MBE Advance Access originally published online on December 2, 2005
Molecular Biology and Evolution 2006 23(3):644-654; doi:10.1093/molbev/msj072
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Research Article |
Molecular Evolution of the Primate Developmental Genes MSX1 and PAX9
* School of Human Evolution and Social Change, Arizona State University and Center for Evolutionary Functional Genomics, The Biodesign Institute and School of Life Sciences, Arizona State University
E-mail: acstone{at}asu.edu.
 |
Abstract |
|---|
 TOP Abstract IntroductionMaterials and MethodsResultsDiscussionConclusionSupplementary MaterialAcknowledgementsReferences |
|---|
In primates, the craniofacial skeleton and the dentition are marked by high levels of interspecific variation. Despite this, there are few comparative species studies conducted at the molecular level to investigate this functional diversity. We have determined nucleotide sequences of MSX1 and PAX9, two developmental genes, in a sample of 27 diverse primate species in order to identify coding or regulatory variation that may be associated with phenotypic diversity. Our analyses have identified four highly conserved noncoding sequences, including one that is conserved across primates and with dogs but not with mice. Although we find that substitution rates vary significantly across MSX1 exons, comparisons of nonsynonymous and synonymous substitution rates (dN/dS) suggest that, as a whole, MSX1 and PAX9 amino acid sequences have been under functional constraint throughout primate evolution. Compared to all other primates in our sample, our analysis of exon 1 in MSX1 finds an unusual pattern of amino acid substitution for Tarsius syrichta, a member of a lineage (tarsiers) that has many unique features among primates. For example, tarsiers are the only extant primates without deciduous incisors, and MSX1 is expressed exclusively in the incisor regions during the earliest stages of dental development. Our overall results provide insight into the utility of comparative species analyses of highly conserved developmental genes and their roles in the evolution of complex phenotypes.
Key Words: comparative genomics • conserved noncoding sequence • homeobox • craniofacial skeleton • dentition • tarsier • callitrichid • strepsirrhine
|
Introduction |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionConclusionSupplementary MaterialAcknowledgementsReferences |
|---|
During growth and development, a suite of genes interact under tight spatiotemporal control to ultimately form complex organs (Levine and Davidson 2005
), and many developmental genes are each involved in the formation of more than one organ system. Because of this, amino acid sequences of these genes are often under strong functional constraint. Regulatory elements can be system specific (e.g., Woolfe et al. 2005) and thus may be less constrained by requirements for the development of multiple organs. However, little is known about the evolutionary genetics of developmental genes in general, although the studies that have been conducted (e.g., engrailed, HLX, and Hox genes) have provided preliminary insight into the evolution of development (Logan et al. 1992; Kappen and Ruddle 1993; Eizinger, Jungblut, and Sommer 1999; Santini, Boore, and Meyer 2003; Bates, Wells, and Venkatesh 2005).
The potential evolutionary significance of nucleotide substitutions, both coding and regulatory, in developmental genes is obvious, yet we lack straightforward approaches for linking nucleotide-level differences to phenotypic variation observed across a group of nonmodel organisms, such as primates. Given the many and complex roles of most developmental genes, it is rarely possible to make solid evolutionary conclusions based on comparative sequence data alone. Evolutionary interpretation can be facilitated by well-designed experimental studies (see Belting, Shashikant, and Ruddle 1998; Gompel et al. 2005), but such studies (and the direct conclusions to be made from them) are largely limited to model organisms. One reasonable approach is to identify unusual molecular evolutionary histories in developmental genes through sequence comparisons among groups of closely related but phenotypically diverse organisms and then later use experimental methods (e.g., transgenic experiments) and population genetics to test hypotheses generated from these data.
Extant primates show a wide range of phenotypic adaptations to diverse environmental conditions, including substrate and diet (Fleagle 1999). The evolution of this phenotypic variation likely involved significant changes in the functions and expression levels and patterns of developmental genes, which in some cases may have left detectable molecular genomic signatures. Msh homeobox homolog 1 (MSX1) and paired box gene 9 (PAX9) are two genes that have been implicated in the development of the craniofacial skeleton and the dentition (Satokata and Maas 1994; Peters et al. 1998), which in primates are marked by high levels of interspecific variation (Fleagle 1999; Swindler 2002). Additionally, a number of MSX1 and PAX9 mutations are associated with human diseases of the craniofacial skeleton and the dentition, including the agenesis of six or more permanent teeth and cleft lip and palate (e.g., Vastardis et al. 1996; Stockton et al. 2000; Lidral and Reising 2002; Jezewski et al. 2003; Lammi et al. 2003; Klein et al. 2005), and knowledge of the disease phenotypes provide us with an additional level of comparison to patterns of interspecific molecular variation. Besides development of the craniofacial skeleton and dentition, MSX1 is also known to be involved in limb, muscle, and nail development (Jumlongras et al. 2001; Lee, Habas, and Abate-Shen 2004; Lallemand et al. 2005) and PAX9 is known to play an integral role in organs derived from pharyngeal pouches (e.g., parathyroid glands) and in the development of tongue, vertebral column, and limbs (Peters et al. 1998, 1999; Jonker et al. 2004). In this study, we determined and compared nucleotide sequences of MSX1 and PAX9 in a sample of 27 diverse primate species, in order to investigate patterns of molecular genetic variation in coding and regulatory regions that may shed light on phenotypic diversity associated with developmental genes.
|
Materials and Methods |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionConclusionSupplementary MaterialAcknowledgementsReferences |
|---|
Sampling and DNA Extraction
Blood, tissue, or genomic DNA was obtained from various primate centers and from the Integrated Primate Biomaterials and Information Resource. Table 1 lists the species included in our study and the source for all samples. One individual from each species was included. From blood and tissue samples, DNA was isolated using a standard phenol/chloroform extraction method (Sambrook and Russell 2001
).
