Archives of Oral Biology 46 (2001) 249–260

Isolation and partial sequencing of potentially

odontoblast-specific/enriched rat cDNA clones obtained by

suppression subtractive hybridization

Rama Dey a, Ho-Hyun Son b, Moon-Il Cho a,*a Department of Oral Biology, School of Dental Medicine, State Uni6ersity of New York at Buffalo, B-10 Foster Hall, Buffalo,

NY 14214-3092, USAb Department of Conser6ati6e Dentistry, College of Dentistry, Seoul National Uni6ersity, Seoul 110-744, South Korea

Accepted 6 September 2000

Abstract

Odontoblasts, which are responsible for dentine formation, are known to synthesize unique gene products such as

dentine sialophosphoprotein. To further identify and clone novel odontoblast-specific genes, a suppression subtractive

hybridization technique was used here. Differentially or predominantly expressed cDNAs in odontoblasts of rat

incisors were obtained by subtracting the common cDNAs expressed in odontoblasts, osteoblasts and pulp cells.

Clones were then partially sequenced and analysed for nucleotide sequence homology by the basic local alignment

search tool program. From a total of 1290 clones analysed, 538 odontoblast-enriched clones were identified in the

subtracted cDNA library. Out of 538 clones, 498 clones (92.6%) demonstrated high identity with genes in the

GenBank database. In contrast, 31 clones (5.7%) showed low sequence identity with known genes, among which 18

clones (3.3%) were observed more than once, thereby possibly representing odontoblast-specific/enriched genes. The

majority (390 clones; 72.5%) of the clones with high homology to known genes were found to be the rat/mouse

dentine sialophosphate by dot-blot analysis (326 clones) and sequencing (64 clones). The second highest enrichment

(39 clones) was for phosphate-regulating gene with homology to endopeptidase on the X-chromosome, which codes

for a neutral endopeptidase. After suppression subtractive hybridization, several cDNAs that are commonly present

in osteoblasts and odontoblasts appeared unsuppressed. Therefore, a rat odontoblast-specific/enriched subtraction

cDNA library has been created from which a number of potentially novel genes for odontoblasts could be identified.

© 2001 Elsevier Science Ltd. All rights reserved.

Keywords: Subtraction cDNA library; Sequencing of cDNA clones; Suppression subtractive hybridization; Rat odontoblasts; Genes

in odontoblasts

www.elsevier.com/locate/archoralbio

Abbre6iations: BLAST, basic local alignment search tool; BSA, bovine serum albumin; DMEM, Dulbecco modified Eagle

medium; FCS, fetal calf serum; G3PDH, glyceraldehyde-3-phosphate dehydrogenase; IPTG, isopropylthio-b-D-galactoside; PEX,

phosphate-regulating gene with homology to endopeptidase on the X-chromosome; Poly(A)+ , polyadenine residues; RT-PCR,

reverse transcriptase-polymerase chain reaction; SDS, sodium dodecyl sulphate; SSPE, ; X-gal, 5-bromo-4-chloro-3-indolyl-b-D-

galactoside.

* Corresponding author. Tel.: +1-716-8292605; fax: +1-716-8293942.

E-mail address: mooncho@buffalo.edu (M.-I. Cho).

0003-9969/01/$ - see front matter © 2001 Elsevier Science Ltd. All rights reserved.

PII: S 0003 -9969 (00 )00117 -5

R. Dey et al. / Archi6es of Oral Biology 46 (2001) 249–260250

1. Introduction

Dentine and bone are mineralized tissues of the teeth

and the skeleton, respectively. Dentine is formed by

odontoblasts that are differentiated from ectomes-

enchymal cells (Ten Cate, 1998), while bone is synthe-

sized by osteoblasts derived from mesenchymal cells

(Friedenstein, 1976; Owen and Friedenstein, 1988). In

spite of distinct differences in the developmental origin

and functions of these tissues, both have a similar

matrix composition. Their organic matrices are com-

posed primarily of type I collagen and several non-col-

lagenous proteins, some of which are present in both;

these include osteonectin, osteopontin, osteocalcin and

bone sialoprotein (Butler, 1998). However, dentine also

contains unique proteins such as dentine phospho-

protein (Veis and Perry, 1967; Dickson et al., 1975;

George et al., 1998) and dentine sialoprotein (Butler,

1987; Ritchie et al., 1998). Recently, both were found to

be the product of a single gene for dentine sialophos-

phoprotein MacDougall et al., 1997). It remains un-

clear whether dentine has other unique matrix

components that are not present in bone.

While studies on the expression of osteoblast-specific

genes have provided important insights (Chen et al.,

1992; Lian and Stein, 1992; Lian et al., 1998), leading

to the identification of the osteogenic master gene,

Cbfa1 (Ducy and Karsenty, 1998), relatively little is

known about odontoblast differentiation and

odontogenesis.

As a first step to better understanding the regulatory

mechanisms responsible for odontoblast differentiation

and dentinogenesis, we have attempted to identify and

clone odontoblast-specific genes of the rat. There are

several forms of subtractive hybridization that can be

used to identify and isolate cDNAs of differentially

expressed genes (Duguid and Dinauer, 1990; Hara et

al., 1991; Lee et al., 1991). To selectively amplify differ-

entially expressed target cDNA fragments and simulta-

neously suppress non-target DNA amplification, a

suppression subtractive hybridization technique was in-

troduced (Diatchenko et al., 1996). This is a PCR-based

cDNA subtraction method that includes two important

procedures, normalization and subtraction. The nor-

malization step is designed to enrich differentially ex-

pressed cDNAs in the target sample, while the

subtraction step excludes commonly expressed cDNAs

between the target and the non-target samples. This

technique, therefore, can generate a target sample-spe-

cific cDNA library. We have applied this technique to

obtain differentially expressed cDNAs in odontoblasts

by subtracting those cDNAs that are common among

odontoblasts, osteoblasts, and pulp cells. The differen-

tially expressed cDNAs were then cloned to create a rat

odontoblast subtraction cDNA library. Here, we report

a compilation of partial sequences and dot-blot analysis

of 538 cDNA clones, some of which could be identified

as potentially novel odontoblast-specific genes.

