SEARCH

SEARCH BY CITATION

Keywords:

  • mouse;
  • tooth development;
  • microarray;
  • gene expression;
  • mRNA in situ hybridization

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

To identify genes heretofore undiscovered as critical players in the biogenesis of teeth, we have used microarray gene expression analysis of the developing mouse molar tooth (DMT) between postnatal day (P) 1 and P10 to identify genes differentially expressed when compared with 16 control tissues. Of the top 100 genes exhibiting increased expression in the DMT, 29 were found to have been previously associated with tooth development. Differential expression of the remaining 71 genes not previously associated with tooth development was confirmed by quantitative reverse transcription-polymerase chain reaction analysis. Further analysis of seven of the latter genes by mRNA in situ hybridization found that five were specific to the developing tooth in the craniofacial region (Rspo4, Papln, Amtn, Gja1, Maf). Of the remaining two, one was found to be more widely expressed (Sp7) and the other was found to be specific to the nasal serous gland, which is close to, but distinct from, the developing tooth (Vrm). Developmental Dynamics 236:2245–2257, 2007. © 2007 Wiley-Liss, Inc.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

The developing dentition is a paradigm for organogenesis, with morphogenesis and patterning (position, shape, and size), cell differentiation (enamel, dentin, cementum, bone), and cell function (mineralization, homeostasis, and regeneration) all having essential roles for its correct formation. Numerous genes are already known to function during the early stages of tooth development (Chai and Maxson, 2006) through both the genetic analysis of families presenting with a dental phenotype, such as PAX9 and third molar agenesis (Stockton et al., 2000), as well as direct investigation of the developing tooth. However, many of the genes involved in the latter stages of amelogenesis, root formation, and tooth eruption remain unknown (Chai and Maxson, 2006). Tooth development in non–human-like rodents is largely the same as that in humans (Peterkova et al., 2006). The mouse is, therefore, an excellent model system within which to investigate the molecular mechanisms of tooth development.

Tooth development involves the reciprocal interactive events of epithelial and mesenchymal tissues, which develop through the soft tissue phases: initiation, bud, cap, and bell stages (Chai and Maxson, 2006). This period is followed by the hard tissue formation stages of dentinogenesis and amelogenesis, or enamel formation. Tooth development involves the complex interplay of regulatory molecules that include transcription factors, growth factors, as well as cell surface and extracellular matrix molecules (Tucker and Sharpe, 2004). Many of these signaling molecules are also used recurrently in a variety of organs throughout vertebrate development and are remarkably well conserved throughout evolution (Peters and Balling, 1999; Pispa and Thesleff, 2003). Considerable insight into the identity and role of genes during the tooth development process has been gained from studies in both animal models and in humans (Thesleff and Nieminen, 1996; Stockton et al., 2000; Thesleff, 2003). There are currently more than 300 genes known to function during tooth development: 40 growth factors, 50 receptors, 105 signaling molecules, 57 transcription factors, 92 intracellular molecules, 114 extracellular molecules, and 81 plasma membrane proteins (genes may be categorized into more than one group; Thesleff, 2006). Despite many of these genes being used recurrently in the development of other organs, there are likely many additional genes that play a critical role in the tooth development process that are heretofore undiscovered.

To identify novel genes involved in tooth development for the purpose of furthering our knowledge of this developmental pathway and to provide a database of genes to aid future genetic dissection of dental phenotypes, we have investigated gene expression in the developing mouse molar tooth between postnatal days (P) 1 and P10. This period of tooth development primarily encompasses the amelogenesis stages where the enamel structures are formed, but also includes the early stages of root development (Chai and Maxson, 2006). Mouse molar teeth were chosen as their development is finite and, hence, more closely resembles that of human tooth development, unlike mouse incisor teeth, which grow continually. To this end, we carried out microarray gene expression, quantitative reverse transcription-polymerase chain reaction (qRT-PCR), and mRNA in situ hybridization analysis to identify genes that are specifically expressed in the developing tooth.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Microarray Analysis of Mouse Molar Tooth Gene Expression

To identify genes that function during mouse molar tooth development, the expression of the 34,000 mouse genes present on the Mouse Genome Expression 430 2.0 microarray (Affymetrix, Santa Clara, CA) were interrogated in a pool of mRNA from molar teeth extracted from Swiss Webster mouse pups between 1 and 10 days postnatal. These data were compared with a control set of 16 tissues obtained from the National Center for Biotechnology (NCBI) Gene Expression Omnibus (GEO) database (accession no. GSE1986), whose expression data were derived from tissue obtained from Swiss Webster mice between 2 and 4 weeks postnatal.

Analysis of the genes differentially expressed between the dental molar teeth (DMT) and control tissues (Fig. 1) found that the genes that exhibited increased expression in DMT were far more significant (10−9–10−23) than those that exhibited decreased expression (10−4–10−8). It was also apparent that more genes exhibited increased expression in the DMT over the controls (Fig. 1).

thumbnail image

Figure 1. Volcano plot showing differential expression between the dental molar teeth (DMT) and control tissues. A positive Log2 ratio between DMT and control tissues indicates increased expression of the gene in the DMT, a negative value indicates reduced expression.

Download figure to PowerPoint

Analysis of the top 100 genes that exhibited decreased expression in the DMT (Table 1) found that none had been previously associated with tooth development. However, analysis of the top 100 genes that exhibited increased expression in the DMT found 29 that had been previously associated with tooth development (Table 2). Eight of these genes have also been associated with anomalies in tooth development. All five genes previously identified as causing amelogenesis imperfecta, a disease characterized by enamel defects (Aldred et al., 2003), when disrupted, are present within these 29 genes: amelogenin (Amelx; Lagerstrom et al., 1991), kallikrein 4 (Klk4; Hart et al., 2004), matrix metalloproteinase 20 (MMP20; enamelysin; Ozdemir et al., 2005), enamelin (Enam; Rajpar et al., 2001), and distal-less homeobox 3 (Dlx3; Price et al., 1999). Other genes present included ameloblastin (Ambn), a gene that when inactivated in the mouse presents with a similar enamel phenotype (Paine et al., 2003), as well as the dentin sialophosphoprotein (Dspp) gene associated with dentinogenesis imperfecta (Zhang et al., 2001) and the paired-like homeodomain transcription factor 2 (Pitx2) associated with Rieger Syndrome (Semina et al., 1996).

