RAPID COMMUNICATION
Clinical
D.A. McKnight1, J.P. Simmer2, P.S. Hart3,
T.C. Hart4, and L.W. Fisher1*
1Craniofacial and Skeletal Diseases Branch, NIDCR, NIH,
DHHS, 9000 Rockville Pike, Bldg. 30, Room 228, Bethesda,
MD 20892, USA; 2Department of Biologic and Materials
Sciences, University of Michigan School of Dentistry,
Dental Research Lab, Ann Arbor, MI 48108, USA; 3Office
of the Clinical Director, NHGRI, NIH, DHHS, Bethesda,
MD 20892, USA; and 4Section of Human and Craniofacial
Genetics, NIDCR, NIH, DHHS, Bethesda, MD 20892,
USA; *corresponding author, lfisher@dir.nidcr.nih.gov
J Dent Res 87(12):1108-1111, 2008
Abstract
Dentinogenesis imperfecta (DGI) and dentin
dysplasia (DD) are allelic disorders due to
mutations in DSPP. Typically, the phenotype
breeds true within a family. Recently, two reports
showed that 3 different net -1 bp frameshift
mutations early in DSPP’s repeat domain caused
DD, whereas 6 more 39 frameshift mutations
were associated with DGI. Here we identify a DD
kindred with a novel -1 bp frameshift (c.3141delC)
that falls within the portion of the DSPP repeat
domain previously associated solely with the DGI
phenotype. This new frameshift mutation shows
that overlapping DSPP mutations can give rise to
either DGI or DD phenotypes. Furthermore, the
consistent kindred presentation of the DD or DGI
phenotype appears to be dependent on an as-yet-
undescribed genetic modifier closely linked to
DSPP.
Key words: DSPP, dentinogenesis imperfecta,
dentin dysplasia, DPP.
Received May 15, 2008; Last revision September 23, 2008;
Accepted September 24, 2008
1108 Downloaded from jdr.sagepub.com at The University of Iowa Libraries on June 19, 2015 For personal use only. No other uses without permission.
International and American Associations for Dental Research
Overlapping DSPP Mutations
Cause Dentin Dysplasia
and Dentinogenesis Imperfecta
Introduction
D SPP (dentin sialophosphoprotein) is the largest member of the SIBLING
(Small, Integrin-Binding LIgand, N-linked Glycoprotein) gene family,
due to the presence of ~250 tandem copies of a nominal 9-base-pair repeat
encoding the SerSerAsp tripeptide (Fisher et al., 2001; Bellahcene et al.,
2008). DSPP encodes a ~1300-amino-acid protein that is cleaved into
the two major non-collagenous proteins found in human dentin, dentin
sialoprotein (DSP) and the repeat-containing dentin phosphoprotein (or
phosphophoryn, DPP) (MacDougall et al., 1997; George et al., 1999).
Two classes of dentin disorders have been directly correlated with DSPP
mutations. Dentinogenesis imperfecta (DGI types II and III: OMIM 125490
and OMIM 125500) and dentin dysplasia (DD type II: OMIM 125420) are
dominantly inherited, non-syndromic diseases of dentin. Clinically, these
diseases are characterized by opalescent, amber-brown teeth with enamel
that has an increased propensity to fracture during mastication, due to the
defective underlying dentin, thus leaving the exposed dentin vulnerable
to rapid wear and caries. Both diseases are also often characterized by the
radiographic appearance of bulbous crowns and obliterated pulps. One
major clinical difference between DGI and DD is that the visible portions of
the crowns of the secondary teeth appear relatively normal in DD, although
thistle-shaped pulps, pulp stones, and pulpal obliteration may be observed
radiographically.
