Genetics Driving Epigenetics Terrence S. Furey and Praveen Sethupathy Science 342, 705 (2013); DOI: 10.1126/science.1246755
If you wish to distribute this article to others, you can order high-quality copies for your colleagues, clients, or customers by clicking here. Permission to republish or repurpose articles or portions of articles can be obtained by following the guidelines here. The following resources related to this article are available online at www.sciencemag.org (this information is current as of November 9, 2013 ): Updated information and services, including high-resolution figures, can be found in the online version of this article at: http://www.sciencemag.org/content/342/6159/705.full.html A list of selected additional articles on the Science Web sites related to this article can be found at: http://www.sciencemag.org/content/342/6159/705.full.html#related This article cites 9 articles, 3 of which can be accessed free: http://www.sciencemag.org/content/342/6159/705.full.html#ref-list-1
Science (print ISSN 0036-8075; online ISSN 1095-9203) is published weekly, except the last week in December, by the American Association for the Advancement of Science, 1200 New York Avenue NW, Washington, DC 20005. Copyright 2013 by the American Association for the Advancement of Science; all rights reserved. The title Science is a registered trademark of AAAS.
Downloaded from www.sciencemag.org on November 10, 2013
This copy is for your personal, non-commercial use only.
PERSPECTIVES GENETICS
DNA sequence variants are associated with factors and epigenetic states that modify gene expression and contribute to quantitative differences in traits.
Genetics Driving Epigenetics Terrence S. Furey1,2,3,4 and Praveen Sethupathy 2,3,4
CREDIT: V. ALTOUNIAN/SCIENCE
H
umans vary according to a plethora of traits, such as height, hair color, behavior, and susceptibility to disease. Both genetics (nature) and environment (nurture) contribute to this variation. Recent large-scale genetic studies have identified thousands of specific DNA variations in the human population that are associated with different traits. However, these studies do not answer a key question: By what means do most DNA variants alter cellular behavior and contribute to differences in specific traits, such as height? A trio of papers in this issue by Kasowski et al. on page 750 (1), Kilpinen et al. on page 744 (2), and McVicker et al. on page 747 (3) provide a framework for exploring the mechanistic link between genetic and trait variation in the human population. Specifically, they find that DNA variants influence a layer of gene regulation called epigenetics through the sequence-specific activity of transcription factors. One of the most important discoveries in genetics in the last 10 years is that the vast majority of trait-associated DNA variations occur in regions of the genome that were once labeled as “junk DNA” because they do not code for proteins. We now know that these regions harbor genetic elements that control where, when, and to what extent specific genes are expressed to make functional RNA and protein products. Therefore, most trait-associated DNA variants are thought to alter not the gene itself, but rather, the regulatory elements that control the process of gene expression. In the last 3 years, several generegulatory variants have been strongly implicated in traits such as blood cholesterol concentrations (4) and diseases such as diabetes (5, 6), osteoarthritis (7), and prostate cancer (8). Despite these advances, precisely how most regulatory variants alter gene expression has been poorly understood. Epigenetic mechanisms are known to control heritable gene expression but have been generally viewed as independent of
the underlying DNA sequence. DNA is packaged in a three-dimensional structure, chromatin, whose basic repeating unit, the nucleosome, consists of ~146 nucleotides of DNA wrapped around an octamer of specialized proteins called histones. Each of the eight histone proteins has amino acid “tails” that stick out from the nucleosome. Specific amino acids in these tails are subject to a vast array of chemical modifications, such as methylation, acetylation, or phosphorylation, which are carried out by a variety of nuclear enzymes. The “histone code hypothesis” (9) proposes that specific combinations of histone tail modifications (epigenetic marks) are associated with transcription factors
1
Human trait variation. (A) Differences in a human trait (such as height) are partly due to the combined effects of genetic variants that alter the expression of multiple genes. (B) At a specific genomic position, a nucleotide [such as adenine (A)] is associated with accessible DNA, which facilitates transcription factor binding. This step leads to histone tail modifications that promote a chromatin environment favorable for the expression of neighboring genes. (C) At the same genomic position, a nucleotide variant [such as guanine (G)] has low affinity for transcription factor binding, which leads to a chromatin environment unfavorable for gene expression. Pol II, polymerase II
Department of Biology, University of North Carolina, Chapel Hill, NC 27599, USA. 2Department of Genetics, University of North Carolina, Chapel Hill, NC 27599, USA. 3Carolina Center for Genome Sciences, University of North Carolina, Chapel Hill, NC 27599, USA. 4Lineberger Comprehensive Cancer Center, University of North Carolina, Chapel Hill, NC 27599, USA. E-mail: [email protected]; [email protected]
that increase or decrease gene expression. Kasowski et al., Kilpinen et al., and McVicker et al. perform integrative analysis of diverse data types generated from lymphoblastoid cell lines across numerous individuals and family trios. They demonstrate that histone tail modifications are highly variable in the human population and that they are heritable across generations. The studies identified hundreds of DNA variants that are associated with both histone tail modification and gene expression variation, indicating that genetics coordinates epigenetic effects on gene regulation. But how does variation in the DNA sequence influence chemical modification at histone tails?