|
Sequencing and Alignment
Human (build 35.1, accession NT_006051.17 for MSX1, NT_026437.11 for PAX9) and mouse (Mus musculus; build 33.1, NT_039303.3 for MSX1, NT_039551.3 for PAX9) nucleotide sequences were obtained from GenBank. Our goal was to analyze representational coding and noncoding (including intron and regulatory) regions from each gene. Polymerase chain reaction (PCR) primers were originally designed in 100% conserved regions (no nucleotide differences) between humans and mice. Following initial PCR amplification and DNA sequencing, species-specific primers and sequencing primers were designed as needed. Primers and PCR conditions are listed in Supplementary Table 1 (Supplementary Material online). Following amplification, PCR products were purified with Shrimp Alkaline Phosphatase and Exonuclease I (USB Corporation, Cleveland, Ohio), cycle sequenced with the Applied Biosystems BigDye Terminator Cycle Sequencing Kit version 3.1, cleaned with isopropanol, and run on a 3730 automated capillary sequencer (Applied Biosystems, Foster City, Calif.). All regions (fig. 1) were sequenced in both forward and reverse directions for 14 primate species (table 1). Sequence data were assembled using the Lasergene SeqMan program (DNAStar, Madison, Wisc.) and visually checked for accuracy.
|
A total of 888 bp of MSX1 coding sequence (lengths are based on the human sequence and exclude start and stop codons) and 1,435 bp of MSX1 intron sequence were obtained. Tarsius syrichta sequence is reported only for the first exon of MSX1 (449 bp). Using degenerate primers designed from the primate consensus MSX1 exon 2 amino acid sequence, we did obtain putative exon 2 nucleotide sequence for T. syrichta. However, this exon 2 amino acid sequence is strongly conserved not only across species but also across other genes (e.g., MSX2). Therefore, the orthology of this T. syrichta sequence is in question, and repeated attempts to amplify this region with other primers were unsuccessful. Thus, this region for T. syrichta was omitted from all analyses.
For PAX9, the following regions were sequenced: 677 bp of the region directly upstream of the first exon (T. syrichta sequence was not obtained for this region), 174 bp of the 5' untranslated region (UTR), 1,020 bp of coding sequence, and 1,210 bp of the introns. Sequences generated for this study have been deposited in GenBank with the accession numbers DQ067468–DQ067556. Coding region sequences were aligned with the SeqMan program. Intronic and other untranslated (i.e., upstream region and UTR) regions were aligned using the MultiPipMaker program (Schwartz et al. 2003) and then checked visually.
Analytical Methods
Phylogenetic relationships were inferred from combined data sets of introns and third-position exon-coding sites with the neighbor-joining method and 10,000 bootstrap replications using MEGA3 (Kumar, Tamura, and Nei 2004). Based on this analysis and others (e.g., Goodman et al. 1998; Pastorini et al. 1998; Pastorini, Thalmann, and Martin 2003; Roos, Schmitz, and Zischler 2004), we constructed a consensus tree as depicted in figure 2.
|
Identification and analysis of potential regulatory elements were facilitated with the phylogenetic shadowing method of Boffelli et al. (2003)
, as implemented in the eShadow program (Ovcharenko, Boffelli, and Loots 2004). Traditionally, conserved noncoding sequences (CNS) have been identified with comparisons of sequences from pairs or trios of distantly related species (e.g., humans, mice, and rats). However, cumulative divergence among a large number of more closely related sequences can be greater than that of pairs or trios of relatively distantly related species, and therefore, we can also identify regions of conserved evolution using sequence data from a greater number of more closely related taxa. Phylogenetic shadowing identifies regions with significantly greater levels of conservation than would be expected by chance, given the phylogenetic relationships among the sampled species. Our use of this method has the additional benefit of highlighting potential regulatory elements that are primate lineage specific. Using a hidden Markov model (HMM) implemented in eShadow, we scanned upstream and intron region sequences for CNS. To fit the analysis to the nucleotide divergence levels observed among the primate species in our study, we used a maximum likelihood method to "train" the HMM parameters using our exon sequence data, as in Ovcharenko, Boffelli, and Loots (2004). Levels of conservation were visualized by creating sequence similarity plots with the Phylo-VISTA program (Shah et al. 2004).
To determine whether amino acid substitutions fit a neutral model of molecular evolution, we computed lineage-specific ratios of nonsynonymous substitutions (N) per nonsynonymous site to synonymous substitutions (S) per synonymous site (dN/dS) using codon-based maximum likelihood methods contained in the codeml program of the software package PAML (Yang 1997). This analysis uses the codon substitution model of Goldman and Yang (1994). We used a model that assumed unequal transition and transversion substitution rates and unequal codon frequencies (codon frequency parameters were calculated using observed nucleotide frequencies at the three codon positions). To calculate lineage-specific ratios, dN/dS was allowed to vary freely among branches (free-ratio model). Then, to test the null hypothesis that dN/dS is equal among lineages, we used the likelihood ratio test (tables 2 and 3) to determine whether the free-ratio model has a significantly higher likelihood than does a one-ratio model in which each branch has the same (i.e., background) dN/dS (Yang 1998). For each lineage with dN/dS > 1, we tested the null hypothesis of neutral evolution for amino acid substitutions and whether dN/dS on that lineage was different from the background dN/dS ratio, using likelihood ratio tests as described by Yang (1998).