Despite these two diseases having some distinguishing clinical
characteristics, all mutations to date have been identified within the DSPP
gene. Mutations in the signal peptide have been shown to cause both DD and
DGI phenotypes (Rajpar et al., 2002; Malmgren et al., 2004). Recently, in
two DD and five DGI families in whom no mutations could be found within
the DSP portion of the DSPP gene, we identified frameshift mutations in the
~2.4-kb repeat domain of exon 5 (DPP) in all affected persons (McKnight
et al., 2008). These -1-bp or -4-bp (i.e., net -1 bp) frameshift mutations did
not result in premature truncation of the DSPP products, but immediately
changed DPP’s normally hydrophilic carboxy-terminal domain into one that
was hydrophobic. This class of mutations was proposed to cause a dominant-
negative disorder through disruption of the cell’s protein trafficking system.
Interestingly, the DD families had their frameshift mutations within the 59
portion of the repeat, resulting in the two longest hydrophobic domains, while
all of the persons with DGI had more 39 frameshifts with correspondingly
shorter hydrophobic carboxy termini. At the same time, Song et al. (2008)
reported related (net -1 bp) frameshift mutations in one DD and four DGI
families. The combination of the two results continued to suggest that DD
frameshifts were 59 to all DGI mutations in DPP.
In this paper, a new DD-causing -1-bp frameshift mutation is identified
within the hypothesized DGI-associated portion of DSPP. This new
mutation is discovered in a previously described four-generation DD family
in whom the mutation was unknown (Beattie et al., 2006).
J Dent Res 87(12) 2008 Overlapping DSPP Mutations Cause DGI/DD 1109

METHODS
Participant Information
The four-generation family
segregating autosomal-dominant
DD was previously linked to the
DSPP region of chromosome
4q21, but no specific mutation was
identified (Beattie et al., 2006).
Clinical findings for affected
family members included a
severe phenotype in the primary
dentition, with amber discoloration
and coronal translucency of the
crowns, as well as radiographic
appearance of bulbous crowns
with pulpal obliteration. The
secondary dentition exhibited a
much milder clinical presentation,
with only a slight grey Figure 1. Clinical phenotype and mutation results. (A) Radiograph of primary dentition. (B) Radiograph of
discoloration, and radiographic secondary dentition. (C) Photograph of secondary teeth for one affected individual (III-9) (Beattie et al., 2006).
Note that only the 4 mandibular incisors are in their natural state; the maxillary teeth have porcelain veneers.
evidence of thistle-shaped pulps
Representative samples of the sequenced cloned plasmids showing: (D) normal allele and (E) c.3141delC
and pulpal mineralization. One mutation analysis. (F) Kyte-Doolittle hydrophathy(AQ)plot (Kyte and Doolittle, 1982) of hydrophobicity was
affected family member (III-9) was generated for the wild-type DSPP (GenBank accession no. EU278640) and (G) the C.3141delC mutation
described as having a more severe (GenBank accession no. EU709728). Portions below the 0.0 y-axis value (bold yellow line) are hydrophilic,
DD phenotype in the secondary demonstrating that wild-type DSPP is normally hydrophilic, even while not accounting for the increased
hydrophilicity that the phosphate groups would add. Immediately after the -1-bp frameshift mutation, the protein
teeth, but not severe enough to be becomes highly hydrophobic (between +1.0 and +3.0) compared with the normal (-1.5) DSPP protein.