Table 1. Top 100 Genes Exhibiting Decreased Expression in the DMT Over Control Samples
Gene symbolGene IDMAdj. P valueGene name
Cox1712856−1.631.29 × 10−08cytochrome c oxidase, subunit XVII assembly protein homolog (yeast)
Nat1170999−1.896.76 × 10−08N-acetyltransferase 11
Akap1056697−1.393.35 × 10−07A kinase (PRKA) anchor protein 10
Farslb23874−1.295.56 × 10−07phenylalanine-tRNA synthetase-like, beta subunit
Mtch256428−1.865.76 × 10−07mitochondrial carrier homolog 2 (Caenorhabditis elegans)
Edn313616−1.61.10 × 10−06endothelin 3
Slc25a28246696−1.421.38 × 10−06solute carrier family 25, member 28
Flcn216805−2.431.89 × 10−06Folliculin
Dtwd169185−1.292.57 × 10−06DTW domain containing 1
H2-D114964−3.654.30 × 10−06histocompatibility 2, D region locus 1
Actb11461−1.695.46 × 10−06actin, beta, cytoplasmic
Ccdc28a215814−2.575.56 × 10−06coiled-coil domain containing 28A
H2-L14980−3.495.82 × 10−06histocompatibility 2, D region
Nudt1367725−2.26.05 × 10−06nudix (nucleoside diphosphate linked moiety X)-type motif 13
Slc12a6107723−2.596.08 × 10−06solute carrier family 12, member 6
Frap156717−1.286.23 × 10−06FK506 binding protein 12-rapamycin associated protein 1
Gm71207965−1.68.30 × 10−06gene model 71, (NCBI)
Idh2269951−1.578.77 × 10−06isocitrate dehydrogenase 2 (NADP+), mitochondrial
Ly6e17069−3.419.22 × 10−06lymphocyte antigen 6 complex, locus E
Reps119707−1.649.89 × 10−06RalBP1 associated Eps domain containing protein
Nagk56174−1.111.05 × 10−05N-acetylglucosamine kinase
D4Wsu53e27981−2.481.25 × 10−05DNA segment, Chr 4, Wayne State University 53, expressed
Nup54269113−1.691.52 × 10−05nucleoporin 54
Rabif98710−1.871.68 × 10−05RAB interacting factor
Rtp467775−2.532.09 × 10−05receptor transporter protein 4
Nup98269966−1.422.10 × 10−05nucleoporin 98
Bcas268183−1.82.16 × 10−05breast carcinoma amplified sequence 2
Impa155980−2.412.18 × 10−05inositol (myo)-1(or 4)-monophosphatase 1
ORF553858−1.282.29 × 10−05open reading frame 5
March6223455−1.082.31 × 10−05membrane-associated ring finger (C3HC4) 6
6330407J23Rik67412−0.772.32 × 10−05RIKEN cDNA 6330407J23 gene
Dvl113542−1.992.79 × 10−05dishevelled, dsh homolog 1 (Drosophila)
0610012G03Rik106264−2.12.88 × 10−05RIKEN cDNA 0610012G03 gene
5230400G24Rik75734−2.522.90 × 10−05RIKEN cDNA 5230400G24 gene
AI506816433855−2.272.99 × 10−05expressed sequence AI506816
Fbxo2271999−1.453.15 × 10−05F-box only protein 22
Mat2a232087−1.953.42 × 10−05methionine adenosyltransferase II, alpha
Cox5a12858−1.283.46 × 10−05cytochrome c oxidase, subunit Va
Ncor120185−1.283.73 × 10−05nuclear receptor co-repressor 1
BC005624227707−1.284.21 × 10−05cDNA sequence BC005624
1190005F20Rik98685−1.584.65 × 10−05RIKEN cDNA 1190005F20 gene
Herpud280517−1.334.65 × 10−05HERPUD family member 2
Atad3a108888−1.995.46 × 10−05ATPase family, AAA domain containing 3A
Ngrn83485−0.935.59 × 10−05neugrin, neurite outgrowth associated
Gtpbp469237−2.375.84 × 10−05GTP binding protein 4
Dtx3l209200−3.055.89 × 10−05deltex 3-like (Drosophila)
Saps352036−2.266.89 × 10−05SAPS domain family, member 3
Tatdn169694−1.126.91 × 10−05TatD DNase domain containing 1
Pcgf576073−2.37.19 × 10−05polycomb group ring finger 5
Metap175624−0.917.42 × 10−05methionyl aminopeptidase 1
5730536A07Rik68250−1.548.10 × 10−05RIKEN cDNA 5730536A07 gene