2. Materials and methods

2.1. Extraction of mRNA

Total RNA was isolated from the odontoblasts/pulp

cells of rat mandibular incisors, and from cultured

dental papilla cells as well as osteoblastic cells to use as

tester and driver RNA samples, respectively. To isolate

total RNA from the odontoblasts/pulp cells, the

mandibular incisors were dissected out from 4-week-old

female Sprague–Dawley rats (Fig. 1). To eliminate

possible contamination with RNAs from enamel organ

and the periodontal ligament (Fig. 2), the unmineral-

ized proximal portion of the incisor was removed (Fig.

1), and all surrounding tissues were scraped off from

the root surface. The odontoblasts/pulp tissues were

carefully removed under a dissecting microscope after

splitting the incisors longitudinally into two halves. To

obtain total RNA from osteoblastic cells, calvariae of

2-day-old rat pups were collected, kept in liquid nitro-

gen and finely minced. To prepare total RNA from

dental papilla cells, cells were cultured as described

below. After homogenization of cultured cells or tis-

sues, total RNA was first isolated with Trizol reagent

Fig. 1. Diagrammatic representation of a longitudinal view of a rat mandible. To eliminate possible contamination with RNAs from

enamel organ and Hertwig’s epithelial root sheath, the unmineralized apical portion (dotted line) was removed.

R. Dey et al. / Archi6es of Oral Biology 46 (2001) 249–260 251

Fig. 2. Light micrographs of cross-sections (1 mm, toluidine

blue stain) of a rat mandible containing a mandibular incisor.

A low-magnification micrograph (a) shows the pulp chamber

(P), and surrounding tissues such as the alveolar bone (AB),

enamel organ (EO) and periodontal ligament (PDL). These

tissues were removed to prevent contamination with RNAs

from the tissues ×50. A high-magnification micrograph (b) of

the rectangular area in (a) demonstrates odontoblasts (O),

pulp cells (P), and their associated predentine (PD) and den-

tine (D). Odontoblasts and pulp cells were collected for the

preparation of a test RNA sample ×400.

were cultured in the absence of 10 mM b-glycerophos-

phate and 50 mg/ml ascorbic acid. At 5 days after

treatment, the expression of dentine sialophospho-

protein in total RNA was examined by RT-PCR as

described below.

2.3. RT-PCR of total RNA from odontoblasts/pulp

cells and cultured dental papilla cells

Total RNA from each cell type was digested with 6

units of RNase-free DNase (Promega, Madison, WI,

USA) in 100 mM Tris–HCl (pH 7.5), containing 50

mM MgCl2 at 37°C for 1 h. Using 1 ml of 0.5 mg/ml

oligo (dT) 12–18 primer, 2 mg of digested total RNA

was reverse-transcribed with 200 units of M-MLV

RNaseH− reverse transcriptase (SuperScript II RT;

Life Technologies. Gaithersburg, MD, USA) according

to the manufacturer’s instructions. In brief, the oligo

(dT) primer was annealed to 2 mg of total RNA from

each cell type in a total volume of 12 ml with water by

heating the mixture at 70°C for 10 min and cooling on

ice for 2 min. The annealed primer–RNA was then

mixed with 200 units of SuperScript II RT in 20 ml of

preheated (42°C) 20 mM Tris–HCl (pH 8.4), contain-

ing 50 mM KCl, 2.5 mM MgCl2, 10 mM dithiothreitol,

0.5 mM each of dATP, dCTP, dGTP and dTTP. The

mixture was incubated at 42°C for 50 min, terminated

at 70°C for 15 min and chilled on ice. The use of this

SuperScript preamplification system and subsequent

RNase H digestion (2 units) for 20 min at 37°C,

followed by PCR amplification using dentine sialophos-

phoprotein primers [dentine sialophosphoprotein sense

primer: 5%-CACATCCAGGAACCGCAGCACA-3%

(nucleotide 136–157) and antisense primer: 5%-CCT-

TACTCTCCTTTGCTTCCTC-3% (nucleotide 445–

466)], made it possible to detect dentine

sialophosphoprotein cDNA in these samples. Amplifi-

cation was performed under the following conditions:

initial denaturation at 94°C for 2 min; 25 cycles of

amplification at 94°C for 45 s, 50°C for 30 s, 72°C for

45 s; a final extension at 72°C for 7 min.

2.4. Tester double-stranded cDNA preparation

To generate an odontoblast-specific cDNA library,

tester double-stranded cDNAs were prepared using the

PCR-Select cDNA subtraction kit (Clontech Laborato-

ries Inc., Palo Alto, CA, USA) according to the manu-

facturer’s manuals. In order to synthesize first-strand

cDNA, 10 pmol of a cDNA synthesis primer was

annealed to 2 mg odontoblast/pulp cell poly(A)+

RNA.

Ligation efficiency of the adaptors to the tester dou-

ble stranded cDNAs was tested according to the manu-

facturer’s instructions, and products were analysed by

agarose gel electrophoresis.

(Life Technologies, Grand Island, NY, USA), and

poly(A)+ RNA was then extracted using the Oligotex

mRNA purification kit (Qiagen, Valencia, CA, USA)

with some modifications (Chien et al., 1999). Integrity

of the RNA was determined by electrophoresis on an

1% formaldehyde denaturing agarose gel.

2.2. Culture of dental papilla mesenchymal cells

To prepare mRNA from cultured dental papilla mes-

enchymal cells, dental papillae were isolated from the

developing first mandibular molars of 2-day-old

Sprague–Dawley rat pups. After dissecting out the

molars, approx. 1 mm thick mesiodistal pieces were cut

from the midplane of the molars. To eliminate a possi-

bility of contamination with odontoblasts, ameloblasts

and Hertwig’s root sheath epithelial cells, the newly

formed crown and its associated cells, and root sheath

epithelial cells in the apical portion of a developing

tooth were carefully removed under a dissecting micro-

scope. The remaining dental papilla was then finely

chopped and digested in 1 ml of DMEM containing 3

mg of collagenase/dispase (Sigma Chemical Co., St

Louis, MO) at 37°C for 30 min. Dissociated cells were

collected by centrifugation, plated at a density

of 5×103 cells/cm2, and cultured in DMEM supple-

mented with 10% FCS and antibiotics. Confluent cells

R. Dey et al. / Archi6es of Oral Biology 46 (2001) 249–260252

2.5. Dri6er double-stranded cDNA preparation

Driver double stranded cDNAs were synthesized us-

ing 2 mg of poly(A)+ RNA each from calvarial os-

teoblasts and cultured dental papilla mesenchymal cells.