|
|
|
Results |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionConclusionSupplementary MaterialAcknowledgementsReferences |
|---|
Conserved Noncoding Sequences
Phylogenetic shadowing analyses of noncoding regions identified four CNS (fig. 3): one in MSX1 intron 1, one in the region upstream of the PAX9 5' UTR, and two within the first intron of PAX9. MSX1 CNS 1 is 178 bp and overlaps the MSX1 intronic human-mouse homology region (MIHR) (Jezewski et al. 2003
). Human-mouse conservation is 87% (no indels) in MSX1 CNS 1, including a stretch of 56 bp that is 100% conserved. Within MSX1 CNS 1, there are only a few substitutions across the different primate lineages. Within the 56-bp region that is 100% conserved between humans and mice, we identified only one substitution, in Saimiri boliviensis.
|
PAX9 CNS 1 covers the 41 bp immediately upstream of the 5' UTR and is likely part of the PAX9 promoter. Two CNS are also found in the first intron of PAX9. In PAX9 CNS 2 (80 bp), Perodicticus potto, Propithecus verreauxi, and Daubentonia madagascariensis lineage substitutions disrupt otherwise 100% conserved stretches of 14, 17, and 13 bp, respectively. Within the 31-bp PAX9 CNS 3, there are only two substitutions, one each inferred to have occurred along the D. madagascariensis and T. syrichta lineages. We conducted an a posteriori test to assess whether the CNS substitution rate, dC, in any lineage is greater than the nonconserved intron region substitution rate, dI, determined using the baseml program in PAML (Yang 1997
). A dC/dI ratio > 1 may suggest that positive selection has led to the fixation of CNS substitutions. This is analogous to a test developed by Hahn et al. (2004) for the comparison of the substitution rate at transcription factor–binding sites to the substitution rate at nonbinding sites. With this analysis, we found only one CNS and only one lineage within that CNS, for which the two substitution rates are similar (P. verreauxi PAX9 CNS 1 dC/dI = 1.05, not significantly greater than 1). In all other cases, dC/dI < 1. Thus, there is no evidence to suggest that CNS substitutions in any lineages occurred at a faster rate than substitutions in nonconserved intron regions.
PAX9 Amino Acid Substitution Patterns
For the coding region of PAX9, the cumulative (across all primate lineages) dN/dS = 0.029. Log likelihood values and dN/dS estimates from each maximum likelihood model considered are given in table 2, and likelihood ratio test results are presented in table 3. There are no lineages for which dN/dS > 1. In fact, for any one lineage the largest dN/dS = 0.135 (Pan troglodytes; N = 1.0, S = 2.1). The null hypothesis of dN/dS homogeneity among lineages was not rejected. To determine whether this pattern of low dN/dS ratios is consistent across additional lineages, we sequenced the largest PAX9 exon (exon 2; 627 bp) in another 13 primate species (table 1) and observed similarly low rates of amino acid substitution (see Supplementary Figure 1, Supplementary Material online).
Deduced amino acid sequences for PAX9 are depicted in Supplementary Figure 1 (Supplementary Material online). The paired box–binding domain of PAX9 is 100% conserved at the amino acid level across the primates we sampled. All other regions of the gene are also largely devoid of substitutions.
MSX1 Amino Acid Substitution Patterns
Considering the entire coding region of MSX1, only the platyrrhine/catarrhine ancestral lineage has dN/dS > 1 (N = 1, S = 0). This dN/dS is not different from the background ratio (data not shown) and the cumulative dN/dS = 0.057. We failed to reject the null hypothesis of dN/dS homogeneity among lineages.
The deduced MSX1 amino acid sequences suggested a difference in selective pressures between the two MSX1 exons, which we evaluated by separating the first and second exons and repeating the dN/dS analyses (the codon split between exon 1 and exon 2 was removed). The second exon of MSX1 contains the homeobox-binding domain (180 bp). This domain is 100% conserved at the amino acid level across all the primates we sequenced. We also found a high level of amino acid conservation in other exon 2 regions. We infer the occurrence of only one exon 2 nonsynonymous substitution among all primate lineages (N = 1, N sites = 346.7, substitution inferred to have occurred in the S. boliviensis lineage). The cumulative dN/dS of exon 2 is 0.002, and dN/dS does not vary significantly among branches. Compared to exon 2, nonsynonymous substitutions in MSX1 exon 1 (cumulative dN/dS = 0.154) are relatively more frequent per nonsynonymous site (excluding T. syrichta: N = 41.92, N sites = 346, Fisher's exact test; P < 0.001). The synonymous substitution rate is not significantly different between the two exons (exon 1 S = 91.51, exon 1 S sites = 101, exon 2 S = 117.56, exon 2 S sites = 91.3, Fisher's exact test; P = 0.09). The hypothesis of dN/dS homogeneity among branches was rejected for exon 1 (P < 0.05).
For MSX1 exon 1, dN/dS > 1 in four lineages. In each case, dN/dS = 
, where N is nonzero and S is zero: P. verreauxi (N = 3.1), Leontopithecus rosalia (N = 3.2), the Callimico goeldii/Callithrix jacchus common ancestor (N = 1.1), and the platyrrhine/catarrhine common ancestor (N = 1). The dN/dS ratios of the P. verreauxi and L. rosalia branches are significantly greater than the background MSX1 exon 1 dN/dS (P < 0.05). However, neither dN/dS is significantly greater than 1. To better assess whether the elevated dN/dS ratios in the P. verreauxi and L. rosalia lineages are unusual compared to MSX1 exon 1 dN/dS ratios in other primate lineages, we sequenced exon 1 in an additional 13 primate species (listed in table 1) and incorporated these data into our analyses ("large tree" in tables 2 and 3). The results are depicted in figure 2. With these additional data, we observed that the Lemur catta and Gorilla gorilla lineages both have dN/dS = (N = 2.1 and N = 2.0, respectively). In both cases, dN/dS is significantly greater than the background dN/dS (P < 0.05) but not significantly greater than 1.