Pcm118536−0.728.26 × 10−05pericentriolar material 1
Stk22s122116−2.558.26 × 10−05serine/threonine kinase 22 substrate 1
Tmem1967226−1.098.42 × 10−05transmembrane protein 19
Lamp116783−1.418.84 × 10−05lysosomal membrane glycoprotein 1
Sepp120363−1.71.03 × 10−04selenoprotein P, plasma, 1
Sod120655−1.551.04 × 10−04superoxide dismutase 1, soluble
H2-K114972−3.991.07 × 10−04histocompatibility 2, K1, K region
6030443O07Rik226151−0.831.07 × 10−04RIKEN cDNA 6030443O07 gene
2310051F07Rik108745−1.721.09 × 10−04RIKEN cDNA 2310051F07 gene
Rrs159014−1.451.11 × 10−04RRS1 ribosome biogenesis regulator homolog (Saccharomyces cerevisiae)
Camta1100072−0.851.20 × 10−04calmodulin binding transcription activator 1
Stim2116873−1.081.20 × 10−04stromal interaction molecule 2
Pknox118771−1.341.27 × 10−04Pbx/knotted 1 homeobox
Prosc114863−1.921.33 × 10−04proline synthetase co-transcribed
9130023D20Rik268706−0.971.38 × 10−04RIKEN cDNA 9130023D20 gene
Mrps969527−1.331.43 × 10−04mitochondrial ribosomal protein S9
Cox6c12864−1.411.45 × 10−04cytochrome c oxidase, subunit Vic
Eif3s254709−1.371.46 × 10−04eukaryotic translation initiation factor 3, subunit 2 (beta)
5730596B20Rik77580−0.721.51 × 10−04RIKEN cDNA 5730596B20 gene
Phf17269424−2.021.53 × 10−04PHD finger protein 17
Chchd366075−1.091.54 × 10−04coiled-coil-helix-coiled-coil-helix domain containing 3
Tob122057−2.261.54 × 10−04transducer of ErbB-2.1
Tfb1m224481−1.481.60 × 10−04transcription factor B1, mitochondrial
3300001M20Rik66926−1.21.67 × 10−04RIKEN cDNA 3300001M20 gene
Tmem6296957−1.531.68 × 10−04transmembrane protein 62
4930431P19Rik73886−0.661.83 × 10−04RIKEN cDNA 4930431P19 gene
A130007E14NA−1.431.93 × 10−04RIKEN cDNA A130007E14 gene
Gdap1014546−1.691.99 × 10−04ganglioside-induced differentiation-associated-protein 10
Setd1b208043−1.412.01 × 10−04SET domain containing 1B
Exosc766446−1.092.01 × 10−04exosome component 7
Prdx118477−1.222.10 × 10−04peroxiredoxin 1
Cdv3321022−0.832.17 × 10−04carnitine deficiency-associated gene expressed in ventricle 3
5330406M23Rik76671−2.962.18 × 10−04RIKEN cDNA 5330406M23 gene
BC03391570661−2.012.19 × 10−04cDNA sequence BC033915
Polr3a218832−1.872.25 × 10−04polymerase (RNA) III (DNA directed) polypeptide A
D17Wsu92e224647−0.812.27 × 10−04DNA segment, Chr 17, Wayne State University 92, expressed
Ing566262−0.972.28 × 10−04inhibitor of growth family, member 5
A230062G08Rik231326−1.562.34 × 10−04RIKEN cDNA A230062G08 gene
7630402D21NA−0.792.38 × 10−04RIKEN cDNA 7630402D21 gene
Sympk68188−0.882.42 × 10−04Symplekin
Hsf2bp74377−0.552.51 × 10−04heat shock transcription factor 2 binding protein
Crebl2232430−2.182.53 × 10−04cAMP responsive element binding protein-like 2
AW320013103448−2.182.54 × 10−04expressed sequence AW320013
Irak2108960−1.372.58 × 10−04interleukin-1 receptor-associated kinase 2
H2-Q815019−4.822.64 × 10−04histocompatibility 2, Q region locus 8
Serbp166870−1.322.67 × 10−04Serpine1 mRNA binding protein 1
H3056E02NA−0.52.72 × 10−04RIKEN cDNA H3056E02 gene
Als2cr2227154−2.382.85 × 10−04amyotrophic lateral sclerosis 2 (juvenile) chromosome region, candidate 2 (human)
5730589K01Rik268741−1.182.87 × 10−04RIKEN cDNA 5730589K01 gene
Table 2. Top 100 Genes Exhibiting Increased Expression in the DMT Over Control Samplesa
Gene symbolGene IDMAdj. P valueGene name
  • a