considered DGI. DNA samples
from one unaffected family
member (III-8) and two affected
family members (III-2 and III-9)
were obtained in accordance with study protocols approved by RESULTS
Institutional Review Boards at the University of Michigan and Complete sequence analysis of the DPP portion of the
the National Institutes of Health, and informed consents were DSPP gene was performed for the persons with DD and the
obtained. unaffected family member. The unaffected family member was
Sequencing of the DSPP Repeat (DPP portion) homozygotic for DSPP HVRR haplotype 1 (GenBank accession
The ~2.4-kb DPP and the 39 hypervariable ~1.2-kb (DSPP HVRR) no. EU278640) and had no frameshifts. Both persons with
repetitive portions of exon 5 were separately amplified by PCR, DD were heterozygotic, with a single normal DSPP HVRR
cloned into pCR4-TOPO plasmids (Invitrogen, Carlsbad, CA, USA) haplotype (III-9, haplotype 1; and III-2, haplotype 2; GenBank
with MAX Efficiency® Stbl2TM cells (Invitrogen), sequenced, and accession no. EU278641) and a new variant of haplotype 1
analyzed as previously described (McKnight et al., 2008). (3 additional SNPs), which included a novel -1-bp frameshift
mutation (c.3141delC or p.S1047fsX223; GenBank accession
Sequencing of DSPP Promoter no. EU709728). Human DPP is normally rich in serines (Ser)
and 39 Untranslated Region (UTR) and aspartic acids (Asp) to form a nominal Ser-Ser-Asp repeat
The person with DD (III-9), originally described in Beattie et in which many of the serines are made fully hydrophilic by
al. (2006), as well as three previously described individuals phosphorylation (Takagi and Veis, 1984). The loss of the single
(DD-3, DGI-2, and DGI-3) (McKnight et al., 2008) were used nucleotide results in a translated carboxy-terminal domain rich in
to correlate the disease phenotype with normal variation in the hydrophobic amino acids (predominantly Val, Ile, and Ala) until
DSPP promoter and 39 UTR. The basic DSPP promoter (626 a stop codon is reached 11 amino acids past the normal in-frame
bp 59 to exon 1) was amplified with oligonucleotides (Forward, termination. The c.3141delC frameshift would encode a DPP-
59 GTCTTTATGCACCTTTGGAC 39 and Reverse; 59 GTTTG repeat with ~ 500 normal hydrophilic amino acids, ending with
AAAGCCCAAGGTGG 39), while 480 bp 39 to the stop codon ~ 200 new, hydrophobic ones (Fig. 1). No frameshift mutations
(UTR) were amplified with (Forward, 59 AGTCCATGCAAGG have been found in 160 previously reported normal control
AGATGATCC 39 and Reverse, 59 CAAGTTGGCTATCACAGTAC individuals (McKnight et al., 2008; Song et al., 2008).
39) based on the DSPP (GenBank accession number NM_014208) The haplotype 4-6-2 (markers D4S1534-D4S414-
and surrounding sequences on the UCSC Genome Browser (http:// D4S1572) segregates with DD (Fig. 2 of Beattie et al., 2006).
genome.ucsc.edu/). PCR conditions for both were 94°C for 5 This haplotype was present in all affected individuals, but
min and then 35 cycles of 94°C for 30 sec, 55°C for 30 sec, 3.5 not in unaffected individuals. DSPP is located in this non-
min at 72°C, and a final 5 min at 72°C. The amplicons were recombinant region, specifically between markers D4S1534
directly sequenced with the same oligonucleotides used for PCR and D4S414. Furthermore, incorporating the DSPP HVRR
and analyzed as described above. alleles identified by this study into the haplotypes demonstrated
Downloaded from jdr.sagepub.com at The University of Iowa Libraries on June 19, 2015 For personal use only. No other uses without permission.

International and American Associations for Dental Research

1110 McKnight et al. J Dent Res 87(12) 2008

Discussion
DGI and the less severe DD have classically been considered two
distinct dominant disorders of dentin, based on consistent clinical/
radiographic differences in the phenotype of the secondary teeth.
While there have been occasional reports in the literature of a
spectrum of clinical severity within a kindred, the final diagnoses
remained uniform for all affected family members. For example,
although one of the affected DD family members in our kindred
(III-9) had gray-colored anterior secondary crowns, the authors
specifically noted that the dental phenotype of this individual
remains clearly DD (Beattie et al., 2006).