They were then blunt-ended by digesting with RsaI and

no adaptor was ligated to their ends. The RsaI-digested

driver cDNAs from the two cell types were mixed

together in equal proportions and used as driver

cDNA.

2.6. Subtracti6e hybridization

In order to obtain the differentially expressed cD-

NAs, two hybridizations were performed according to

the manufacturer’s instructions. The first hybridization

was to achieve equalization and enrichment of differen-

tially expressed sequences, while the second hybridiza-

tion was to form double-stranded tester molecules with

different adapters on each end, that correspond to

differentially expressed cDNAs.

2.7. Suppression PCR amplification

Two PCR amplifications were carried out according

to manufacturer’s protocols. The first was to amplify

exponentially double-stranded cDNAs with different

adaptor sequences on each end, and the second was to

enrich the differentially expressed cDNAs, and further

reduce background.

2.8. Efficiency of dentine sialophosphoprotein

amplification

In order to investigate the efficiency of enrichment of

odontoblast-specific/enriched cDNAs, the abundance of

dentine sialphosphoprotein in subtracted and unsub-

tracted secondary PCR products was assessed by PCR

amplification using the dentine sialphosphoprotein

primers. Samples were collected at 18th, 23rd, 28th and

33rd cycle intervals and analysed by gel electrophoresis.

We expected to observe a band representing the ap-

prox. 330-bp fragment of dentine sialophosphoprotein.

2.9. Cloning of the subtracted cDNA

The subtracted cDNAs obtained after secondary

PCR were cloned with a TA cloning kit (Invitrogen

Corporation, Carlsbad, CA, USA). The cDNAs were

inserted into the PCR 2.1 vector by incubating 3 ml of

the secondary PCR amplification and 2 ml of the vector

(25 ng/ml) with 4 Weiss units of T4 DNA ligase at 14°C

overnight. Escherichia coli (TOP 10F%) cells were trans-

formed with 2 ml of the plasmid, plated and incubated

at 37°C overnight on five 2× YT agar plates, contain-

ing 50 mg/ml ampicillin, IPTG and X-gal. Thus, a

library of putative odontoblast-specific/enriched cD-

NAs was constructed. The cDNA inserts were flanked

on both sides by EcoRI sites to facilitate their isolation.

2.10. Differential screening of subtracted cDNA library

In an attempt to screen the colonies with cDNAs

specific to odontoblasts, a differential screening proce-

dure was applied. For this purpose, dot-blotted mem-

branes with PCR reaction products from individual

colonies were prepared and screened by hybridization

with subtracted, unsubtracted and reverse cDNAs

probes.

2.10.1. Preparation of dot-blot membranes

Individual colonies on five selective 2×YT agar

plates were isolated and grown in 96-multiwell plates

containing 250 ml of 2×YT medium and 50 mg/ml

ampicillin. For PCR amplification, 2 ml from each well

was mixed with nested PCR primers 1 and 2 in 20 ml

reactions. PCR was performed with an initial denatura-

tion at 94°C for 30 s and then 25 cycles at 95°C for 30

s and 68°C for 3 min. A volume (15 ml) of each PCR

reaction product was mixed with 10 ml of 0.3 M NaOH/

15% Ficoll/0.5% bromophenol blue, and 0.2 ml of each

mixture was dot-blotted on to nylon membranes (Schle-

icher and Schuell, Inc., NH, USA) and cross-linked

with an ultraviolet linking device under 120 mJ.

2.10.2. Preparation of probes

2.10.2.1. Subtracted cDNA probes. Reaction products of

the secondary PCR of the subtracted cDNAs were

phenol-extracted, precipitated and digested with RsaI

and EaeI (New England Biolabs, Beverly, MA, USA)

to remove the adaptors from the ends of cDNAs. The

adaptors were then separated by agarose gel elec-

trophoresis, while cDNAs were purified using the Qia-

gen DNA purification kit (Qiagen, Valencia, CA,

USA). These served as the positive probes for screening

colonies.

2.10.2.2. Unsubtracted cDNA probes. In the same way,

the unsubtracted cDNAs from odontoblasts/pulp cells

were prepared and served as the unsubtracted probes

for screening colonies.

2.10.2.3. Re6erse cDNA probes. To prepare the reverse

cDNA probes, the entire procedure for subtraction was

repeated in reverse. In this case, the tester (odonto-

blasts/pulp cells) and the driver (calvarial osteoblasts

and cultured dental papilla mesenchymal cells) that

have been used in the previous subtraction procedures

served as the driver and the tester, respectively. Thus,

the recombinants with transcripts in the actual driver

cells could be identified and eliminated.

R. Dey et al. / Archi6es of Oral Biology 46 (2001) 249–260 253

2.10.3. Hybridization

Probes were labelled with [a-32P] dCTP according to

protocol (107 counts per min per 100 ng) using the

DECAprime II DNA labelling kit (Ambion, Austin,

TX, USA), and were hybridized to the DNA dot-blot

membranes in hybridization buffer containing 50% for-

mamide, 5× Denhardt’s solution, 0.5% SDS, 100 mg/

ml single-stranded DNA (Life Technologies, Grand

Island) and 5× SSPE at 42°C overnight. The blots

were washed twice with 2× SSPE/0.1% SDS at 37°C,

1× SSPE/0.1% SDS at 65°C, 0.1× SSPE/0.1% SDS at

65°C, and then autoradiographed.

2.11. Dentine sialophosphoprotein probe preparation

and hybridization

In order to avoid unnecessary sequencing of dentine

sialophosphoprotein clones, we attempted to identify

dentine sialphosphoprotein clones using dot-blot analy-

sis and eliminate them before their sequencing. Thus,

dentine sialophosphoprotein probes were prepared

from five clones that had been identified as dentine

sialophosphoprotein by sequence analysis. The inserts

were amplified using nested PCR primers 1 and 2 for 35

cycles at 94°C for 30 s, 55°C for 45 s, and 72°C for 45

s. The adaptors were removed by digestion with RsaI

and EaeI, and the amplified products were analysed by

1% agarose gel electrophoresis. The products were then

cut, purified using the Qiaex II gel extraction kit (Qia-

gen), labelled with [a-32P] dCTP using the DECAprime

II DNA labelling kit, and used to hybridize the same

dot-blot membranes that had previously been used for

differential screening under the conditions described

above.