    Known tooth development genes are highlighted in boldface type.

Ambn116989.193.49 × 1023Ameloblastin
Amelx117047.65.15 × 1020amelogenin X chromosome
Dspp135177.139.95 × 1020dentin sialophosphoprotein
Mmp20308007.122.41 × 1019matrix metallopeptidase 20 (enamelysin)
Enam138017.434.57 × 1018Enamelin
Dmp1134066.642.02 × 1017dentin matrix protein 1
Sp6833955.822.87 × 1017trans-acting transcription factor 6
Gal3st43302175.852.89 × 10−17galactose-3-O-sulfotransferase 4
Klk4566405.299.37 × 1017kallikrein 4 (prostase, enamel matrix, prostate)
Dlx5133954.723.34 × 1016distal-less homeobox 5
Dsp1096203.984.86 × 10−16Desmoplakin
Col12a1128163.71.21 × 1015procollagen, type XII, alpha 1
Amtn714214.631.63 × 1015RIKEN cDNA 5430427O21 gene
Ibsp158916.693.10 × 1015integrin binding sialoprotein
Slc13a52378313.984.83 × 10−15solute carrier family 13 (sodium-dependent citrate transporter), member 5
Csn3129946.357.55 × 10−15casein kappa
Gja1146096.628.52 × 1015gap junction membrane channel protein alpha 1
Igsf3789083.295.90 × 10−14immunoglobulin superfamily, member 3
Panx32080983.48.14 × 10−14pannexin 3
Msi2766264.62.13 × 10−13Musashi homolog 2 (Drosophila)
Mmp13173864.583.96 × 1013matrix metallopeptidase 13
Dkk1133803.014.60 × 1013dickkopf homolog 1 (Xenopus laevis)
Sox11206663.154.73 × 10−13SRY-box containing gene 11
C230013L11Rik3197123.727.11 × 10−13RIKEN cDNA C230013L11 gene
Papln1707215.317.11 × 10−13papilin, proteoglycan-like sulfated glycoprotein
Hapln1129502.38.08 × 10−13hyaluronan and proteoglycan link protein 1
Mrc2175343.188.08 × 10−13mannose receptor, C type 2
Sp71705743.118.08 × 10−13trans-acting transcription factor 7
Phex186752.341.24 × 10−12phosphate regulating gene with homologies to endopeptidases on the X chromosome (hypophosphatemia, vitamin D resistant rickets)
Susd53821113.961.47 × 10−12sushi domain containing 5
Wnt10a224093.62.01 × 1012wingless related MMTV integration site 10a
Man1a2171562.382.34 × 10−12mannosidase, alpha, class 1A, member 2
Cspg2130032.933.37 × 10−12chondroitin sulfate proteoglycan 2
Sema5a203562.543.62 × 10−12sema domain, seven thrombospondin repeats (type 1 and type 1-like), transmembrane domain (TM) and short cytoplasmic domain (semaphorin) 5A
Trp63220612.763.85 × 1012transformation related protein 63
Dlx4133942.94.60 × 1012distal-less homeobox 4
Dlx2133924.34.93 × 1012distal-less homeobox 2
Pcdh18731732.644.93 × 10−12protocadherin 18
Wnt6224203.065.98 × 1012wingless-related MMTV integration site 6
5430413K10Rik714251.827.37 × 10−12RIKEN cDNA 5430413K10 gene
Maf171323.658.51 × 10−12avian musculoaponeurotic fibrosarcoma (v-maf) AS42 oncogene homolog
Vangl2938403.368.51 × 10−12vang-like 2 (van gogh, Drosophila)
Timp2218582.831.43 × 1011tissue inhibitor of metalloproteinase 2
Twsg1659603.361.47 × 10−11twisted gastrulation homolog 1 (Drosophila)
Wif1241176.321.47 × 10−11Wnt inhibitory factor 1
Cxadr130522.282.14 × 10−11coxsackievirus and adenovirus receptor
Msx2177024.142.14 × 1011homeo box, msh-like 2
Rnd3741944.052.30 × 10−11Rho family GTPase 3
Igbp1185183.522.60 × 10−11immunoglobulin (CD79A) binding protein 1
Npnt1142492.773.10 × 10−11Nephronectin
Col11a2128154.43.11 × 10−11procollagen, type XI, alpha 2
Edg3136103.063.30 × 10−11endothelial differentiation, sphingolipid G-protein-coupled receptor, 3
6430550H21Rik2453861.994.13 × 10−11RIKEN cDNA 6430550H21 gene
Chic1122121.974.27 × 10−11cysteine-rich hydrophobic domain 1
Ccdc50675011.976.02 × 10−11coiled-coil domain containing 50
Mmp14173874.386.18 × 1011matrix metallopeptidase 14 (membrane-inserted)
Nebl741032.196.18 × 10−11Nebulette
Nedd4179994.66.18 × 10−11neural precursor cell expressed, developmentally down-regulated gene 4
Ell137162.446.47 × 10−11elongation factor RNA polymerase II
Rspo42287702.646.82 × 10−11R-spondin family, member 4
E130306M17Rik3208252.158.34 × 10−11RIKEN cDNA E130306M17 gene
Nfatc4731812.69.72 × 10−11nuclear factor of activated T-cells, cytoplasmic, calcineurin-dependent 4
5730601F06Rik775192.511.14 × 10−10RIKEN cDNA 5730601F06 gene
D13Ertd787e526803.451.14 × 10−10DNA segment, Chr 13, ERATO Doi 787, expressed
Clec11a202562.861.21 × 10−10C-type lectin domain family 11, member a
2810417H13Rik680261.731.29 × 10−10RIKEN cDNA 2810417H13 gene
Shh204232.961.38 × 1010sonic hedgehog
Fscn1140862.821.42 × 10−10fascin homolog 1, actin bundling protein (Strongylocentrotus purpuratus)
Epha3138371.971.46 × 10−10Eph receptor A3
Msrb33201832.051.56 × 10−10methionine sulfoxide reductase B3
Fras12314704.51.58 × 10−10Fraser syndrome 1 homolog (human)
Rgs5197373.291.82 × 10−10regulator of G-protein signaling 5
Bmp7121623.221.87 × 1010bone morphogenetic protein 7
Cldn1127373.761.87 × 10−10claudin 1
Sema3d1081512.172.04 × 10−10sema domain, immunoglobulin domain (Ig), short basic domain, secreted, (semaphorin) 3D
Fjx1142212.562.30 × 10−10four jointed box 1 (Drosophila)
Pitx2187413.672.56 × 1010paired-like homeodomain transcription factor 2
AL033314564632.072.57 × 10−10expressed sequence AL033314
Slc20a2205162.592.59 × 10−10solute carrier family 20, member 2
D1Ertd471e278772.632.75 × 10−10DNA segment, Chr 1, ERATO Doi 471, expressed
Slc39a8675474.022.81 × 10−10solute carrier family 39 (metal ion transporter), member 8
Pard6g937372.843.14 × 10−10par-6 partitioning defective 6 homolog gamma (Caenorhabditis elegans)
Asxl32119611.943.15 × 10−10additional sex combs like 3 (Drosophila)
Dlx6os13200382.773.21 × 10−10Dlx6 opposite strand transcript 1
Dlx3133933.393.44 × 1010distal-less homeobox 3
Itgb6164202.394.09 × 10−10integrin beta 6
1200003M09Rik717182.194.66 × 10−10RIKEN cDNA 1200003M09 gene
Stx18711161.985.22 × 10−10syntaxin 18
Fzd6143682.26.70 × 10−10frizzled homolog 6 (Drosophila)
Atp6v0a1119752.686.73 × 1010ATPase, H+ transporting, lysosomal V0 subunit A1
Zfp64227222.446.73 × 10−10zinc finger protein 64
Sox4206773.457.43 × 1010SRY-box containing gene 4
Lrrc15744882.597.96 × 10−10leucine rich repeat containing 15
4632425D07Rik2082131.798.47 × 10−10RIKEN cDNA 4632425D07 gene
Frem22420224.148.49 × 10−10Fras1 related extracellular matrix protein 2
Slc13a52378312.789.84 × 10−10solute carrier family 13 (sodium-dependent citrate transporter), member 5
Nes180082.981.03 × 10−09Nestin
Grb10147831.891.10 × 10−09growth factor receptor bound protein 10
Igfbp51601121.10 × 10−09insulin-like growth factor binding protein 5
Krt17166671.751.12 × 10−09keratin 17

After the removal of known tooth development genes, 56 (of 71) functionally characterized genes were identified as exhibiting increased expression in the DMT (Table 2) compared with 78 (of 100) exhibiting decreased expression (Table 1). Additionally, there were 14 and 23 additional unknown genes, respectively. Bioinformatics analysis of these unknown genes found that very few possessed a known domain within their sequence (data not shown) and that they were distributed between the different cellular compartments, making it difficult to assign a putative function.