Studies have localized DD to the interval of chromosome
4q21 containing the DGI locus, and it has been proposed that
Figure 2. Chromosome 4 haplotype analysis incorporating the DSPP
DD and DGI represent allelic mutations of a common gene
HVRR haplotypes generated by this study. The STR marker alleles are (Dean et al., 1997; Beattie et al., 2006). The identification of
from Beattie et al. (2006). DSPP occurs in the non-recombinant portion mutations in the DSP portion of the DSPP gene (i.e., in the
of the haplotype between markers D4S1534 and D4S414. The DSPP first 2% of the translated gene product) confirmed the allelic
HVRR 1* indicates the disease-causing frameshift (c.3141delC) mutation nature of DD/DGI. However, in many cases, mutations could
identified by this study.
not be identified in these early DSPP exons. We found, within
our original cohort, that complete analysis of the DSPP gene
can account for mutation detection in most and perhaps all true
cases of non-syndromic DGI and DD (McKnight et al., 2008).
that the frameshift mutation occurs on the 4-1*-6-2 haplotype DGI and DD mutations can be categorized into 3 types of
associated with DD (Fig. 2). dominant DSPP mutations: (1) changes in the signal peptide; (2)
We analyzed the 626-bp DSPP 59 proximal promoter changes in the highly conserved first 3 amino acids of the mature
domain and the 480-bp 39 UTR (extended by 199 bp to include a protein, Ile-Pro-Val (including exon-3-skipping mutations); and
highly conserved motif 39 to the polyA site) for two persons with (3) deletions resulting in net -1-bp frameshifts in the repetitive
DD and two with DGI, to find any variation in the promoter or domain of DPP (McKnight et al., 2008). Both the signal peptide
message stability domains that may contribute to their different and the Ile-Pro-Val sequence appear to be important in the
phenotypes. No sequence variation was observed in either correct processing of the protein into and through the rough
domain (data not shown), suggesting that the hypothesized cis- endoplasmic reticulum (rER), Golgi apparatus, and/or out of the
acting, DSPP-linked genetic modifier causing the DD vs. DGI odontoblast. Mutations in either of these motifs are hypothesized
phenotypes resides elsewhere within this area of chromosome 4. to interfere with the cell’s normal protein processing. The
-1-bp frameshift mutations cause
a change in the normally soluble
hydrophilic repeat into a large
hydrophobic domain, most likely
resulting in self-aggregation and
precipitation within the rER or
subsequent processing organelles.
These 3 classes of mutations are
hypothesized to have dominant-
negative effects on the dentin
matrix assembly/mineralization,
because odontoblasts are unable
to make and secrete normal
DSPP and/or type I collagen in an
organized fashion or perhaps even
in sufficient quantities to make the
rapidly forming and mineralizing
matrix. Support for this hypothesis
exists in the literature, with Takagi
et al. showing that an extract
of one DGI-II person’s tooth
Figure 3. Similar DSPP mutations that are associated with both DD (top arrows) and DGI (bottom arrows).
The 4 coding exons of DSPP are shown approximately to scale, separated by introns (solid bars) not to
apparently lacked any identifiable
scale. Open triangle shows approximate location of DSP/DPP cleavage site. There are 12 highlighted DPP (Takagi et al., 1983). We also
positions along the DSPP gene where mutations have been identified for either DD or DGI. In the signal presented a corollary hypothesis
peptide (exon 2), c.16T>G was associated with DD, while c.44C>T was associated with DGI. Within the that true null alleles of DSPP or
DPP repeat domain (stippled), 3 DD-associated -1-bp frameshift mutations were previously reported to be nonsense mutations that result
59 to the 6 DGI-associated -1-bp frameshift mutations. The new DD-associated (c.3141delC, *) mutation
is located 39 to the previous DD frameshifts and within the region that was recently associated solely with in non-functional DSP and/or
DGI frameshifts. Reference for each mutation is noted. DPP (for example, all net +1-bp
Downloaded from jdr.sagepub.com at The University of Iowa Libraries on June 19, 2015 For personal use only. No other uses without permission.

International and American Associations for Dental Research

J Dent Res 87(12) 2008 Overlapping DSPP Mutations Cause DGI/DD
frameshift mutations within the repeat immediately result in stop
codons) are predicted to be extremely rare recessive mutations
with a phenotype similar to that reported for the homozygotic
Dspp-null mouse (Sreenath et al., 2003; McKnight et al., 2008).
