Skip to Main Content
Text size: SmallMediumLargeExtra-Large


April 2013

Larry Fisher, Ph.D.NIDCR scientist Dr. Larry Fisher discovered the five-gene cluster in the late 1990s looking through the first maps of the human and mouse genomes. All mapped to the same region of human chromosome four. All encoded phosphoproteins, or proteins made with a phosphate group bound to one of its amino-acid side chains. All of the proteins were present to various degrees in the original soft extracellular matrices that will harden into new bone, dentin, and cementum. But collagen, the main structural ingredient in the matrices of most tissues, does not calcify into hard mineralized tissue. The assumption among many researchers was the phosphoproteins, which are unique to bones and teeth, must initiate the mineralization process.

Fisher wondered whether the clustering of genes might just be a strange coincidence. His first answer was maybe. Its five proteins didn’t appear to be related. Their amino-acid sequences, the first clue that proteins share a common evolution, were seemingly as dissimilar as the words Beetle and Bailey. But the final answer remained out until Fisher could compare more precisely the five gene sequences and parse their chemical blueprints. The parsing came in 2000, and Fisher and colleague Neal Fedarko noticed that the five proteins shared several short stretches of just a few amino acids. Moreover, short stretches of their DNA, stitched together to direct the genes, shared properties that weren’t obvious when examining amino acids alone. A family they no doubt were. But they were related in the same sense that human communities in diaspora are loosely but inextricably linked.

As Fisher, Fedarko, and colleagues later explained, each member of the cluster “appears to be able to drift substantially [in amino acid sequence] as long as it remains hydrophilic [has an affinity for water] and flexible, and retains a number of [binding] motifs and member-related short amino acid sequences.”

Aware they still had far more questions than answers about the protein family, Fisher and Fedarko steered clear of speculating on the function of its proteins. They opted for a safe, biochemically descriptive name: Small Integrin-Binding LIgand N-linked Glycoproteins, with the easier-to-remember acronym SIBLINGs.

Fisher and Fedarko, however, were willing to propose that the SIBLING cluster originated from a single ancient gene encoded back in the mists of time. This 375,000 base-pair run of redundant gene sequence, like Darwin’s finches, faced the genomic equivalent of survival of the fittest. The SIBLINGs were free to spin much of their duplicated sequence in new directions and happen upon encoding unique molecules that benefited the organism in its quest for survival. Otherwise, the freshly duplicated gene would remain redundant, face gradual mutational decay, and ultimately disappear from the genome. That’s where their shared evolutionary traits seem to have won the day in at least three key ways. One, as acidic molecules, SIBLINGs have an affinity for binding calcium ions at low concentration. Two, their flexibility also allows them to bind multiple proteins. Three, a conserved binding motif in SIBLINGs allowed them to dock their payloads to a subset of integrins. The latter being a key type of cell-surface protein that mediates cell adhesion and serves as a start switch for myriad internal signaling pathways. The assumption is these phosphoproteins must have had a motif or two in generating the new biology that allowed certain vertebrates to develop teeth, jaws, and otherwise survive the drag of gravity as they rose from the water and colonized the land.

The finding wasn’t The Holy Grail that explained why bones and teeth mineralize, while similar matrices do not in the skin, tendon, and ligaments. But it was a genomic treasure, one with the potential to greatly enrich our understanding of the evolutionary origins of bone, teeth, and mammalian biology. Fisher, a biochemist, continued studying the individual SIBLINGs, working his way down to the most problematic one, dentin sialophosphoprotein, or DSPP. It was like the unruly kid at the back of the class. DSPP is so acidic that it doesn’t behave in most standard laboratory assays. “It’s the last protein that most people want to work on,” explained Fisher. He persevered, got lucky in the laboratory to produce enough DSPP to study, and began making a series of important discoveries. Recently, Fisher sat down to talk about what he’d learned about this difficult protein and how the making of dentin can elucidate not only the mysteries of oral biology but Biology with a capital “B".


If the SIBLINGs weren’t good about conserving their original amino-acid sequences, how is it possible to explore their evolutionary origins? How do you study proteins in very different animals when they change so rapidly as to be virtually unrecognizable?

But they can be found. You just have to use a different vantage point. We know that the SIBLING cluster is found between two more highly conserved [and therefore more easily recognized in distant species] genes called SPARCL21 and PDK2. So we can look at the sequence, locate these two genes, and work our way between them to find sequence motifs that are indicative of genes in the cluster. That allows us to perform comparative genomics.

Meaning, in this case, comparing sequenced SIBLING genes from one species to the next.