Web-based gene ontology analysis of the top 100 genes exhibiting increased expression in the DMT found a greater than expected number of genes conferring compression resistance in an extracellular matrix and in integrin binding. There were also a greater than expected number of genes functioning in osteoblast differentiation, regulation of cell differentiation and signal transduction, skeletal development, tissue remodeling, anatomical structure formation, mesoderm development, biomineral formation, organ and appendage morphogenesis, all of which are processes that would be expected to be important during tooth development. Interestingly, similar analysis of those genes exhibiting decreased expression in the DMT found that a greater than expected number of genes were involved in the catalysis of redox reactions that use heme groups as a hydrogen or electron donors for the generation of a proton electrochemical gradient across a membrane and that are located in the organelle inner and outer membrane and in the organelle or mitochondrial envelope.

qRT-PCR Confirmation of Microarray Results

On the premise that genes involved in the amelogenesis and tooth development processes would most likely exhibit an increased expression in our DMT mRNA when compared with other control mRNA samples, which is strongly supported by our analysis above, we concentrated on the top 100 genes that were identified as exhibiting a significantly increased expression level in the DMT over the control samples for confirmation of their expression by qRT-PCR analysis. The 71 genes not previously associated with tooth development that were identified in the top 100 genes identified by microarray analysis to show increased expression in the DMT, as well as three control genes known to function in tooth development (Amelx, Enam, Shh), were chosen for qRT-PCR analysis. Of these 74, expression levels of 68 were found to be in agreement with the microarray data, of which six were unknown genes (data not shown). Based on bioinformatics and knowledge about pathways or functions of related genes, seven genes were selected for further analysis: Rspo4 (R-spondin family, member 4; Cristin-4), Papln (papilin, proteoglycan-like sulfated glycoprotein), Vrm (Vomeromodulin precursor), Amtn (Amelotin), Gja1 (gap junction protein, alpha 1; Connexin 43), Sp7 (Sp7 transcription factor; osterix), and Maf (avian musculoaponeurotic fibrosarcoma [v-maf] homolog), and the results of their qRT-PCR can be found in Table 3.

Table 3. qRT-PCR Results for Rspo4, Papln, Amtn, Gja1, Maf, Vrm, and Sp7, and the Amelx and Enam Controlsa
GeneSampleCtSDΔCtSDΔΔCt
Rep. 1Rep. 2Rep. 3Average
  • a

    The ΔCt is the difference in cycle number between sample and negative data, and ΔΔCt is the Log2 fold difference between the DMT and either the Liver or Pool samples (ud = undetermined, nd = not determined). DMT, dental molar teeth; qRT-PCR, quantitative reverse transcription-polymerase chain reaction.

AmelxDMT14.8214.8214.7314.790.04−5.980.07 
 Liver33.6333.2633.5633.480.1613.140.16−19.13
 Pool33.4433.5633.2133.400.1513.500.16−19.48
 Negative34.9235.8334.9335.23    
EnamDMT17.8517.7117.8517.800.06−2.970.08 
 Liver36.0734.9334.9035.300.67ndnd
 Pool34.8633.6234.9134.460.73ndnd
 Negative32.4532.2732.0232.25    
Rspo4DMT23.4623.1123.7623.440.271.640.27 
 Liver34.9433.03ud33.990.9611.940.96−10.29
 Pool29.3229.6329.4929.480.128.850.28−7.21
 Negative33.6833.9333.9333.850.12   
PaplnDMT20.6820.7120.9020.760.100.380.12 
 Liver29.5729.5730.6529.930.519.840.52−9.47
 Pool27.2327.2226.8727.110.177.880.35−7.50
 Negative32.8532.7433.3532.98    
AmtnDMT25.9925.9124.8225.570.545.190.54 
 Liver33.5933.3633.7733.570.20ndnd
 Pool32.9832.3932.2432.530.39ndnd
 Negative33.2632.7932.7732.94    
Gja1DMT21.3021.2821.3721.320.04−0.480.04 
 Liver29.9129.5929.8329.780.147.730.16−8.21
 Pool25.8224.8124.9425.190.454.560.51−5.04
 Negative33.7733.6336.9534.7833.77   
MafDMT23.1023.1423.4723.240.171.440.17 
 Liver27.2926.9527.3627.200.185.150.19−3.71
 Pool26.6926.4126.6226.570.125.940.27−4.51
 Negative36.7335.4935.4135.8836.73   
VrmDMT39.7026.7626.5726.670.106.280.12 
 Liver45.2036.8937.1039.734.73ndnd
 Pool35.0536.2337.2236.171.09ndnd
 Negative36.5036.1236.5336.38    
Sp7DMT25.6225.4725.4225.510.084.740.10 
 Liver35.4537.3835.0235.951.0315.611.03−10.88
 Pool33.3433.3433.9533.540.2913.640.29−8.90
 Negative41.0736.9536.9538.32    

Verification of Tooth Specific Expression of Genes by mRNA In Situ Hybridization

Seven genes were selected based on the microarray and qRT-PCR results, and the pattern of their expression in the developing incisor (Fig. 2) and molar (Fig. 3) tooth was analyzed in coronal sections of the heads of newborn mice using mRNA in situ hybridization. The results identified five genes that are expressed specifically in tooth tissues in the craniofacial region, including one expressed in the stratum intermedia (Gja1; Figs. 2A, 3A), two in the ameloblasts (Maf and Papln; Figs. 2B, 3B, and 2C, 3C, respectively), one only in incisor epithelia and the basal lamina of molar, but not incisor, ameloblasts (Amtn; Figs. 2D, 3D), and one in the dental papilla (Rspo4; Figs. 2E, 3E). Of the remaining two genes, Sp7 was found not to be expressed specifically in tooth tissues and the remaining gene (Vrm) was found to be expressed in the nasal serous gland, which is located close to, but distinct from, the developing tooth (data not shown).

thumbnail image

Figure 2. AE: In situ hybridization of Gja1 (A), and Maf (B), Papln (C), Amtn (D), Rspo4 (E) in incisor tooth germs. “a” and “c” are hematoxylin and eosin–stained sections, and “b” and “d” are stained for mRNA in situ analysis. “a” and “b” are at ×10 magnification, and “c” and “d” are at ×40 magnification. The asterisk indicates the position of the ameloblasts (A–D) or odontoblasts (E). Arrows indicate the localization of gene expression.