The deduced MSX1 exon 1 amino acid sequences of all primate species of our study and mice are depicted in figure 4. Most substitutions cluster in two discrete regions of the protein. Included in these clusters are 12 positions at which we have inferred the occurrence of amino acid substitution homoplasy. Outside these regions, there are few amino acid substitutions, even between primates and mice. We infer many such substitutions to have occurred along a single lineage, T. syrichta. Compared to other primates, T. syrichta lineage substitutions occur more evenly across MSX1 exon 1. Several substitutions occur at positions or in regions that are otherwise 100% conserved across primates and with mice. The T. syrichta exon 1 dN/dS of 0.203 is not elevated relative to the cumulative dN/dS of 0.151 (fig. 2), which is a function of many synonymous substitutions (S = 35.6) along this long branch rather than a lack of nonsynonymous substitutions (N = 25.0). Interestingly, the average number of amino acid differences between T. syrichta and other primates (25.72; range 24–32) is greater than the number of amino acid differences between mice and primates other than T. syrichta (17.8; range 16–25).
|
|
Discussion |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionConclusionSupplementary MaterialAcknowledgementsReferences |
|---|
The MSX1 and PAX9 genes have critical roles in the development and function of many organs. Perhaps partly reflecting this, we found that, on the whole, the coding sequences of MSX1 and PAX9 suggest a history of strong functional constraint throughout primate evolution. However, we identified a region of MSX1 (exon 1) that is less conserved at the amino acid level than other gene regions. Additionally, amino acid substitution rates at this exon vary across primates, and we find a particularly unusual substitution pattern in the T. syrichta lineage. Here, we discuss the implications of these comparative sequence data for understanding the evolutionary history of primate MSX1 and PAX9 genes, hypotheses concerning their role in primate phenotypic evolution, and insights into MSX1- and PAX9-associated human diseases.
Conserved Noncoding Sequences
Our phylogenetic shadowing analyses of MSX1 and PAX9 noncoding sequences identified four CNS that may potentially contain regulatory elements. Ideally, CNS would be known a priori to avoid phylogenetic shadowing type II error. For instance, substitutions in a CNS may have occurred at a higher rate than the background substitution rate (i.e., they were positively selected for) if they were involved in adaptive change. In such a scenario, a reduction in average conservation would result, and the phylogenetic shadowing method may fail to identify an evolutionarily important CNS. This does not mean that the CNS we identified should be disregarded because the above scenario should not have a large effect on type I error rates.
MSX1 CNS 1 may be directly involved in the regulation of MSX1 expression or could be indirectly involved by way of an antisense mRNA transcript that initiates within the MSX1 3' UTR and terminates in MSX1 CNS 1 (Blin-Wakkach et al. 2001; Berdal et al. 2002). The 2.2-kb antisense product has been detected in developing dental and bone tissues and may block expression of MSX1 by binding to the sense mRNA transcript (Berdal et al. 2002). Future experimental studies may be able to determine whether any of the primate MSX1 CNS 1 nucleotide substitutions affect expression levels of the antisense transcript and if this might be involved in phenotypic diversity.
The two CNS found in the first intron of PAX9 have not, to our knowledge, been described previously and are of unknown functional significance. Conserved elements are found preferentially in first introns, and it has been suggested that first intron CNS likely have a role in transcriptional regulation (Majewski and Ott 2002). Interestingly, PAX9 CNS 2 (80 bp) does not appear to be conserved between primates and mice, with 59% human-mouse conservation including indels and 71% human-mouse conservation when indels are excluded. This level of conservation is similar to a 67% conservation estimate previously reported for genome-wide human-mouse noncoding (nonconserved) regions (Waterston et al. 2002). We further investigated the evolutionary history of this CNS by examining the dog (Canis familiaris) nucleotide sequence for this region using the human-dog alignment from the University of California Santa Cruz genome browser (http://genome.ucsc.edu/). In contrast with the mouse, PAX9 CNS 2 human-dog conservation is apparently greater, 88% (no indels), than general levels of human-dog conservation, 78% (for this estimate, we used MultiPipMaker to align an 
400-kb nucleotide sequence region that included 200 kb upstream and 200 kb downstream from the PAX9 locus [human accession NT_026437.11, dog NW_139865.1]; of 336,174 aligned nucleotide sites [indels excluded], 263,849 were similar in humans and dogs). Although we are not able to make a statistical statement regarding the significance of these comparisons, this result suggests that the lack of human-mouse conservation at PAX9 CNS 2 reflects changes along the mouse lineage, which may or may not be associated with PAX9 gene regulation in mouse.
Amino Acid Substitution Patterns
Our analyses identified few nonsynonymous substitutions at the PAX9 gene across the primate species we sampled. The cumulative PAX9 dN/dS = 0.029. There are no lineages for which dN/dS > 1, and dN/dS ratios do not vary among branches. Together, these data suggest a history of strong functional conservation for PAX9 throughout primate evolution. Given the strong levels of amino acid conservation at this gene, the few substitutions that were observed may be of functional relevance. However, there are no primate lineages with an unusual PAX9 amino acid substitution pattern (i.e., more substitutions than in other lineages).