2.12. DNA sequencing and homology search

For DNA sequencing, plasmid mini-preps of all po-

tentially positive clones were prepared using the Wizard

Plus Minipreps DNA Purification System (Promega,

Madison, WI, USA) and subjected to automated se-

quencing (Applied Biosystems, Foster City, CA, USA)

at the CAMBI facility, SUNY at Buffalo. Nucleic acid

homology search was performed against the GenBank

DNA database (non-redundant) using the advanced

BLAST program (Altschul et al., 1997) via a network

connection to the National Center for Biotechnology

Information (National Institute of Health, Bethesda).

Partial sequences ranging from 200 to 750 nucleotides

were used in the nucleic acid homology search. Quality

of match is given as % identity in base pairs for the

stretches of highest nucleotide match. Clones have been

divided into low homology (scores B100) for small

stretches, intermediate homology (scores 100–225) for

intermediate stretches, and high homology (scores \

225) for long stretches of highest nucleotide homology

to known genes in the GenBank DNA database.

2.13. Northern blot analysis for clones OD-314 and

OD-654 in rat tissues

It was interesting to know whether some of the

differentially expressed cDNAs that have been iden-

tified by suppressive subtractive hybridization were ac-

tually expressed specifically in odontoblasts/pulp cells.

To investigate this aspect, we selected clones OD-314

and OD-654 (that were detected more than once and

showed low homology) and studied their expression in

various rat tissues using Northern blot analysis as

described by Chien et al. (1999). Blot membranes were

prepared in our laboratory from agarose gels contain-

ing 10 mg total RNA in each lane from rat odontoblast/

pulp, calvaria, heart, kidney and liver. In order to

examine their expression in additional tissues, a multi-

ple-tissue blot membrane containing 1 mg mRNA each

from rat brain, heart, kidney, liver, lung, skeletal mus-

cle and spleen was purchased from Clontech. cDNA

probes used for blotting included the approx. 600-bp

insert from clone OD-314 and the approx. 650-bp insert

from clone OD-654, the rat G3PDH insert of 1.2-kb,

and the b-actin cDNA (Clontech). Both G3PDH and

b-actin cDNA probes were used to normalize the actual

amount of RNA loaded.

3. Results

3.1. Collection of odontoblasts/pulp tissues

The odontoblasts/pulp tissues within the fully miner-

alized dentine only were isolated (Fig. 1) and subjected

to total RNA isolation. Morphologically, these odonto-

blasts are fully differentiated and actively engaged in

dentinogenesis (Fig. 2).

3.2. Expression of dentine sialophosphoprotein in

odontoblasts/pulp cells and cultured dental papilla cells

To determine if dental papilla mesenchymal cells

differentiate into odontoblasts during a 5-day culture in

the absence of b-glycerophosphate and ascorbic acid,

we examined the amplification of dentine sialophospho-

protein in total RNA from these cells using RT-PCR,

and compared it with that of the odontoblast/pulp cells

as a positive control. The odontoblast/pulp tissue

clearly showed a distinct band representing an approx.

330-bp dentine sialophosphoprotein fragment, while the

cultured dental papilla cells did not reveal the approx.

330-bp dentine sialophosphoprotein fragment (Fig. 3).

These results indicate that dental papilla cells remain

undifferentiated under the culture condition and were

thus used to prepare one of the driver cDNAs for

creating a subtracted odontoblast-specific cDNA

library.

R. Dey et al. / Archi6es of Oral Biology 46 (2001) 249–260254

3.3. Ligation efficiency of adaptors to tester

double-stranded cDNA

Efficiency of ligation of the adaptors one or two to

the tester double stranded cDNAs was examined using

(1) G3PDH 3% and 5% primers as control, and (2)

G3PDH 3% primer and PCR primer 1 spanning the

adaptor-ligated junctions. As expected, an approx. 500-

bp band was observed using G3PDH 3% and 5% primers

(Fig. 4). When the G3PDH 3% primer and PCR primer

1 were used, an approx. 1.2 kb band in either adaptor

1- or adaptor 2-ligated tester cDNAs was amplified

similar (Fig. 4), indicating an efficient ligation of the

adaptors to tester cDNAs.

3.4. Efficiency of dentine sialophosphoprotein

amplification

The amplification of odontoblast-specific genes as

evidence of efficient enrichment was investigated by

assessing dentine sialophosphoprotein amplification in

subtracted as well as unsubtracted samples. In the

subtracted samples, a weak band appeared as early as

at the 18th cycle, which increased in intensity from the

23rd through the 33rd cycles. In contrast, no amplifica-

tion was observed in the unsubtracted samples (Fig. 5).

This is a clear indication that the subtracted sample was

highly enriched in odontoblast-specific cDNAs.

Fig. 4. PCR analysis of ligation efficiency of adaptor 1 (Ad 1)

or 2 (Ad 2) to tester double-stranded cDNAs (odontoblasts

and pulp cells). When G3PDH 5% and 3% primers were used as

control, an approx. 500-bp band with high intensity was

observed (lanes 1 and 3). When G3PDH 3% primer and PCR

primer 1 were used, an approx. 1.2-kb band (lanes 2 and 4)

was amplified with similar intensity to that of the control.

Lane M, DNA size markers.

3.5. Cloning and differential screening by blotting

A total of 1290 white bacterial colonies were iden-

tified on five selective 2× YT plates, suggesting highly

efficient cloning of the secondary PCR products into

PCR 2.1 vector. A few blue colonies were counted. To

identify odontoblast-specific/enriched clones, amplified

Fig. 5. Enriched dentine sialophosphoprotein (DSPP) amplifi-

cation in the subtracted samples by PCR. When PCR amplifi-

cation of DSPP was assessed in the subtracted sample using

antisense and sense DSPP primers, a weak band of An approx.

330-bp DSPP fragment appeared at the 18th, and its intensity

increased thereafter at the 23rd, 28th and 33rd cycle (lanes

1–4, respectively). The band was not observed at the same

cycles (lanes 5–8, respectively) in the unsubtracted sample.

Lane M, DNA size markers.