Download figure to PowerPoint

thumbnail image

Figure 3. AE: In situ hybridization of Gja1 (A), and Maf (B), Papln (C), Amtn (D), Rspo4 (E) in molar tooth germs. “a” and “c” are hematoxylin and eosin–stained sections, and “b” and “d” are stained for mRNA in situ analysis. “a” and “b” are at ×10 magnification, and “c” and “d” are at ×40 magnification. The asterisk indicates the position of the ameloblasts (A–D) or odontoblasts (E). Arrows indicate the localization of gene expression.

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Genetic dissection of the tooth development process over the past two decades has identified signaling molecules and factors that partner with them to allow precise development of the tooth at the appropriate developmental stage, with the correct pattern, shape, and size. The availability of the complete genomic sequence for many vertebrate species including rodents and humans as well as their representative transcriptomes, has provided a platform for large-scale interrogation of genes in specific organs. To identify genes heretofore undiscovered as critical players in the biogenesis of teeth, we have used microarray gene expression analysis of the developing mouse molar tooth between P1 and P10 to identify genes differentially expressed when compared with 16 control tissues. Our aim was to identify tooth-specific developmental genes. As many genes involved in developmental processes in the tooth are likely required during the development of other tissues, we wanted to minimize the inclusion of these more general developmental genes in our list of candidate tooth-specific genes. These control tissues (see the Experimental Procedures section) were chosen to provide a wide cross-section of tissues with which to compare DMT to ensure that the genes exhibiting the greatest differential expression, and by inference genes most likely to be tooth-specific, would be identified ahead of those genes that function more generally in tissue development.

Analysis of the top 100 genes showing increased expression in the DMT compared with control samples found that 29 had been previously associated with tooth development. However, analysis of the top 100 genes found to exhibit decreased expression in the DMT identified no genes previously associated with tooth development. This finding is striking, given that the repression of genes and signaling pathways during developmental processes is likely as important as their activation. These genes exhibiting decreased expression may, therefore, offer novel gene families that are involved in the regulation of tooth development, but further research is required to clarify their role, if any, in tooth development.

The high expression levels of 68 of 71 genes, which were identified as exhibiting a significantly increased expression level in the DMT and which had not been previously associated with tooth development, were confirmed by qRT-PCR analysis, which included six previously uncharacterized genes, indicating that the combination of microarray and qRT-PCR analyses is a good screening method to identify tissue/organ specific genes. Only five of seven genes analyzed by mRNA in situ hybridization were tooth-specific, indicating that microarray and qRT-PCR analyses may not be accurate enough to characterize particular tissue/organ specific genes. Our results suggest that the combination of microarray, qRT-PCR, and mRNA in situ hybridization is a better strategy to characterize the global tissue/organ specific expression pattern and to characterize accurately the specificity and expressing cell type of each target gene.

The expression of Maf and Amtn was found to be appreciably lower in the molar (Fig. 3B and D, respectively) compared with incisor (Fig. 2B and D, respectively) tooth by mRNA in situ hybridization. However, their expression had been found to be appreciable by microarray and qRT-PCR analysis of the molar tooth. The microarray and qRT-PCR analysis was performed on molar teeth from mouse pups between P1 and P10 and at P2, respectively, and the mRNA in situ hybridization was performed on tooth germs from P0. Their lower expression in the molar compared with incisor tooth germs at P0, but their appreciable expression at later time points in the molar tooth, would suggest that their expression is increasing from P0 to P10 in the molar tooth and that they are expressed earlier in incisor tooth development.

Amtn is reported to be an ameloblast-specific gene (Iwasaki et al., 2005; Moffatt et al., 2006; Trueb et al., 2007) whose expression is restricted to maturation-stage ameloblasts in developing murine molars and incisors (Iwasaki et al., 2005; Moffatt et al., 2006). Its greatest expression is found in the basal lamina of the ameloblasts in both the incisors and molars (Moffatt et al., 2006) from which it is secreted and is believed to play a role in cell adhesion during the latter stages of dental enamel formation (Iwasaki et al., 2005; Moffatt et al., 2006). However, we have found that Amtn is localized only in the incisor epithelia and the basal lamina of molar ameloblasts at P0 stage. Much of the previous research has been performed at later stages of tooth development (>P5), and our result would suggest that the localization of Amtn changes during tooth development. The temporospatial expression pattern of Amtn during tooth development needs further examination.

Gja1 has been previously reported to be expressed in the cells of the stratum intermedium and the inner enamel epithelium (Pinero et al., 1994; Kagayama et al., 1995) where it is distributed exclusively at the sites of contact between odontoblasts, suggesting a function in cell to cell signaling (Fried et al., 1996). Mutations in Gja1 have been found to cause oculodentodigital dysplasia and Hallermann–Streiff syndrome, conditions that include dental anomalies (Paznekas et al., 2003; Pizzuti et al., 2004). We have also found Gja1 to be expressed in the stratum intermedium, supporting the previous observations and putative function.

Rspo4 is a member of the R-spondin (Rspo) protein family that has purported essential activities in vertebrate development and their ligand-type activities overlap with those of the canonical Wnt ligands where both are capable of the activation of beta-catenin (Kim et al., 2006). However, there are likely independent receptor/signaling pathways for Rspo proteins that intersect those of Wnt at the level of beta-catenin (Kim et al., 2006). Both Wnt signaling (Sarkar and Sharpe, 1999) and beta-catenin (Obara et al., 2006) function during tooth development. However, the localization of beta-catenin (Obara et al., 2006) is different from that which we identified for Rspo4, suggesting that their interaction may have a role in cell–cell signaling between the dental papilla and the inner dental epithelium during tooth development. Plans are under way to inactivate this gene in the mouse and will lead to important insights on the role of this gene in various developmental pathways.

Papln is an extracellular matrix glycoprotein with an N-terminal region, called the papilin cassette, which shows similarity to the C-terminal region of matrix-associated metalloproteinases (Kramerova et al., 2000). This finding has led to a suggested role in influencing cell rearrangements and the modulation of metalloproteinases during organogenesis (Kramerova et al., 2000). There are many known metalloproteinases that function during tooth development, and the localized expression of Papln in the ameloblasts identified in this study would support this suggested role.