The hypothesis that DSPP haploinsufficiency in humans will
have no disease phenotype is consistent with the observed lack
of phenotype in the heterozygotic Dspp-null mice.
While all identified mutations for non-syndromic DD and
DGI in humans occur within the DSPP gene, the locations of
these mutations are not specific to a particular phenotype (Fig. 3).
For example, a DD phenotype is caused by a change in the DSPP
signal peptide at c.16T>G in a European individual, while a
Central American person with DGI had a c.44C>T signal peptide
mutation (Rajpar et al., 2002; Malmgren et al., 2004). While
both mutations change the DSPP signal peptide (presumably
resulting in similar improper processing), they result in different
phenotypes within these two unrelated kindreds. Another
example of this interesting genotype/phenotype relationship is
within the third class of DSPP mutations described in this report.
Recently, 2 reports showed that 3 DD-associated frameshift
mutations were 59 to 6 DGI-associated frameshifts (McKnight
et al., 2008; Song et al., 2008), suggesting that persons with
DD encoded longer stretches of mutant hydrophobicity than the
more severely affected persons with DGI. In this current report,
a DD family was discovered to have a -1-bp frameshift mutation
located much more 39 than in previous DD kindreds and, more
importantly, within the DGI cluster defined by the combined
McKnight et al. (2008) and Song et al. (2008) data. Therefore,
this classic DD family demonstrates that the length of mutation-
derived hydrophobicity of DSPP does not correlate with the
specific phenotype.
These results raise an interesting question about why
similar mutations in both the signal peptide of DSPP and
-1-bp frameshift mutations within the repeat domain can result
in either DD or DGI phenotypes that remain stable within
extended kindreds. The presence of a distinct genetic modifying
element very closely linked to DSPP that acts with or upon
the DSPP mutation to determine the specific DD or DGI
phenotype could explain this phenotypic variance. (A less
closely linked or unlinked genetic modifier would result in both
DD and DGI phenotypes appearing within extended kindreds.)
One hypothesis is that this modifying element is directly
involved in controlling the amount of DSPP expressed from the
mutant allele. For example, one variant of the DSPP promoter
within the normal population may result in higher levels of
DSPP mRNA production. When a mutant DSPP is associated
with this more active promoter variant, the increased amount
of precipitating mutant protein may result in a more highly
affected odontoblast population in that person/kindred. Any
of the -1-bp frameshift mutations, for example, when made in
larger quantities, may cause the more severe DGI phenotype.
A corollary hypothesis can be made for normal variants in
the 39 UTR of the DSPP mRNA that may result in higher
message stability and therefore more mutant protein production/
accumulation. The hypothesized promoter, message-stability
motif, or other unidentified linked modifier may or may not
result in any observable phenotypic differences within the
normal population. Our limited evaluation did not identify
sequence differences between the basic promoters (626 bp 59 to
exon 1) or 39 UTR of DSPP among a small set of persons with
DD and DGI. There are, however, many potential regulatory
Downloaded from jdr.sagepub.com at The University of Iowa Libraries on June 19, 2015 For personal use only. No other uses without permission.
International and American Associations for Dental Research
1111
elements more 59 to exon 1 or within the introns that remain to
be investigated. It is also possible that the hypothesized DSPP-
linked genetic modifier may involve another nearby gene (e.g.,
another SIBLING member) or even a currently unknown, non-
coding functional RNA syntenic to DSPP. Since both DD and
DGI phenotypes are observed in many geoethnic areas, it is
reasonable to conclude that both normal DSPP-linked variants
exist in much of the human population.
In summary, a novel frameshift mutation (c. 3141delC) in
a well-documented DD family is described. We suggest that
DGI and DD, when considered together with other previously
reported mutations, are the phenotypic result of a combination
of similar or perhaps even identical mutations in DSPP and a
yet-to-be-described DSPP-linked genetic modifier.