Fig. 3. PCR amplification of dentine sialophosphoprotein in

tester (odontoblasts and pulp cells) and driver (cultured dental

papilla mesenchymal cells) using DSSP sense and antisense

primers. PCR amplification of DSSP in cultured dental papilla

mesenchymal cells produced no amplification was found (lane

1). The odontoblasts/pulp tissue revealed amplification of an

approx. 330-bp DSSP fragment (lane 2). Lane M, DNA size

markers.

R. Dey et al. / Archi6es of Oral Biology 46 (2001) 249–260 255

Fig. 6. A chart demonstrating homology of the partially

sequenced cDNA clones. The clones were grouped according

to the nucleic acid homology to previously reported genes in

the GenBank database. Non-recombinant clones are those

without any inserts.

cific/enriched clones, some of which appeared more

than once (Table 1) or once (Table 2). The second

group represented those showing more limited identity

(scores 100–225) (Table 3). This group occupied 1.1%.

The third group included clones showing low identity

(scores B100), thereby probably representing previ-

ously unidentified genes. They made up 5.7% and ap-

peared either many times (Table 4) or only once (Table

5). The remaining clones (0.6%) contained no insert.

The most frequently detected cDNAs were the den-

tine sialophosphoprotein (390 clones: 326 by blotting

and 64 by sequencing) which made up 72.5% of the

potentially odontoblast-specific/enriched clones. PEX

represents the second highest redundancy (39 clones),

followed by clones with very low homology to metallo-

panstimulin (nine clones) and a putative MtN3 protein

(seven clones). Redundancy was also observed in clones

with high homology to E25B protein (five clones),

osteocalcin (four clones), mitochondrial cytochrome ox-

idase (four clones), a1(I) and a2(I) collagen, lysyl ox-

idase, a receptor-type protein 88, odorant-binding

protein, monoamine oxidase (three clones each). Clones

found twice included those showing high homology to

mitochondrial genes, protein phosphatase, phosphodi-

esterase I and very low homology to placental villi

mRNA (clone SMAP41).

cDNAs were dot-blotted and hybridized with sub-

tracted, unsubtracted and reverse cDNA probes. Odon-

toblast-specific/enriched clones were positive with the

subtracted probe, but negative with both unsubtracted

and reverse cDNA probes (results not shown). From

this differential screening, a total of 538 odontoblast-

enriched clones were identified. In an attempt to elimi-

nate the clones expressing dentine sialophosphoprotein

before DNA sequencing, they were further subjected to

dot-blot analysis. For this purpose, five representative

dentine sialophosphoprotein clones showing complete

sequence identity to different regions of dentine

sialophosphoprotein cDNAs were selected and used as

probes. As a result, 326 dentine sialophosphoprotein

clones were identified and the remaining clones were

subjected to DNA sequencing.

3.6. Analysis of DNA sequence homology search

The partial sequences of the selected cDNA clones

were searched for sequence homologies in the GenBank

DNA database by BLAST. The cDNAs were divided

into four groups based on the degree of sequence

homologies observed (Fig. 6). The first group contained

clones whose sequences show high identity (scores \

225). They may represent genes that are either identical

or nearly identical to previously reported genes. This

represented 92.6% of the potentially odontoblast-spe-

Table 1

Clones detected more than once and showing high homology

to known genes (scores \225)a

SpeciesGene Frequency

R/M/H 390Dentine sialophosphoproteinb

R/H/MPEX protein 39

ME25B protein 5

RAcidic phosphoprotein 4

Osteocalcin R 4

Mitochondrial cytochrome oxidase R 4

R/M/Ha1(I) collagen 3

3Ma2(I) collagen

3RLysyl oxidase

3RReceptor-type protein

R 3Odorant-binding protein

RMonoamine oxidase B (Maobf3) 3

Mitochondrial genes for 16S R 2

rRNA, tRNA-Leu and

tRNA-Val-Phe

MProtein phosphatase, 2AB%alpha3 2

regulatory subunit

R/MPhosphodiesterase I 2

a Species: H, Homo sapiens; M, Mus musculus; R, Rattus

norvegicus.b Dentine sialophosphoprotein includes dentine phospho-

protein and dentine sialoprotein (MacDougall et al., 1997).

R. Dey et al. / Archi6es of Oral Biology 46 (2001) 249–260256

Table 2

Clones detected once and showing high homology to known genes (scores \225)a

Length sent Score SpeciesGeneClone

OD-13 34 kDa Phosphoprotein 415- 575 P

739 254CollagenOD-65 a1(XI) R

OD-82 Matrix metalloproteinase 663 357 S

491 686D6.1A proteinOD-124 R

255OD-126 32912q24.1 BAC RPCI11-412D9 H

350Dmx-like 1 464OD-127 H

675OD-167 Alpha subunit kidney-type Na+, K+-ATPase 718 R

656 844Sulfated glycoproteinOD-268 R

HelicaseOD-284 540 716 M

P190-BOD-310 487 652 M

656Ribosomal protein L5 pseudogene 250OD-330 H

Sulfonylurea receptor 2BOD-342 658 359 R

657NRD convertase 932OD-392 R

647OD-529 5733 b-hydroxysteroid dehydrogenase isomerase type II.2 R

642Cell-binding bone sialoprotein 492OD-588 R

640 654OD-596 L1ALB gene for LINE 1 repetitive seq R

225 375Mitochondrial cytochrome BOD-598 R

G-protein coupled thrombin receptorOD-618 631 944 R

Antizyme inhibitorOD-742 626 975 R

613Mitochondrial genome fragment 589OD-912 R

5%nucleotidaseOD-977 447 912 R

43034-kDa mov34 homologue 383OD-988 H

645 916OD-1008 Ribophorin II R

596Amelogenin 785OD-1033 R

603 700 ROD-1056 Ribosomal protein S4

653Heat shock protein 86 731OD-1161 M

Glycogen phosphorylase brainOD-1246 631 777 R

OD-1277 Osteonectin 644 337 R

a Species: H, Homo sapiens; M, Mus musculus; P, Plasmodium berghei; R, Rattus rattus; S, Sus scrofa.

3.7. Northern blot analysis

The expression of clones OD-314 and OD-654 was

found only in odontoblasts/pulp cells at high levels, but

was not observed in bone, brain, heart, kidney, liver,

lung, and skeletal muscle (Fig. 7). Interestingly, the

radioautograph of clone OD-314 demonstrated two

distinct bands (approx. 2.2 and 1.5 kb), whereas that of

OD-654 revealed several bands (ranging from approx.