MSX1 dN/dS ratios suggest that, as a whole, the amino acid sequence of this gene has also been under functional constraint throughout primate evolution. We observed that the nonsynonymous substitution rate in MSX1 exon 1 is significantly higher than that in exon 2, suggesting that there have been different selection pressures acting on amino acid changes across different regions of this gene. In a clinical (case/control) study of human MSX1 variation in relation to cleft lip and/or cleft palate, Jezewski et al. (2003) observed a significantly higher proportion of synonymous single-nucleotide polymorphisms in exon 1 compared to exon 2, which they proposed could reflect the presence of regulatory elements within exon 2. Purifying selection to maintain the specific sequence of exonic regulatory elements could have altered the rate of synonymous substitution (see Pagani, Raponi, and Baralle 2005). If regulatory elements leading to a reduced synonymous substitution rate are preferentially present in the second exon of MSX1, then this same mechanism could also have affected the nonsynonymous substitution rate. However, we did not find a significant difference in synonymous substitution rates between the two exons (the exon 2 synonymous substitution rate was even slightly higher); thus, the greater level of amino acid conservation in exon 2 cannot easily be explained by a general difference in the neutral substitution rates between the two exons. This result also does not support the suggestion that regulatory elements may be preferentially present in exon 2 (Jezewski et al. 2003), at least across the primates included in our sample.
Unlike for exon 2, the hypothesis of dN/dS homogeneity among branches was rejected for exon 1 (P < 0.05), which could reflect either positive selection or a relaxation of functional constraint for amino acid substitutions along one or more lineages. Although several lineages with relatively elevated dN/dS were identified, no dN/dS is significantly greater than 1, and thus, we cannot reject a null hypothesis of neutral evolution for any single lineage. Therefore, the pattern of dN/dS heterogeneity across lineages is more consistent with relaxed functional constraint than with positive selection.
Finally, the pattern of MSX1 exon 1 amino acid substitution along the T. syrichta lineage is unique among primates. First, the number of substitutions is large; there is a mean of 25.7 amino acid differences between T. syrichta and other primates, compared to a mean of 17.8 differences between mouse and any primate other than T. syrichta. Second, many of the T. syrichta nonsynonymous substitutions occur in regions that are 100% conserved across other primates and with mice. Despite these unusual patterns, the rate of nonsynonymous relative to synonymous substitutions along the tarsier branch is not greater than observed in the tree as a whole.
Functional Implications
Experimental studies to characterize evolutionarily and functionally important regulatory regions will benefit from comparative analyses such as ours that highlight different evolutionary classes of conservation in noncoding sequences. PAX9 CNS 2 is conserved across primates and with dogs, suggesting that the lack of conservation with mice is due to changes on the mouse lineage. This CNS may be involved in the development of morphological features common to the dog and primate lineages but not mouse. For example, PAX9 is involved in the development of the dentition (Peters et al. 1998), and the dentition of mice, that have no canines or premolars and only one set of teeth, is highly derived compared to dogs and most primates (Tucker and Sharpe 2004).
Among primates, the tarsier lineage is one of the most phenotypically derived, including unique features in the craniofacial skeleton, dentition, and limbs (Fleagle 1999), all organ systems in which MSX1 plays a key developmental role. The unusual T. syrichta MSX1 exon 1 substitution pattern may be related to tarsier phenotypic evolution. In mice, MSX1 is expressed in the incisor regions during the earliest dental development stages, and homozygous knockouts do not form any teeth, with development arrested at an earlier stage in incisors than in molars. These mice also have abnormal palates, mandibles, maxillas, and other bones of the craniofacial skeleton, as well as defects in nail development (Satokata and Maas 1994; Jumlongras et al. 2001). Tarsiers have a reduced number of lower permanent incisors and are the only extant primate lineage with no deciduous incisors (Swindler 2002). It is possible that one or more of the MSX1 exon 1 T. syrichta substitutions are directly involved in the derived incisor phenotype of tarsiers. One alternative explanation is that functional constraint on the MSX1 protein was relaxed following craniofacial, dental, or limb phenotype changes. Further studies are needed to establish any causative links between genotype and phenotype and to distinguish between evolutionary scenarios. Nonetheless, it is clear that this highly derived lineage has a unique pattern of molecular evolution at MSX1.
Implications for Human Disease
Levels of conservation or divergence among primate species can be used to assess the potential severity of human mutations. Several heterozygous (autosomal dominant) MSX1 coding region mutations in humans are associated with oligodontia (OMIM 604625
[OMIM]
), the absence of six or more permanent teeth, excluding third molars. Some of these mutations, along with others that do not expressly result in tooth agenesis, are also associated with cleft lip with or without cleft palate (OMIM 119530
[OMIM]
) or cleft palate only (OMIM 119540
[OMIM]
). One mutation has been associated with tooth and nail syndrome, including both nail dysplasia and oligodontia (OMIM 189500
[OMIM]
).
All but one of the MSX1 oligodontia-associated mutations occur within or disrupt (i.e., gene deletion, frameshift mutation, or premature stop codon) the homeobox-binding domain (Vastardis et al. 1996; van den Boogaard et al. 2000; Jumlongras et al. 2001; Nieminen et al. 2003; De Muynck et al. 2004), which is 100% conserved across all the primates we sampled. The one nonhomeobox nonsynonymous mutation associated with oligodontia occurs in MSX1 exon 1 (Lidral and Reising 2002, see fig. 4). This mutation occurs within a region that is 100% conserved at the amino acid level across all primates (including T. syrichta) and mice.
In contrast to the MSX1 oligodontia-associated mutations, of which all but one occur in exon 2, nearly all of the MSX1 mutations that have been linked to cleft lip and/or cleft palate but not oligodontia are found in either exon 1 or in noncoding regions (Lidral et al. 1998; Jezewski et al. 2003; Suzuki et al. 2004, see fig. 4). For each of these cases, there is 100% conservation at the position of interest across the primates included in our study. Jezewski et al. (2003) also identified two mutations within the MIHR region of MSX1 intron 1 that may be involved in cleft palate. While one of these positions (451 + 887G > T) is 100% conserved across the primates we sequenced, the other (451 + 1046C > T) is not as all strepsirrhines but P. potto were observed to have the same nucleotide at this position (T) as the cleft palate patient.
