscopy

and by labeling with subunit-speciﬁc antibodies. J.

Biol. Chem., 253(12):4389–4395.

Gabr, A. A., Reed, M., Newman, D. R., Pohl, J., Khosla,

J., and Sannes, P. L. (2007). Alterations in cytoskele-

tal and immune function-related proteome proﬁles in

whole rat lung following intratracheal instillation of

heparin. Respir. Res., 8:36.

Gervin, K., Vigeland, M. D., Mattingsdal, M., Hammer?,

M., Nygard, H., Olsen, A. O., Brandt, I., Harris,

J. R., Undlien, D. E., and Lyle, R. (2012). DNA

methylation and gene expression changes in monozy-

gotic twins discordant for psoriasis: identiﬁcation

of epigenetically dysregulated genes. PLoS Genet.,

8(1):e1002454.

Heijmans, J., Buller, N. V., Hoff, E., Dihal, A. A., van der

Poll, T., van Zoelen, M. A., Bierhaus, A., Biemond,

I., Hardwick, J. C., Hommes, D. W., Muncan, V., and

van den Brink, G.R. (2012). Rage signalling promotes

intestinal tumourigenesis. Oncogene.

Javierre, B. M., Fernandez, A. F., Richter, J., Al-

Shahrour, F., Martin-Subero, J. I., Rodriguez-Ubreva,

J., Berdasco, M., Fraga, M. F., O’Hanlon, T. P., Rider,

L. G., Jacinto, F. V., Lopez-Longo, F. J., Dopazo,

J., Forn, M., Peinado, M. A., Carreno, L., Sawalha,

A. H., Harley, J. B., Siebert, R., Esteller, M., Miller,

BIOINFORMATICS2013-InternationalConferenceonBioinformaticsModels,MethodsandAlgorithms

266

F. W., and Ballestar, E. (2010). Changes in the pat-

tern of DNA methylation associate with twin discor-

dance in systemic lupus erythematosus. Genome Res.,

20(2):170–179.

Jin, T., Perry, A., Jiang, J., Smith, P., Curry, J. A., Unter-

holzner, L., Jiang, Z., Horvath, G., Rathinam, V. A.,

Johnstone, R. W., Hornung, V., Latz, E., Bowie, A. G.,

Fitzgerald, K. A., and Xiao, T. S. (2012). Struc-

tures of the HIN Domain:DNA Complexes Reveal

Ligand Binding and Activation Mechanisms of the

AIM2 Inﬂammasome and IFI16 Receptor. Immunity,

36(4):561–571.

Latonen, L., Jarvinen, P. M., Suomela, S., Moore, H. M.,

Saarialho-Kere, U., and Laiho, M. (2010). Ultravi-

olet B radiation regulates cysteine-rich protein 1 in

human keratinocytes. Photodermatol Photoimmunol

Photomed, 26(2):70–77.

Lin, J. X., Li, P., Liu, D., Jin, H. T., He, J., Rasheed, M. A.,

Rochman, Y., Wang, L., Cui, K., Liu, C., Kelsall,

B. L., Ahmed, R., and Leonard, W. J. (2012). Critical

Role of STAT5 Transcription Factor Tetramerization

for Cytokine Responses and Normal Immune Func-

tion. Immunity, 36(4):586–599.

Meylan, E., Burns, K., Hofmann, K., Blancheteau, V., Mar-

tinon, F., Kelliher, M., and Tschopp, J. (2004). RIP1 is

an essential mediator of Toll-like receptor 3-induced

NF-kappa B activation. Nat. Immunol., 5(5):503–507.

O-Sullivan, I., Chopra, A., Carr, J., Kim, T. S., and Co-

hen, E. P. (2008). Immunity to growth factor receptor-

bound protein 10, a signal transduction molecule, in-

hibits the growth of breast cancer in mice. Cancer

Res., 68(7):2463–2470.

O’Hanlon, T. P., Li, Z., Gan, L., Gourley, M. F., Rider,

L. G., and Miller, F. W. (2011a). Plasma proteomic

proﬁles from disease-discordant monozygotic twins

suggest that molecular pathways are shared in mul-

tiple systemic autoimmune diseases*. Arthritis Res.

Ther., 13(6):R181.

O’Hanlon, T. P., Rider, L. G., Gan, L., Fannin, R., Paules,

R. S., Umbach, D. M., Weinberg, C. R., Shah,

R. R., Mav, D., Gourley, M. F., and Miller, F. W.

(2011b). Gene expression proﬁles from discordant

monozygotic twins suggest that molecular pathways

are shared among multiple systemic autoimmune dis-

eases. Arthritis Res. Ther., 13(2):R69.

Pan, H. F., Leng, R. X., Tao, J. H., Li, X. P., and Ye, D. Q.