Maf contains a C-terminal conserved basic region leucine zipper (bZIP) domain, which likely mediates its dimerization and DNA binding property in its role as a transcription factor. c-Maf has been reported to be involved in lens development (Reza and Yasuda, 2004), the control of chondrocyte fate and differentiation (Lefebvre and Smits, 2005), and in the differentiation of TH2-type lymphocytes (Lavender et al., 2000; Agnello et al., 2003). We have found it to be expressed in the ameloblasts of the developing mouse incisor and molar tooth, which would suggest a putative role in the transcriptional regulation of genes specific to the ameloblast during amelogenesis.

Sp7 is a zinc-finger transcription factor and a member of the SP gene family, which is highly expressed in proliferating epithelial cells of hair follicles, limbs, and teeth (Nakamura et al., 2004), and it has also been reported to be ubiquitously expressed in adult tissues (Scohy et al., 2000). This is in agreement with the more general localization found by mRNA in situ hybridization in this study. During tooth development, Sp7 is expressed in proliferating dental epithelium and differentiated odontoblasts, and it has been proposed that it may function in the regulation of cell growth (Nakamura et al., 2004). Other members of the SP family, such as Sp4 and Sp6, are also known to be involved in tooth development (Supp et al., 1996; Nakamura et al., 2004).

Vrm is a novel glycoprotein originally identified in the lateral nasal gland where it is highly concentrated in the mucus of the vomeronasal organ, which has led to the hypothesis that vomeromodulin participates in perireceptor events that facilitate the process of pheromone access and detection as a chemosensory stimulus transporter in olfactory transduction (Khew-Goodall et al., 1991; Krishna et al., 1994). In this study, Vrm was identified by mRNA in situ hybridization to be located in the nasal serous gland that is located close to, but distinct from, the developing tooth.

We compared the developing tooth structure (P1–P10) against developed control tissues (2–4 weeks), an approach that may at first glance appear to have significant limitations. Given that microarray analysis is amenable to sequential experimentation, we reasoned that we could conduct the analysis with the DMT initially and compare the data obtained with those already available in the public domain, thereby enabling an economical experimental test. Our results demonstrate that, despite that the tissues being compared were not obtained from age-matched animals, this approach was indeed valid given that we identified a large number of known tooth development genes as most significantly differentially expressed between the DMT and control samples. Further validation of our approach was obtained by qRT-PCR analysis of these genes in RNA extracted from six control tissues obtained from P2 mice which supported the microarray data analysis for almost all genes. The final and most credible validation was obtained by mRNA in situ hybridization analysis of a subset of novel genes that confirmed the expression of five to seven genes as being specifically expressed in the tooth in the craniofacial region, and one in the nasal serous gland that is located close to the tooth. The presence of this latter gene in our analysis can possibly be explained by the structure within which it is expressed having been attached to the tooth structures extracted from the P1–P10 mouse pups and, therefore, its inclusion in the mRNA analyzed in both the microarray and qRT-PCR analysis. Future analysis would include the use of age-matched control tissues to allow for an improved comparison that may identify additional genes. The use of specific time points rather than the pool of time points used here may also allow for additional unknown and/or tooth-specific genes to be identified that may have been diluted in this analysis. Further analysis of other genes showing differential expression between the tooth and control samples but to a lesser degree than the 100 concentrated on here may also identify additional tooth development genes.

We have shown that microarray gene expression analysis of the developing tooth is a viable method for the identification of novel genes underlying this developmental process. Importantly, this approach has identified novel genes of unknown function, genes related to those known to have a role in tooth development, and genes known to be expressed in other tissues but not known to be expressed at significant levels in the developing tooth.

EXPERIMENTAL PROCEDURES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Isolation of Mouse Molar Tooth RNA

Molar teeth were manually extracted from 60 Swiss Webster mouse pups between 1 and 10 days postnatal (USC IACUC Protocol No. 7225). The teeth were pooled and ground in the presence of liquid nitrogen using a pestle and mortar, and the RNA were extracted using the RNeasy RNA purification kit (Qiagen, Valencia, CA) following the manufacturer's recommended protocol.

Microarray Analysis of Gene Expression in Mouse Molar Teeth

The mRNA was labeled and hybridized to the Affymetrix Mouse Genome Expression (MGE) 430 2.0 microarrays following the manufacturer's recommended protocols and reagents (Affymetrix, San Clara, CA). Microarrays hybridized with mouse molar tooth RNA were analyzed in quadruplicate and the raw microarray data, both CEL and data files, are available online at the NCBI-GEO (http://www.ncbi.nlm.nih.gov/geo/; Edgar et al., 2002) database, accession number GSE7164. The DMT microarray data was compared with a control set of 16 tissues (brain, CD4+ T-cells, eye, heart, kidney, liver, lung, lymph node, ovary, placenta, smooth muscle, spleen, stomach, submaxillary gland, testis, and thymus) whose data were obtained from the NCBI-GEO database (accession no. GSE1986), which were also obtained from Swiss Webster mice, but between 2 and 4 weeks postnatal. The comparison was performed using the Bioconductor open-source analysis program run within the R-script environment (Gentleman et al., 2004) and the limma package (Smyth, 2004) to identify genes exhibiting differential expression between DMT and control tissues. Briefly, raw microarray data was normalized using the log-additive robust-multichip-average (RMA) algorithm (Irizarry et al., 2003). The data from the four DMT microarray replicates were compared with that of the 16 control tissues by fitting a linear model to the expression data for each probe. The resulting coefficients for each probe were then subjected to a pair-wise comparison between the “DMT” vs. “control” data and differential expression identified by calculating moderated t statistics and log-odds of differential expression by empirical Bayes shrinkage of the gene-wise sample variances toward a common value (Smyth, 2004).