Heterozygous PAX9 mutations have also been associated with human oligodontia. All but one of these mutations (Frazier-Bowers et al. 2002) occur within or disrupt the paired box–binding domain of PAX9 (Stockton et al. 2000; Nieminen et al. 2001; Das et al. 2002, 2003; Lammi et al. 2003; Mostowska et al. 2003; Jumlongras et al. 2004; Klein et al. 2005). Again, the binding domain of PAX9 was 100% conserved across all primates in our study. The only nonbinding domain mutation is a frameshift insertion (Frazier-Bowers et al. 2002) that disrupts several long regions of amino acid conservation among primates and mice (Supplementary Figure 1, Supplementary Material online). In a recent case/control study, Peres et al. (2005) identified an association between two upstream PAX9 mutations and the agenesis of fewer than six permanent teeth. In many of the agenesis individuals, only third molars were absent. The region containing these mutations was sequenced in our study in chimpanzee, rhesus macaque (Macaca mulatta), and five New World monkey species, three of which do not have permanent third molars (Swindler 2002). The nonagenesis-associated nucleotides were 100% conserved across these species.
|
Conclusion |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionConclusionSupplementary MaterialAcknowledgementsReferences |
|---|
As our understanding of regulatory elements and their function advances, comparative analyses of developmental genes will become increasingly valuable for studies of phenotypic variation and evolution. Molecular evolutionary analyses of PAX9 and MSX1 across a diverse range of primates reveal lineage- and gene-specific patterns of variation. Together with our comparative species data, the locations of human disease mutations at MSX1 and PAX9 highlight specific gene regions that reflect strong functional constraint and a history of purifying selection. It is possible that, during primate evolution, PAX9 and MSX1 disease-related phenotypes conferred a fitness reduction and were removed by natural selection.
|
Supplementary Material |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionConclusionSupplementary MaterialAcknowledgementsReferences |
|---|
Supplementary Figure 1 and Supplementary Table 1 are available at Molecular Biology and Evolution online (http://www.mbe.oxfordjournals.org/).
Supplementary Figure 1—PAX9 deduced amino acid sequences.
Supplementary Table 1—PCR primers and conditions.
|
Acknowledgements |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionConclusionSupplementary MaterialAcknowledgementsReferences |
|---|
We thank Nathaniel Dominy, William Kimbel, Peter Jezewski, Jeffrey Murray, and Gary Schwartz for helpful discussions and comments pertaining to this work. Several anonymous reviewers provided critical feedback that helped to improve the manuscript. This study would not have been possible without the samples generously provided by Duke University Primate Center, Integrated Primate Biomaterials and Information Resource (http://www.ipbir.org), New Iberia Research Center, H. Vasken Aposhian, and Jennifer Pastorini. This work was supported by a Sigma Xi Grant-in-Aid of Research, an Arizona State University Department of Anthropology Research and Development Grant, and by National Institutes of Health – National Center for Research Resources Grant No. 3 U42 RR015087-05S1 to the University of Louisiana at Lafayette New Iberia Research Center.
|
Footnotes |
|---|
John McDonald, Associate Editor
|
References |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionConclusionSupplementary MaterialAcknowledgementsReferences |
|---|
Bates, M. D., J. M. Wells, and B. Venkatesh. 2005. Comparative genomics of the Hlx homeobox gene and protein: conservation of structure and expression from fish to mammals. Gene 352:45–56.[CrossRef][Web of Science][Medline]
Belting, H. G., C. S. Shashikant, and F. H. Ruddle. 1998. Modification of expression and cis-regulation of Hoxc8 in the evolution of diverged axial morphology. Proc. Natl. Acad. Sci. USA 95:2355–2360.
Berdal, A., F. Lezot, L. Pibouin, D. Hotton, S. Ghoul-Mazgar, C. Teillaud, B. Robert, M. MacDougall, and C. Blin. 2002. Msx1 homeogene antisense mRNA in mouse dental and bone cells. Connect. Tissue Res. 43:148–152.[Web of Science][Medline]
Blin-Wakkach, C., F. Lezot, S. Ghoul-Mazgar et al. (9 co-authors) 2001. Endogenous Msx1 antisense transcript: in vivo and in vitro evidences, structure, and potential involvement in skeleton development in mammals. Proc. Natl. Acad. Sci. USA 98:7336–7341.
Boffelli, D., J. McAuliffe, D. Ovcharenko, K. D. Lewis, I. Ovcharenko, L. Pachter, and E. M. Rubin. 2003. Phylogenetic shadowing of primate sequences to find functional regions of the human genome. Science 299:1391–1394.
Das, P., M. Hai, C. Elcock, S. M. Leal, D. T. Brown, A. H. Brook, and P. I. Patel. 2003. Novel missense mutations and a 288-bp exonic insertion in PAX9 in families with autosomal dominant hypodontia. Am. J. Med. Genet. 118A:35–42.
Das, P., D. W. Stockton, C. Bauer, L. G. Shaffer, R. N. D'Souza, T. Wright, and P. I. Patel. 2002. Haploinsufficiency of PAX9 is associated with autosomal dominant hypodontia. Hum. Genet. 110:371–376.[CrossRef][Web of Science][Medline]
De Muynck, S., E. Schollen, G. Matthijs, A. Verdonck, K. Devriendt, and C. Carels. 2004. A novel MSX1 mutation in hypodontia. Am. J. Med. Genet. 128A:401–403.[CrossRef]
Eizinger, A., B. Jungblut, and R. J. Sommer. 1999. Evolutionary change in the functional specificity of genes. Trends Genet. 15:197–202.[CrossRef][Web of Science][Medline]
Fleagle, J. G. 1999. Primate adaptation and evolution. Academic Press, San Diego.