(2011). Ets-1: a new player in the pathogenesis of sys-

temic lupus erythematosus? Lupus, 20(3):227–230.

Quek, A. M., Britton, J. W., McKeon, A., So, E., Lennon,

V. A., Shin, C., Klein, C. J., Watson, R. E., Kotse-

nas, A. L., Lagerlund, T. D., Cascino, G. D., Wor-

rell, G. A., Wirrell, E. C., Nickels, K. C., Aksamit,

A. J., Noe, K. H., and Pittock, S. J. (2012). Autoim-

mune Epilepsy: Clinical Characteristics and Response

to Immunotherapy. Arch Neurol, 69(5):582–593.

Samadani, U. and Costa, R. H. (1996). The transcriptional

activator hepatocyte nuclear factor 6 regulates liver

gene expression. Mol. Cell. Biol., 16(11):6273–6284.

Sarkar, S., Song, Y., Sarkar, S., Kipen, H. M., Laumbach,

R. J., Zhang, J., Strickland, P. A., Gardner, C. R.,

and Schwander, S. (2012). Suppression of the NF-B

pathway by diesel exhaust particles impairs human an-

timycobacterial immunity. J. Immunol., 188(6):2778–

2793.

Selvaraj, R. K., Shanmugasundaram, R., and Klasing, K. C.

(2010). Effects of dietary lutein and PUFA on PPAR

and RXR isomer expression in chickens during an in-

ﬂammatory response. Comp. Biochem. Physiol., Part

A Mol. Integr. Physiol., 157(3):198–203.

Silva, J. A., Ferrucci, D. L., Peroni, L. A., Abrahao, P. G.,

Salamene, A. F., Rossa-Junior, C., Carvalho, H. F.,

and Stach-Machado, D. R. (2012). Sequential IL-23

and IL-17 and increased Mmp8 and Mmp14 expres-

sion characterize the progression of an experimental

model of periodontal disease in type 1 diabetes. J.

Cell. Physiol., 227(6):2441–2450.

Skeie, G. O., Bentsen, P. T., Freiburg, A., Aarli, J. A., and

Gilhus, N. E. (1998). Cell-mediated immune response

against titin in myasthenia gravis: evidence for the in-

volvement of Th1 and Th2 cells. Scand. J. Immunol.,

47(1):76–81.

Taguchi, Y.-h. (2010). A signiﬁcance test based upon PCA.

SIG-DMSM, A903:38–47.

Tripp, R. A., Barskey, A., Goss, L., and Anderson, L. J.

(2002). Substance P receptor expression on lympho-

cytes is associated with the immune response to res-

piratory syncytial virus infection. J. Neuroimmunol.,

129(1-2):141–153.

Tsukazaki, T., Mori, H., Fukai, S., Ishitani, R., Mori, T.,

Dohmae, N., Perederina, A., Sugita, Y., Vassylyev,

D. G., Ito, K., and Nureki, O. (2008). Conforma-

tional transition of Sec machinery inferred from bac-

terial SecYE structures. Nature, 455(7215):988–991.

Tsunoda, T., Takashima, Y., Tanaka, Y., Fujimoto, T., Doi,

K., Hirose, Y., Koyanagi, M., Yoshida, Y., Okamura,

T., Kuroki, M., Sasazuki, T., and Shirasawa, S. (2010).

Immune-related zinc ﬁnger gene ZFAT is an essential

transcriptional regulator for hematopoietic differenti-

ation in blood islands. Proc. Natl. Acad. Sci. U.S.A.,

107(32):14199–14204.

Umeyama, H. and Iwadate, M. (2002). Fams and famsbase

for protein structure. In Current Protocols in Bioin-

formatics. John Wiley & Sons, Inc.

Watson, F. L., Puttmann-Holgado, R., Thomas, F., Lamar,

D. L., Hughes, M., Kondo, M., Rebel, V. I., and

Schmucker, D. (2005). Extensive diversity of Ig-

superfamily proteins in the immune system of insects.

Science, 309(5742):1874–1878.

Zhou, X., Tan, F. K., Xiong, M., Arnett, F. C., and Feghali-

Bostwick, C. A. (2005). Monozygotic twins clinically

discordant for scleroderma show concordance for ﬁ-

broblast gene expression proﬁles. Arthritis Rheum.,

52(10):3305–3314.

Zoller, K. E., MacNeil, I. A., and Brugge, J. S. (1997).

Protein tyrosine kinases Syk and ZAP-70 display dis-

tinct requirements for Src family kinases in immune

response receptor signal transduction. J. Immunol.,

158(4):1650–1659.

StructurePredictionwithFAMSforProteinsScreenedCriticallytoAutoimmuneDiseasesbaseduponBioinformatics

267