7.2 to 2.4 kb).

4. Discussion

Interest in differentially expressed genes has provided

the impetus for the development of molecular tech-

niques to identify and isolate products of these genes.

Subtractive hybridization (Lee et al., 1991), differential

display of mRNA (Liang and Pardee, 1992), suppres-

sion subtractive hybridization (Diatchenko et al., 1996)

and more recently serial analysis of gene expression

(Velculescu et al., 1995) and DNA array (Nacht et al.,

1999) techniques have been introduced. We have used

the suppression subtractive hybridization method to

identify and clone the differentially expressed genes in

rat odontoblasts by subtracting the common genes

present in both osteoblasts and odontoblasts/pulp cells.

The method is based on the suppression PCR effect of

long inverted terminal repeats. When attached to DNA

fragments, these repeats can selectively suppress am-

plification of undesirable sequences in PCR procedures

(Siebert et al., 1995; Chenchik et al., 1996a,b). This

method can overcome the problem of differences in

mRNA abundance by incorporating a hybridization

step that equalizes sequence abundance during the

course of subtraction by standard hybridization kinetics

(Diatchenko et al., 1996). The major advantages of the

method include the requirement of small quantities of

poly(A)+ RNA from tissues and the ability to enrich

for rare sequences (Diatchenko et al., 1996). Indeed, we

found that this suppression subtractive hybridization is

a powerful method for generating an odontoblast

cDNA library that is highly enriched for differentially

expressed genes. However, the presence of a high pro-

portion of low abundant cDNAs with high nucleotide

sequence identity to known genes indicates that the

R. Dey et al. / Archi6es of Oral Biology 46 (2001) 249–260 257

subtraction may not be very efficient as previously

suggested (Diatchenko et al., 1996), and consequently

the screening was laborious.

Using suppression subtractive hybridization, we were

able to identify 538 odontoblast-enriched cDNA clones.

Knowing that the dentine sialophosphoprotein gene is

unique to odontoblasts, we expected that numerous

dentine sialophosphoprotein clones might be included

among the 538 clones. To avoid unnecessarily laborious

sequencing of dentine sialophosphoprotein clones, we

attempted to identify and eliminate them using dot-blot

analysis before DNA sequencing. For this purpose, five

representative dentine sialophosphoprotein clones that

had demonstrated identical sequences to different re-

gions of the dentine sialophosphoprotein cDNAs were

used as probes. As expected, this additional procedure

allowed us to identify 326 dentine sialophosphoprotein

clones and reduced the number of clones to be se-

quenced. The identification of a total of 390 dentine

sialophosphoprotein clones (326 by blotting and 64 by

sequencing) out of the 538 odontoblast-enriched cDNA

clones also provided strong evidence for the high degree

of selectivity of the suppression subtractive hybridiza-

tion technique in generating differentially expressed

cDNAs.

Interestingly, many clones that occurred multiple

times in this study represented those encoding extracel-

lular matrix proteins as observed in a rat incisor cDNA

library (Matsuki et al., 1995). These include cDNAs for

dentine sialophosphoprotein, osteocalcin, and type I

collagen. In addition, we found several cDNAs for lysyl

oxidase, bone sialoprotein and osteonectin that are

known to be highly enriched in both odontoblasts and

osteoblasts (Butler, 1998). The appearance of these

cDNAs indicates that the subtraction procedure may

not be effectively achieved, particularly for the highly

enriched genes. To eliminate common sequences, the

addition of excess driver cDNAs to the first hybridiza-

tion samples was recommended in the second hy-

bridization step during subtractive hybridization

(Diatchenko et al., 1996). In this study, we initially used

2 mg driver mRNA each from cultured dental papilla

mesenchymal cells and calvarial osteoblastic cells for

the subtraction procedure according to the manufac-

ture’s procedures. We speculate that the amount of

driver mRNA used may not be enough and the applica-

tion of a smaller amount of driver mRNA sample from

osteoblasts may have resulted in the incomplete sub-

traction. Fortunately, only those genes highly enriched

in both cell types were found to appear repeatedly.

Furthermore, we observed ribosomal protein L5

pseudogene and 28S ribosomal RNA that may be

present in various cell types. The reason for the appear-

ance of non-translated clones such as clone OD-330

(ribosomal protein L5 pseudogene) in the library is

unclear. However, our nucleotide sequence analysis of

this particular clone revealed the presence of long se-

quences of As. We speculate that this unique sequence

somehow led to initial recognition by the oligo (dT)

primer and subsequent amplification.

An intriguing finding was the appearance of an amel-

ogenin clone once. As the amelogenin gene is one of

ameloblast-specific genes, this result raises the possibil-

ity that the tester RNA sample (odontoblasts/pulp

cells) was contaminated partially with ameloblast RNA

in spite of our efforts to eliminate this problem. How-

ever, this contamination was not a serious problem, as

no other known ameloblast-specific gene was observed.

One interesting observation was the appearance of

the rat/human PEX gene as the second highest enriched

clone (39 clones) in a rat odontoblast subtraction

cDNA library. The PEX gene belongs to the neutral

endopeptidase family, which is composed of type II

integral membrane glycoproteins. These include neutral

endopeptidase (neprilysin), endothelin-converting en-

zymes and the erythrocyte surface protein and Kell

(Turner and Tanzawa, 1997). Specific defects in the

PEX gene are reportedly linked to human hypophos-

phataemic rickets that are characterized by the abnor-

mal development of bone and dentine (Nikiforuk and

Fraser, 1979; Abe et al., 1988; Shields et al., 1990). A

recent in situ hybridization study confirmed the expres-

sion of PEX mRNA in both osteoblasts and odonto-

blasts of the developing mouse (Ruchon et al., 1998),

although its role in both osteogenesis and odontogene-

sis remains unknown. It is of great interest to know

Table 3

Clones detected once and showing intermediate homology to known genes (scores 100–225)a

SpeciesClone Gene Length sent Score

OD-93 HGenomic DNA of 8p21.3 677 216

634OD-280 Translation Initiation factor 186 H

Major histocompatibility complex class I regionOD-313 316 M126

LOD-442 28S rRNA 508 182

204450 MSYTOD-1091

MOD-1247 F2 clone (C57BL/10XC3H) 147524

a Species: H, Homo sapiens; L, Latimeria chalumnae; M, Mus musculus.