Frazier-Bowers, S. A., D. C. Guo, A. Cavender, L. Xue, B. Evans, T. King, D. Milewicz, and R. N. D'Souza. 2002. A novel mutation in human PAX9 causes molar oligodontia. J. Dent. Res. 81:129–133.
Goldman, N., and Z. Yang. 1994. A codon-based model of nucleotide substitution for protein-coding DNA sequences. Mol. Biol. Evol. 11:725–736.[Abstract]
Gompel, N., B. Prud'homme, P. J. Wittkopp, V. A. Kassner, and S. B. Carroll. 2005. Chance caught on the wing: cis-regulatory evolution and the origin of pigment patterns in Drosophila. Nature 433:481–487.[CrossRef][Web of Science][Medline]
Goodman, M., C. A. Porter, J. Czelusniak, S. L. Page, H. Schneider, J. Shoshani, G. Gunnell, and C. P. Groves. 1998. Toward a phylogenetic classification of primates based on DNA evidence complemented by fossil evidence. Mol. Phylogenet. Evol. 9:585–598.[CrossRef][Web of Science][Medline]
Hahn, M. W., M. V. Rockman, N. Soranzo, D. B. Goldstein, and G. A. Wray. 2004. Population genetic and phylogenetic evidence for positive selection on regulatory mutations at the factor VII locus in humans. Genetics 167:867–877.
Jezewski, P. A., A. R. Vieira, C. Nishimura et al. (14 co-authors). 2003. Complete sequencing shows a role for MSX1 in non-syndromic cleft lip and palate. J. Med. Genet. 40:399–407.
Jonker, L., R. Kist, A. Aw, I. Wappler, and H. Peters. 2004. Pax9 is required for filiform papilla development and suppresses skin-specific differentiation of the mammalian tongue epithelium. Mech. Dev. 121:1313–1322.[CrossRef][Web of Science][Medline]
Jumlongras, D., M. Bei, J. M. Stimson, W. F. Wang, S. R. DePalma, C. E. Seidman, U. Felbor, R. Maas, J. G. Seidman, and B. R. Olsen. 2001. A nonsense mutation in MSX1 causes Witkop syndrome. Am. J. Hum. Genet. 69:67–74.[CrossRef][Web of Science][Medline]
Jumlongras, D., J. Y. Lin, A. Chapra, C. E. Seidman, J. G. Seidman, R. L. Maas, and B. R. Olsen. 2004. A novel missense mutation in the paired domain of PAX9 causes non-syndromic oligodontia. Hum. Genet. 114:242–249.[CrossRef][Web of Science][Medline]
Kappen, C., and F. H. Ruddle. 1993. Evolution of a regulatory gene family: HOM/HOX genes. Curr. Opin. Genet. Dev. 3:931–938.[CrossRef][Medline]
Klein, M. L., P. Nieminen, L. Lammi, E. Niebuhr, and S. Kreiborg. 2005. Novel mutation of the initiation codon of PAX9 causes oligodontia. J. Dent. Res. 84:43–47.
Kumar, S., K. Tamura, and M. Nei. 2004. MEGA3: integrated software for molecular evolutionary genetics analysis and sequence alignment. Brief. Bioinform. 5:150–163.
Lallemand, Y., M. A. Nicola, C. Ramos, A. Bach, C. S. Cloment, and B. Robert. 2005. Analysis of Msx1; Msx2 double mutants reveals multiple roles for Msx genes in limb development. Development 132:3003–3014.
Lammi, L., K. Halonen, S. Pirinen, I. Thesleff, S. Arte, and P. Nieminen. 2003. A missense mutation in PAX9 in a family with distinct phenotype of oligodontia. Eur. J. Hum. Genet. 11:866–871.[CrossRef][Web of Science][Medline]
Lee, H., R. Habas, and C. Abate-Shen. 2004. MSX1 cooperates with histone H1b for inhibition of transcription and myogenesis. Science 304:1675–1678.
Levine, M., and E. H. Davidson. 2005. Gene regulatory networks for development. Proc. Natl. Acad. Sci. USA 102:4936–4942.
Lidral, A. C., and B. C. Reising. 2002. The role of MSX1 in human tooth agenesis. J. Dent. Res. 81:274–278.
Lidral, A. C., P. A. Romitti, A. M. Basart et al. (10 co-authors). 1998. Association of MSX1 and TGFB3 with nonsyndromic clefting in humans. Am. J. Hum. Genet. 63:557–568.[CrossRef][Web of Science][Medline]
Logan, C., M. C. Hanks, S. Noble-Topham, D. Nallainathan, N. J. Provart, and A. L. Joyner. 1992. Cloning and sequence comparison of the mouse, human, and chicken engrailed genes reveal potential functional domains and regulatory regions. Dev. Genet. 13:345–358.[CrossRef][Web of Science][Medline]
Majewski, J., and J. Ott. 2002. Distribution and characterization of regulatory elements in the human genome. Genome Res. 12:1827–1836.
Mostowska, A., A. Kobielak, B. Biedziak, and W. H. Trzeciak. 2003. Novel mutation in the paired box sequence of PAX9 gene in a sporadic form of oligodontia. Eur. J. Oral Sci. 111:272–276.[CrossRef][Web of Science][Medline]
Nieminen, P., S. Arte, D. Tanner, L. Paulin, S. Alaluusua, I. Thesleff, and S. Pirinen. 2001. Identification of a nonsense mutation in the PAX9 gene in molar oligodontia. Eur. J. Hum. Genet. 9:743–746.[CrossRef][Web of Science][Medline]
Nieminen, P., J. Kotilainen, Y. Aalto, S. Knuutila, S. Pirinen, and I. Thesleff. 2003. MSX1 gene is deleted in Wolf-Hirschhorn syndrome patients with oligodontia. J. Dent. Res. 82:1013–1017.