Right. Dee McKnight, a former postdoc in my lab, performed comparative genomics to study the origins of the SIBLING gene dentin sialophosphoprotein, or DSPP. She looked at the repeat domain of DSPP in 34 mammalian species as well as the chicken and a type of lizard called an anole, the relevance of which will be obvious in a moment.

What are some of the lessons learned?

I think one of the most interesting is DSPP probably was the last of the five SIBLINGs to emerge. We found genomic evidence that DSPP seems to have arisen about 300 million years ago via a duplication of the ancient SIBLING gene dentin matrix protein-1 [DMP-1].

I’ve heard that exon five of the DSPP gene is kind of a deadman’s curve of gene sequencing.

Yes, it is really strange genomic territory. In humans, exon five consists of about 230-250 repeats, each containing roughly the same nine base pairs of DNA. If you think of a base pair as a scrabble block, the repeats spell out the same word, say, “Cleveland,” over and over with occasional typos slipped in. This sustained stutter of sequence occurs in all known mammals with teeth, although the length of the repeats varies from species to species. Among those studied so far elephants, whose ivory is pure dentin, have the shortest number of repeats at about 75.

In humans, DSPP’s exon five encodes a chain of 700 to 800 amino acids with a distinctive pattern of serine-aspartates interrupted sometimes by a lysine. Over and over. This unique pattern produces an exceedingly hydrophyllic protein. Or, in lay terms, DSPP is probably the most water-loving protein in the body. It also is probably the most acidic protein.

What does exon five look like in reptiles?

It’s full of serine plus asparate repeats, too. But here’s the key point. We found that the sequence appears to be of different origin than that in their mammalian counterpart. Reptiles seem to have worked from a different assemblage of genomic letters to create a different scrabble word, which is dotted with an occasional arginine-based motif. It’s the same biochemical outcome. But it’s reached by a different evolutionary route.

How could that be?   Green anole

We think it suggests the following scenario. The DMP-1 gene of the ancient shared ancestor of mammals and reptiles underwent a spontaneous duplication. Then in the line that gave us mammals, the DMP1 gene closest to the SPARCL1 gene expanded a specific small set of three amino acids, serine-serine-aspartate, leading to what we now call DSPP and leaving the second copy of DMP1 to retain its more ancient function. The line that gave us the anole, and presumably all other reptiles, left the copy of DMP1 closest to SPARCL1 intact and appears to have independently modified a slightly different set of serine sans aspartate amino acids to select for many repeats that give rise to a protein similar in character to mammalian DSPP. The DNA sequences encoding the two DSPP-like proteins have left us a history of what happened millions of years ago as the teeth of reptiles and mammals evolved into similarly functioning structures.

Meaning DSPP should have a lot to say about mammalian evolutionary biology?

And possibly human migration patterns, too. We noticed different repeat lengths in different ethnicities. For example, some repeat lengths and minor differences with the sequences themselves seemed specific to individuals of Far East Asian descent. Other differences were unique to those of more recent African descent or Northern Europeans. I should say, though, that our data remain more observational than conclusive. I’m a biochemist, not a geneticist. Pursuing the more conclusive data falls outside of my lab’s focus and expertise. But, as I’ve often said, if others would like to pick up the issue, I’m completely open to assisting in any way possible.

Fast Forward to the Present

The interest in SIBLINGs grew out of their presumed role in mineralizing extracellular matrices into bone, dentin, and cementum. Whether that presumption is true or false, I understand remains an open question. Is that also the case for DSPP?

I think so. As the “D” in its name suggests, DSPP is most highly expressed in developing dentin. But more recently, DSPP has been found to be expressed at much lower levels in bone, cartilage, and certain cells in the kidneys, salivary glands, and sweat glands. In bone, for example, others have described DSPP as being expressed at 1/400th the level it is in dentin. So one can ask: Does DSPP function in the same way in these tissues as it does in a tooth? It seems like a long shot, particularly in the soft tissues, which, of course, don’t mineralize. I mention this only to suggest that the function question likely will have more than one final answer.

But what about the function of DSPP in making dentin?

Well, the broad outlines are fairly clear. If you do not make enough DSPP, you do not have a well-mineralized dentin. So, what happens is tooth-forming odontoblast cells express the gene and make the DSPP protein as what’s known as a pro-protein.

That is, a large precursor protein.

Right, the odontoblasts produce the pro-protein and rapidly secrete it into the extracellular space. A highly conserved enzyme clips DSPP into two smaller presumably biochemically active fragments, DSP and DPP. I say presumably because the function of both proteins remains undetermined. Although DSP and DPP are present in the extracellular matrix that will mineralize into dentin, we don’t know when and where they are actually performing their specific functions. So the research story very much continues.