R. Dey et al. / Archi6es of Oral Biology 46 (2001) 249–260258

Table 4

Clones detected once and showing limited regions of homolgy to known genes (scores B100)a

Species Accession LengthGene % IdentitybClone

H gb�U80456 27OD-106 100Transcription factor SIM2 short form

H dbj/AB019565Placental villi mRNA clone SMAP52 21OD-320 100

H gb�U29343OD-325 35Hyaluronan receptor (RHAMM) 94

H NM002132/HMMRHyaluronan-mediated motility receptor (HMMR) 25OD-348 100

Placental villi mRNA clone SMAP8526OD-361 H dbj/ab019570 27 100

H gb/AC004549BAC clone RG459N13 from 7p15 57OD-450 87

H gb/AC000403OD-551 53Genomic sequence Human 13 93

H dbj/AB019572Placental villi mRNA clone SMAP26 26OD-654 100

Chromosome II BAC F14B2OD-805 A gb/AC004450 27 96

H dbj/AB020860Genomic DNA of 8p21.3-p22 79OD-925 91

Clone an01-h01OD-1062 M gb�AF120321.1 20 100

20 of 103 of the complete genomeOD-1064 C gb�AF017104 20 100

H dbj�AB019568 27 100OD-1135 Placental villi mRNA clone SMAP8

a Species: A, Arabidopsis thaliana; C, Chlamydia pneumoniae; H, Homo sapiens; M, Mus musculus.b Quality of match is given as % identity in base pairs for the stretches of highest nucleotide match.

why the PEX clone appeared so many times, particu-

larly in a rat odontoblast subtraction cDNA library. To

understand the regulatory mechanism for the PEX ex-

pression, the sequencing of the gene in its entirety is

under investigation.

Despite the appearance of several clones in the rat

odontoblast cDNA library that are also expressed by

osteoblasts, our compilation of 390 dentine sialophos-

phoprotein clones out of 538 clones strongly suggests

that this technique was effective in identifying potential

odontoblast-specific/enriched cDNAs. Thus, there is a

good possibility that our approach, using the driver

samples from both cultured dental papilla cells and

calvarial osteoblasts, increases the chance of discover-

ing additional odontoblast-specific genes. The identifi-

cation of cDNA clones that appeared many times and

demonstrated low identity with previously reported

genes is indicative of potentially novel odontoblast-spe-

cific genes. This notion was further supported by the

results from our Northern blots. mRNAs for clones

OD-314 and OD-654 were expressed specifically in

odontoblasts/pulp cells, but not in other tissues includ-

ing bone, strongly suggesting that these clones represent

potentially new odontoblast-specific genes. In addition,

the presence of multiple bands in Northern blots sug-

gested that these clones may have multiple members of

each gene family or produce alternative spliced gene

products.

We speculate that novel genes may include a variety

of proteins, such as extracellular matrix proteins, en-

zymes, transmembrane receptors, growth factors,

proteins involved in signal-transduction pathways, and

transcription factors that may have important roles in

odontoblast differentiation and dentinogenesis. These

genes can be identified and characterized by full-length

sequencing of the cDNAs, recombinant protein produc-

tion, RT-PCR, in situ hybridization analysis of their

expression and location in tissues, and immunohisto-

chemical localization of their products using antibodies.

More importantly, functional analysis of the gene prod-

ucts could reveal the mechanisms responsible for odon-

toblast differentiation and dentinogenesis.

Acknowledgements

We thank Dr Jaro Sodek for his critical reading of

the manuscript. This study was supported in part by

USPHS grant DE 07034 and DE 4849.

Table 5

Clones detected more than once and showing low homology to

known genes (scores B100)a

Gene Species Frequency

S 9bMetallopanstimulin

7bDPutative MtN3-like protein

2bHPlacental villi SMAP41

a Species: S, Strongyloides ratti; H, Homo sapiens; D, Di-

anthus caryophyllus.b Some of the clones have been confirmed to be specific to

odontoblasts/pulp cells by Northern blot analyses (unpub-

lished data).

R. Dey et al. / Archi6es of Oral Biology 46 (2001) 249–260 259

Fig. 7. Northern blot analysis of mRNA expression in various

rat tissues using cDNAs derived from clones OD-314 and

OD-654. For agarose gel electrophoresis, 10 mg total RNA (A

and B) or 1 mg mRNA (C) was loaded in each lane. Radioau-

tographs for clones OD-314 (A and C) and OD-654 (B)

revealed their expression specifically in odontoblasts/pulp cells

(O/P), but not in other tissues such as brain (B), calvaria (Ca),

heart (H), kidney (K), liver (L), lung (Lu), and skeletal muscle

(SM). Their expression appeared in multiple bands (A and B).

Expression of G3PDH and b-actin, which are well-known

house-keeping genes, was observed in all the tissues examined.

Diatchenko, L., Lau, Y.F., Campbell, A.P., Chenchik, A.,

Moqadam, F., Huang, B., Lukyanov, S., Lukyanov, K.,

Gurskaya, N., Sverdlov, E.D., Siebert, P.D., 1996. Sup-

pression subtractive hybridization: a method for generating

differentially regulated or tissue-specific cDNA probes and

libraries. Proc. Natl. Acad. Sci. 93, 6025–6030.

Dickson, I.R., Dimuzio, M.T., Volpin, D., Anantha-

narayanan, S., Veis, A., 1975. The extraction of phospho-

proteins from bovine dentin. Calcif. Tissue Res. 19, 51–61.

Ducy, P., Karsenty, G., 1998. Genetic control of cell differen-

tiation in the skeleton. Curr. Opin. Cell. Biol. 10, 614–619.

Duguid, J.R., Dinauer, M.C., 1990. Library subtraction of in

vitro cDNA libraries to identify differentially expressed

genes in scrapie infection. Nucleic Acids Res. 18, 2789–

2792.

Friedenstein, A.J., 1976. Precursor cells of mechanocytes. Int.

Rev. Cytol. 47, 327–359.

George, A., Srinivasan, R., Thotakura, S., Veis, A., 1998. The

phosphophoryn gene family: identical domain structures at

the carboxyl end. Eur. J. Oral. Sci. 106 (Suppl. 1), 221–

226.

Hara, E., Kato, T., Nakada, S., Sekiya, S., Oda, K., 1991.