Analysis of Gene Ontology and Uncharacterized Genes

The identification of putative domains within the uncharacterized genes was performed using the NCBI Conserved Domain Database (CDD) database (http://www.ncbi.nlm.nih.gov/Structure/cdd/wrpsb.cgi; Altschul et al., 1997). The predicted localization of the uncharacterized genes and the identification of sequence motifs that support this were identified using PSORT (Horton and Nakai, 1996, 1997) located on the National Institute for Basic Biology (NIBB) server (http://psort.nibb.ac.jp/). Gene ontology for each set of genes was investigated using the Web-based program WebGestalt (http://bioinfo.vanderbilt.edu/webgestalt/; Zhang et al., 2005).

qRT-PCR Analysis

Six tissues (heart, lung, kidney, liver, spleen, and brain) were collected from 2-day-old Swiss Webster mice and were immersed in liquid nitrogen immediately after dissection. The RNAs were prepared according to the Trizol protocol (Invitrogen, Carlsbad, CA). Equal amounts of RNAs from each of these six tissues were pooled and first-strand cDNA was prepared using the High Capacity cDNA Archive kit (Applied Biosystems, Foster City, CA). Ten micrograms of RNA from either the pooled tissues or the molar tooth RNA from P1–P10 mice were mixed with 10 μl of 10 × buffer, 4 μl of 25 × dNTP, 10 μl of Random Primers, 5 μl of MultiScribe Reverse Transcriptase (50 U/μl) and H2O to 100 μl. The primers for qRT-PCR were designed with Primer Express version 3.1 (Applied Biosystems, Foster City, CA) based on the cDNA sequences deposited in the NCBI Genbank database (Supplementary Table S1, which can be viewed at http://www.interscience.wiley.com/jpages/1058-8388/suppmat). qRT-PCR reactions were prepared in triplicates with 1 μl of a 50 times dilution of first-strand cDNA mixed with 5 μl of 2 × ABsolute QPCR SYBR Green ROX mix (ABgene, Rochester, NY), 1 μl of 700 nM of the primer, and 3 μl of H2O. Reactions underwent thermal cycling as follows: 1 cycle of 95°C for 4 min, followed by 50 cycles of 95°C for 1 min, and 55°C for 1 min, in an ABI 7900HT real-time thermal cycler (Applied Biosystems). Glyceraldehyde-3-phosphate dehydrogenase was used as an internal control to normalize for variation in the RNA concentrations.

mRNA In Situ Hybridization

Riboprobes for mRNA in situ hybridization were prepared using nested/half-nested PCR primers (Supplementary Table S2) designed with Primer3 (Rozen and Skaletsky, 2000) based on the cDNA sequences deposited in the NCBI GenBank database. The first PCR reaction was prepared using 1 μl of a 50 times dilution of the product from first-strand cDNA synthesis, 1 μl of 1.25 mM of dNTPs, 0.5 μl of 3.2 μM each primer, 2 μl of 10 × buffer, 1 unit of Taq DNA polymerase and H2O to 20 μl. The initial PCR reaction was performed as following: 1 cycle of 95°C for 5 min, followed by 10 cycles of 95°C for 45 sec, 65°C for 45 sec (reduced by 1 degree each subsequent cycle), and 72°C for 2 min, then 25 cycles of 95°C for 45 sec, 55°C for 45 sec, and 72°C for 2 min, and a final extension of 72°C for 8 min. The second PCR reaction consisted of 1 μl of a 10 times dilution of the initial PCR products, with the other PCR reagents and conditions as used for the initial PCR. One member of each pair of nested PCR primers included the Sp6 promoter sequence as a 5′ tag sequence so that the transcripts from the PCR products could be complemented with the sense strand of the gene. The nested PCR products were purified from a 1% agarose gel using the QIAquick Gel Extraction kit (Qiagen, Valencia, CA) according to the manufacturer's recommended protocol. The riboprobes were prepared with DIG RNA Labeling mix (Roche Applied Science, Indianapolis, IN) using Sp6 RNA polymerase according to the manufacturer's recommended protocol. The transcripts were examined by electrophoresis on a 1% agarose gel and diluted 2 to 20 times using water based on the intensity of bands.

Newborn mice were anesthetized using chloroform and killed immediately by decapitation. The heads were fixed in 4% paraformaldehyde at 4°C overnight, decalcified in 0.1 M ethylenediaminetetraacetic acid/2% paraformaldehyde/phosphate buffered saline for 1 week, embedded in paraffin, and sectioned. The 8-μm paraffin sections were rehydrated through successive incubations with xylene, ethanol at concentrations of 100%, 90%, 70%, and 50%, and PBS before treatment with 0.1 M of triethanolamine for 10 min and 0.1 M of triethanolamine with 0.25% of acetic anhydride for 10 min. Hybridization was performed in mRNA in situ hybridization solution (Dakocytomation, Carpinteria, CA) at 55°C. The sections were washed at 50°C in 2 × standard saline citrate (SSC)/50% formamide at 50°C for 1 hr and 1 × SSC/50% formamide at 50°C for 30 min, treated with 20 μg/ml of RNase A at 37°C for 30 min, and further washed with 2 × SSC at 55°C for 1 hr and 0.1 × SSC at 55°C for 2 hr. The sections were blocked with 10% lamb serum (Gibco; Invitrogen, Carlsbad, CA) at room temperature for 2 min and incubated with Anti-Digoxigenin-AP (Roche Applied Science; 1:1,000 dilution in 10% lamb serum) at 4°C overnight. The sections were treated with BCIP/NBT Liquid Substrate System (Sigma-Aldrich, St. Louis, MO) at room temperature overnight and counterstained with Methyl Green (Fluka; Sigma-Aldrich).

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

This research was supported by grant DE14102 (P.I.P.) from the National Institute of Dental and Craniofacial Research. M.L.S. is supported by DE06988. This research was conducted partly in a facility constructed with support from Research Facilities Improvement Program Grant No. C06 (RR10600-01, CA62528-01, RR14514-01) from the National Center for Research Resources, National Institutes of Health. Author contributions are as follows: T.J.P. conceived the study. P.I.P. and M.L.S. advised on strategy. M.L.S. harvested the mouse molar teeth. G.A.M. extracted the RNA. Y.H. conducted the microarray experiments under the supervision of R.M.-S. T.J.P. analyzed the microarray data. F.L. performed all RT-PCRs and qRT-PCRs, and in situ hybridizations (with guidance from S.O. in Y.C.'s laboratory on section preparation, and S.O. and P.B. in result interpretation). P.I.P. directed and provided funding for the studies.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

The Supplementary Material referred to in this article can be found at http://www.interscience.wiley.com/jpages/1058-8388/suppmat

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.