At the gene level also. In the early 2000s, the DSPP gene turned out to be mutated in a number of families with histories of inherited dentin abnormalities.

Since that discovery, more than 30 mutations have been identified in DSPP that are associated with nonsyndromic dentinogenesis imperfecta type II [DGI-II] and dentin dysplasia [DD]. DGI-II is the more severe condition. It is characterized by dentin that develops abnormally, mineralizes poorly, and leaves people with fragile teeth and serious dental problems.

What is the inheritance pattern?

As you know, we inherit two copies of DSPP. Each inherited gene is called an allele. In a series of mouse experiments, our NIDCR colleague Ashok Kulkarni knocked out one allele and found the dentin appeared to form just fine. When he knocked out both alleles, the dentin was abnormal. We think that the same is true in humans, but to date no one has found any families that are missing both DSPP alleles. But in humans, DG-II and DD arise when only one allele is altered, not missing. The mutation has more of an effect than if it was entirely missing. It’s called a dominant-negative effect. And the question is why does this dominant-negative effect occur?

And the answer to the question is written in the specifics of the mutations.

DSPP gene sequence That’s right. The mutations fall into three categories. One consists of a lone reported mutation at the beginning of exon two and really represents fascinating biology that we can’t explain at this point. The mechanism by which this single mutation causes the mild form of dentin dysplasia remains a real mystery to us.

We proposed that all of the other mutations can be clustered into just two categories. The first includes all of the other cases, where the loss of one or four DNA bases in the DPP part of the protein [from exon five] causes DSPP to change from being very soluble in water to having a strong propensity to insert itself into the oil-like membranes of the cell. The second category includes all of the mutations that result not in the loss of a DNA base but in the change of a single DNA base. We propose that all of these mutations ultimately end up throwing off the coding of the mature protein’s first three amino acids. They are isoleucine (I)-proline (P)-valine (V) .

You call the three amino acids the IPV motif. Why is it such a mutational hot spot?

Frankly, changes of a single amino acid in most places in the DSPP protein would likely have no consequences. The protein is not highly conserved. But there are some single changes that do have immediate effects. The IPV motif seems to work like a short -but-important barcode. We think a yet-to-be-discovered protein receptor within the endoplasmic reticulum [ER] seems to use the motif to recognize the protein, bind it, and contribute to its rapid removal from the site of first protein synthesis, which is the ER.

Just to define terms. The endoplasmic reticulum is an organelle within the cell. It’s a metaphorical conveyor belt, where proteins destined to be released from the cell first get synthesized. The proteins then are rapidly passed onto the parts of the cell that finish packaging them. Then, they are sent out of the cell to perform their functions.

That sounds right. For my lab, what was so helpful in looking at the immediate consequences of the many mutations in the second category is we could collapse them into one biochemical observation: All cause changes in the simple IPV motif. That gave us the opportunity to simulate the mutational changes in the laboratory and model what likely happens.

Oral Biology, Extraordinary Biology

What happens, at least in silico, takes the traditional concept of a mutation in a new direction. Photos of Wildtype DSPP and Mutant DSPP

We think so. When people talk about mutations, they refer to function. That is, the mutated gene compromises a cell’s ability to perform a specific metabolic function. But in the case of DSPP, our data indicate that the fundamental problem isn’t related to its ultimate function outside the cell at all. The first category of mutations causes these mutant proteins to insert or dissolve into the structure of the ER itself rather than be rapidly shuttled out in preparation for leaving the cell. In the IPV mutations, the mutant protein remains soluble in the ER, but its receptor apparently cannot recognize its changed barcode, grab it, and help it on its way out of the ER.

Meaning, in both cases, mutant DSPP builds up on the conveyor belt?

Right, it builds up on the conveyor belt. But remember that we are pretty sure that just losing the mutant protein within the cell should not cause the dentin diseases. Mice - and very likely humans - that make only half of the normal amount of DSPP due to complete loss of one of their DSPP genes have completely normal teeth. So we had to go one step farther and propose how the mutant proteins stuck in the ER’s conveyor belt system also could cause the loss of some of the normal DSPP being made from the normal gene in all of the families reported with both DD and DGI.

How’d you take that next step?

The ER is a calcium-rich environment. What we think happens is while the mutant DSPP hangs around to be shuttled out of the ER, these molecules do exactly what nature has designed them to do. They bind calcium. The more calcium these molecules bind, the more they start to clump together into complexes that cannot be passed onto the next stage of secretion out of the cell. Our working hypothesis is these complexes also pull in the calcium-coated normal DSPP, and little of it can get out either. What’s really interesting about working with DSPP is that when things start to go wrong in the ER, they go wrong very quickly.