Subtractive cDNA cloning using oligo(dT)30-latex and

PCR: Isolation of cDNA clones specific to undifferentiated

human embryonal carcinoma cells. Nucleic Acids Res. 19,

7097–7104.

Lee, S.W., Tomasetto, C., Sager, R., 1991. Positive selection of

candidate tumor-suppressor genes by subtractive hy-

bridization. Proc. Natl. Acad. Sci. USA 88, 2825–2829.

Lian, J.B., Stein, G.S., 1992. Concepts of osteoblast growth

and differentiation: Basis for modulation of bone cell

development and tissue formation. Crit. Rev. Oral Biol.

Med. 3, 269–305.

Lian, J.B., Stein, G.S., Stein, J.L., van Wijnen, A.J., 1998.

Osteocalcin gene promoter: unlocking the secrets for regu-

lation of osteoblast growth and differentiation. J. Cell

Biochem. Suppl. 30–31, 62–72 Suppl.

Liang, P., Pardee, A., 1992. Differential display of eukaryotic

messenger RNA by means of the polymerase chain reac-

tion. Science 257, 967–970.

MacDougall, M., Simmons, D., Luan, X., Nydegger, J., Feng,

J., Gu, T.T., 1997. Dentin phosphoprotein and dentin

sialoprotein are cleavage products expressed from a single

transcript coded by a gene on human chromosome 4.

Dentin phosphoprotein DNA sequence determination. J.

Biol. Chem. 272, 835–842.

Matsuki, Y., Nakashima, M., Amizuka, N., Warshawsky, H.,

Goltzman, D., Yamada, K.M., Yamada, Y., 1995. A com-

pilation of partial sequences of randomly selected cDNA

clones from the rat incisor. J. Dent. Res. 74, 307–312.

Nacht, M., Ferguson, A.T., Zhang, W., Petroziello, J.M.,

Cook, B.P., Gao, Y.H., Maguire, S., Riley, D., Coppola,

G., Landes, G.M., Madden, S.L., Sukumar, S., 1999.

Combining serial analysis of gene expression and array

technologies to identify genes differentially expressed in

breast cancer. Cancer Res. 59, 5464–5470.

Nikiforuk, G., Fraser, D., 1979. Chemical determinants of

enamel hypoplasia in children with disorders of calcium

and phosphate homeostasis. J. Dent. Res. 58(Spec Issue B),

1014–1015.

Owen, M., Friedenstein, A.J., 1988. Stromal stem cells: mar-

References

Abe, K., Ooshima, T., Lily, T.S., Yasufuku, Y., Sobue, S.,

1988. Structural deformities of deciduous teeth in patients

with hypophosphatemic vitamin D-resistant rickets. Oral

Surg. Oral Med. Oral Pathol. 65, 191–198.

Altschul, S.F., Madden, T.L., Schaffer, A.A., Zhang, J.,

Zhang, Z., Miller, W., Lipman, D.J., 1997. Gapped

BLAST and PSI-BLAST: a new generation of protein

database search programs. Nucleic Acids Res. 25, 3389–

3402.

Butler, W.T., 1998. Dentin matrix proteins. Eur. J. Oral. Sci.

106 (suppl 1), 204–210.

Butler, W.T., 1987. Dentin-specific proteins. Methods Enzy-

mol. 145, 290–303.

Chen, J., Shapiro, H.S., Sodek, J., 1992. Development expres-

sion of bone sialoprotein mRNA in rat mineralized con-

nective tissues. J. Bone Miner. Res. 7, 987–997.

Chenchik, A., Diatchenko, L., Moqadam, F., Tarabykin, V.,

Lukyanov, S., Siebert, P.D., 1996a. Full-length cDNA

cloning and determination of mRNA 5% and 3% ends by

amplification of adaptor-ligated cDNA. BioTechniques 21,

526–534.

Chenchik, A., Moqadam, F., Siebert, P.D., 1996b. A new

method for full-length cDNA cloning by PCR. In: Kreig,

P.A. (Ed.), A Laboratory Guide to RNA Isolation, Analy-

sis and Synthesis. Wiley-Liss Inc., NY, pp. 273–321.

Chien, H.H., Lin, W.L., Cho, M.I., 1999. Interleukin-1b in-

duced release of matrix protein into culture media causes

inhibition of mineralization of nodules formed by peri-

odontal ligaments cells in vitro. Calcif. Tissue Int. 64,

402–413.

R. Dey et al. / Archi6es of Oral Biology 46 (2001) 249–260260

row-derived osteogenic precursors. Ciba Foundation Sym-

posium 136, 42–60.

Ritchie, H.H., Ritchie, D.G., Wang, L.H., 1998. Six decades

of dentinogenesis research. Historical and prospective

views on phosphophoryn and dentin sialoprotein. Eur. J.

Oral Sci. 106 (Suppl. 1), 211–220.

Ruchon, A.F., Marcinkiewicz, M., Siegfried, G., Tenenhouse,

H.S., DesGroseillers, L., Crine, P., Boileau, G., 1998. Pex

mRNA is localized in developing mouse osteoblasts and

odontoblasts. J. Histochem. Cytochem. 46, 459–468.

Shields, E.D., Scriver, C.R., Reade, T., Fujiwara, T.M., Mor-

gan, K., Ciampi, A., Schwartz, S., 1990. X-linked hy-

pophosphatemia: the mutant gene is expressed in teeth as

well as in kidney. Am. J. Hum. Genet. 46, 434–442.

Siebert, P.D., Chenchik, A., Kellogg, D.E., Lukyanov, K.A.,

Lukynov, S.A., 1995. An improved method for walking in

uncloned genomic DNA. Nucleic Acids Res. 23, 1087–

1088.

Ten Cate, A.R., 1998. Dentinogenesis. In: Ten Cate, A.R.

(Ed.), Oral Histology: Development, Structure and Func-

tion. Mosby-Year Book, Inc., St. Louis, Missouri, pp.

128–149.

Turner, A.J., Tanzawa, K., 1997. Mammalian membrane

metallopeptidases: NEP, ECE, KELL, and PEX. FASEB

J. 5, 355–364.

Veis, A., Perry, A., 1967. The phosphoprotein of the dentin

matrix. Biochemistry 6, 2409–2416.

Velculescu, V.E., Zhang, L., Vogelstein, B., Kinzler, K.W.,

1995. Serial analysis of gene expression. Science 270, 484–

487.

.