Ovcharenko, I., D. Boffelli, and G. G. Loots. 2004. eShadow: a tool for comparing closely related sequences. Genome Res. 14:1191–1198.
Pagani, F., M. Raponi, and F. E. Baralle. 2005. Synonymous mutations in CFTR exon 12 affect splicing and are not neutral in evolution. Proc. Natl. Acad. Sci. USA 102:6368–6372.
Pastorini, J., M. R. Forstner, R. D. Martin, and D. J. Melnick. 1998. A reexamination of the phylogenetic position of Callimico (primates) incorporating new mitochondrial DNA sequence data. J. Mol. Evol. 47:32–41.[CrossRef][Web of Science][Medline]
Pastorini, J., U. Thalmann, and R. D. Martin. 2003. A molecular approach to comparative phylogeography of extant Malagasy lemurs. Proc. Natl. Acad. Sci. USA 100:5879–5884.
Peres, R. C., R. M. Scarel-Caminaga, A. R. do Espirito Santo, and S. R. Line. 2005. Association between PAX-9 promoter polymorphisms and hypodontia in humans. Arch. Oral Biol. 50:861–871.[CrossRef][Web of Science][Medline]
Peters, H., A. Neubuser, K. Kratochwil, and R. Balling. 1998. Pax9-deficient mice lack pharyngeal pouch derivatives and teeth and exhibit craniofacial and limb abnormalities. Genes Dev. 12:2735–2747.
Peters, H., B. Wilm, N. Sakai, K. Imai, R. Maas, and R. Balling. 1999. Pax1 and Pax9 synergistically regulate vertebral column development. Development 126:5399–5408.[Abstract]
Roos, C., J. Schmitz, and H. Zischler. 2004. Primate jumping genes elucidate strepsirrhine phylogeny. Proc. Natl. Acad. Sci. USA 101:10650–10654.
Sambrook, J., and D. W. Russell. 2001. Molecular cloning: a laboratory manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y.
Santini, S., J. L. Boore, and A. Meyer. 2003. Evolutionary conservation of regulatory elements in vertebrate Hox gene clusters. Genome Res. 13:1111–1122.
Satokata, I., and R. Maas. 1994. Msx1 deficient mice exhibit cleft palate and abnormalities of craniofacial and tooth development. Nat. Genet. 6:348–356.[CrossRef][Web of Science][Medline]
Schwartz, S., L. Elnitski, M. Li, M. Weirauch, C. Riemer, A. Smit, E. D. Green, R. C. Hardison, and W. Miller. 2003. MultiPipMaker and supporting tools: alignments and analysis of multiple genomic DNA sequences. Nucleic Acids Res. 31:3518–3524.
Shah, N., O. Couronne, L. A. Pennacchio, M. Brudno, S. Batzoglou, E. W. Bethel, E. M. Rubin, B. Hamann, and I. Dubchak. 2004. Phylo-VISTA: interactive visualization of multiple DNA sequence alignments. Bioinformatics 20:636–643.
Stockton, D. W., P. Das, M. Goldenberg, R. N. D'Souza, and P. I. Patel. 2000. Mutation of PAX9 is associated with oligodontia. Nat. Genet. 24:18–19.[CrossRef][Web of Science][Medline]
Suzuki, Y., P. A. Jezewski, J. Machida et al. (10 co-authors). 2004. In a Vietnamese population, MSX1 variants contribute to cleft lip and palate. Genet. Med. 6:117–125.[Web of Science][Medline]
Swindler, D. R. 2002. Primate dentition: an introduction to the teeth of non-human primates. Cambridge University Press, Cambridge.
Tucker, A., and P. Sharpe. 2004. The cutting-edge of mammalian development; how the embryo makes teeth. Nat. Rev. Genet. 5:499–508.[Web of Science][Medline]
van den Boogaard, M. J., M. Dorland, F. A. Beemer, and H. K. van Amstel. 2000. MSX1 mutation is associated with orofacial clefting and tooth agenesis in humans. Nat. Genet. 24:342–343.[CrossRef][Web of Science][Medline]
Vastardis, H., N. Karimbux, S. W. Guthua, J. G. Seidman, and C. E. Seidman. 1996. A human MSX1 homeodomain missense mutation causes selective tooth agenesis. Nat. Genet. 13:417–421.[CrossRef][Web of Science][Medline]
Waterston, R. H., K. Lindblad-Toh, E. Birney et al. (219 co-authors). 2002. Initial sequencing and comparative analysis of the mouse genome. Nature 420:520–562.[CrossRef][Medline]
Woolfe, A., M. Goodson, D. K. Goode et al. (13 co-authors). 2005. Highly conserved non-coding sequences are associated with vertebrate development. PLoS Biol. 3:e7.[CrossRef][Medline]
Yang, Z. 1997. PAML: a program package for phylogenetic analysis by maximum likelihood. Comput. Appl. Biosci. 13:555–556.
———. 1998. Likelihood ratio tests for detecting positive selection and application to primate lysozyme evolution. Mol. Biol. Evol. 15:568–573.[Abstract]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  H. Takahashi, A. Kamiya, A. Ishiguro, A. C. Suzuki, N. Saitou, A. Toyoda, and J. Aruga Conservation and Diversification of Msx Protein in Metazoan Evolution Mol. Biol. Evol., January 1, 2008; 25(1): 69 - 82. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||