Why’s that?

Well, DSPP is THE extreme example of an acidic protein destined to be secreted out of a cell. Proteins don’t get any more acidic than DSPP. In fact, most cells refuse to make it, and that’s been a major challenge for us to generate a supply to study. But in the cells that make DSPP, our surmise is the ER receptor spots the IPV motif in normal DSPP and shoos it out of the ER very, very quickly.

It’s more like a NASCAR pit crew than a car wash.

Right, they want it gone. But the cells end up choking on these DSPP-calcium complexes. We asked whether this phenomenon is unique to DSPP. We first looked at the other SIBLINGs and saw that they had roughly the same problem leaving the ER when their IPV motif was changed. Then we looked more broadly at acidic proteins in general that are destined to be secreted out of the cell and into the calcium-rich environment outside. Same story. If a protein is acidic and has the ability to interact with calcium, cells seem compelled to get it out of the ER very quickly.

Is the IPV motif also in play for the acidic proteins from creatures other than mice and men?

We think so. The early genomic evidence suggests that this mechanism for getting acidic proteins out of the cell may in fact be an extremely ancient one. We looked at the sea urchin, which produces an acidic tooth protein that is possibly an analogue to DSPP. The analogue protein also had the motif at the beginning that may help it get out of the ER. Most recently we have begun to look in the simple, one-celled baker’s yeast. It turns out that one of the most highly studied secreted proteins in several popular species of these fungi is acidic and has an IPV-like motif.

Is the motif already well recognized in the research literature?

No, it’s hardly been recognized at all. Our hunch is a small family of receptors may be responsible for shuttling the various acidic proteins out of the cell. The next task is to find them to gain a clearer picture of the process. But we think the motif will be a common mechanism for molecules in every field, including acidic hormones and growth factors that have this motif. We’re proposing they all would use this same mechanism to get out of the ER.

What does this mean for families with the dentin anomalies?

From a clinical perspective, the question might be is there some way to help the normal proteins out of the ER before they get sidetracked into forming complexes? The answer may be driven by work in the collagen field. Collagens – and there are, of course, many different types -- have the same problem. Cells can make an inappropriate collagen molecule, and they recognize the protein as badly assembled and won’t permit it to exit the ER. When things go wrong, usually the problem is not enough appropriately assembled, normal collagen gets out of the cell. The cell is so busy trying to handle the inappropriately-made collagens that it doesn’t assemble enough of the correct collagen complexes.

You mentioned earlier that DSPP likely is transported out of the ER via a still-undetermined protein receptor. Would the same be true here with collagen?

Yes. In fact, collagen type VII is one of the few molecules whose specific receptor for getting out of the ER is known. That was actually discovered in diseases of Type VII collagen. There was a person who had a Type VII collagen disease, but nothing wrong with the Type VII collagen. When geneticists found the mutation, it turned out to be this receptor in the ER that helps Type VII get out. The person made correct collagen complexes, but not enough could get outside the cell. So the person had the same disorder because of the receptor problem, and I think the DSPP situation could be a variation on this theme.

Biology Knows No Bounds

I’ve heard researchers say that it can be an uphill battle to publish major findings in oral biology in some of the high-impact journals. In other words, the 20th century’s biological divide between the mouth and body remains a stubborn artifact in the early 21st century. True?

I think that there is some truth to that. There seems to be a lingering idea that results or hypotheses that originate in the study of oral tissues, particularly teeth, have no application beyond those tissues. Of course, that’s just not true. Oral biology is extraordinarily rich and can help to untangle the complexity of human biology writ large. By analogy, if you want to learn to speak French, you don’t necessarily have to rent an apartment in Paris for six months. You can visit just about any continent on the globe and challenge your French language skills. It’s the same story with human biology.

The language of biology runs from head to toe.

Exactly. As a biochemist, I consider it an advantage to study oral biology. I really do. The mouth is easily accessible and spans a range of soft and hard tissues, each with its own profound complexity. The NIDCR always has been highly supportive of taking the most compelling discoveries in the mouth and applying them broadly to other systems. That remains a real strength of this institute. Its main focus certainly is on dental-oral-craniofacial health-related questions. But if a discovery has obvious or complimentary implications outside the mouth, the institute’s leadership has been highly supportive of taking the ball and running with it for better public health.

In other words, while discoveries made below the neck can influence the framing of questions in oral research, the opposite also is true. It’s a two-way street.

Yes, and I think my lab’s work on the SIBLINGs illustrates this point.

Share This Page

GooglePlusExternal link – please review our disclaimer

LinkedInExternal link – please review our disclaimer


This page last updated: February 26, 2014