Acknowledgments
This research was supported by the Intramural Research
Program of the NIH, NIDCR and in part (for JPS) by USPHS
Research Grant DE15846, also from the NIH, NIDCR.
REFERENCES
Beattie ML, Kim JW, Gong SG, Murdoch-Kinch CA, Simmer JP, Hu
JC (2006). Phenotypic variation in dentinogenesis imperfecta/dentin
dysplasia linked to 4q21. J Dent Res 85:329-333.
Bellahcene A, Castronovo V, Ogbureke KU, Fisher LW, Fedarko NS (2008).
Small integrin-binding ligand N-linked glycoproteins (SIBLINGs):
multifunctional proteins in cancer. Nat Rev Cancer 8:212-226.
Dean JA, Hartsfield JK Jr, Wright JT, Hart TC (1997). Dentin dysplasia, type
II linkage to chromosome 4q. J Craniofac Genet Dev Biol 17:172-177.
Fisher LW, Torchia DA, Fohr B, Young MF, Fedarko NS (2001). Flexible
structures of SIBLING proteins, bone sialoprotein, and osteopontin.
Biochem Biophys Res Commun 280:460-465.
George A, Srinivasan R, Thotokura SR, Liu K, Veis A (1999). Rat dentin
matrix protein 3 is a compound protein of rat dentin sialoprotein and
phosphophoryn. Connect Tissue Res 40:49-57.
Kyte J, Doolittle RF (1982). A simple method for displaying the hydropathic
character of a protein. J Mol Biol 157:105-132.
MacDougall M, Simmons D, Luan X, Nydegger J, Feng J, Gu TT (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.
Malmgren B, Lindskog S, Elgadi A, Norgren S (2004). Clinical,
histopathologic, and genetic investigation in two large families with
dentinogenesis imperfecta type II. Hum Genet 114:491-498.
McKnight DA, Hart PS, Hart TC, Hartsfield JK, Wilson A, Wright JT, et
al. (2008). A comprehensive analysis of normal variation and disease-
causing mutations in the human DSPP gene. Hum Mutat (in press)
(Epub: June 2008, PMID: 18521831).
Rajpar MH, Koch MJ, Davies RM, Mellody KT, Kielty CM, Dixon MJ
(2002). Mutation of the signal peptide region of the bicistronic gene
DSPP affects translocation to the endoplasmic reticulum and results in
defective dentine biomineralization. Hum Mol Genet 11:2559-2565.
Song YL, Wang CN, Fan MW, Su B, Bian Z (2008). Dentin phosphoprotein
frameshift mutations in hereditary dentin disorders and their variation
patterns in normal human population. J Med Genet 45:457-464.
Sreenath T, Thyagarajan T, Hall B, Longenecker G, D’Souza R, Hong S, et
al. (2003). Dentin sialophosphoprotein knockout mouse teeth display
widened predentin zone and develop defective dentin mineralization
similar to human dentinogenesis imperfecta type III. J Biol Chem
278:24874-24880.
Takagi Y, Veis A (1984). Isolation of phosphophoryn from human dentin
organic matrix. Calcif Tissue Int 36:259-265.
Takagi Y, Veis A, Sauk JJ (1983). Relation of mineralization defects in
collagen matrices to noncollagenous protein components. Identification
of a molecular defect in dentinogenesis imperfecta. Clin Orthop Relat
Res 176:282-290.
 ERRATUM
In “Overlapping DSPP Mutations Cause Dentin Dysplasia and
Dentinogenesis Imperfecta”, by McKnight et al. [J Dent Res
87(12):1108-1111, 2008], a key reference has just been updated.
The complete reference is now: McKnight DA, Hart PS,
Hart TC, Hartsfield JK, Wilson A, Wright JT, et al. (2008). A
DOI: 10.1177/0022034508329738
comprehensive analysis of normal variation and disease-
causing mutations in the human DSPP gene. Hum Mutat 29:
1392-1404.
This is the central reference upon which the authors built the
Rapid Communication.
95