Tall
Short
A Amino acid tails Histone DNA
Gene
B
Histone-modifying enzyme Transcription factor Methyl group Pol II G
A
A
C
T
G
T
C
C
T
T
G
A
C
A
G
A
Tall
Gene expression
Acetyl group
C
Short
G
No gene expression
G
A
A
C
T
G
T
C
C
T
T
G
G
C
A
G
www.sciencemag.org SCIENCE VOL 342 8 NOVEMBER 2013 Published by AAAS
705
PERSPECTIVES The three studies point to transcription factor activity as the missing link. Most transcription factors bind directly to DNA, each with a preference for a particular DNA sequence pattern. Some DNA variants can substantially alter transcription factor binding affinity at particular genomic locations and thereby influence gene transcription. Kasowski et al., Kilpinen et al., and McVicker et al. used computationally predicted and empirically determined transcription factor binding data to identify hundreds of DNA variants that affect the strength of transcription factor binding. Many of these variants were also associated with variation in histone tail modifications. These findings suggest a mechanism in which transcription factor binding to DNA initiates the recruitment of histone-modifying enzymes that set the histone tail modification pattern. Thus, a possible model (see the figure) is that trait-associated variants, most of which are gene regulatory in nature, affect the recruitment and binding of transcription factors to DNA. Differential transcription factor binding leads to variable histone tail modifications that collectively influence gene expression. Gene expression variations can manifest as trait differences.
Among some of the distinct findings of the studies, McVicker et al. showed that a single DNA variant can influence histone modifications at multiple related regions in the genome, providing information on their functional relationships. Kasowski et al. found that an individual’s ancestry can affect what genomic regions exhibit genetically driven variability in chromatin marks. Kilpinen et al. noted that coordinated effects of DNA variation extend beyond transcription factor binding and histone tail modifications to other aspects of gene regulation, such as rate of transcription. Interestingly, all three studies found that many of the DNA variants associated with both transcription factor binding and histone tail modification variability were not associated with gene expression variability. This suggests that there is an abundance of nonconsequential regulatory variation, and/ or that there are widespread mechanisms to compensate for the effects of regulatory variation, and/or that the regulatory effects of some transcription factor binding events are evident only under specific environmental conditions that are not well captured in cell culture. The studies of Kasowski et al., Kilpinen et al., and McVicker et al. provide new insight into genetic mechanisms that affect
complex traits and disease, and also elucidate basic gene-regulatory processes, but by no means is either of these problems solved. For example, as the authors of these three studies emphasize, not every regulatory variant will lead to trait differences or even gene expression differences. Why are some regulatory variants more critical than others for trait variability or disease risk? There is much more to uncover to answer this and related questions, but these studies bring us one step closer and provide a framework for exploring this topic further. References and Notes 1. M. Kasowski et al., Science 342, 750 (2013); 10.1126/ science.1242510. 2. H. Kilpinen et al., Science 342, 744 (2013); 10.1126/ science.1242463. 3. G. McVicker et al., Science 342, 747 (2013); 10.1126/ science.1242429. 4. K. Musunuru et al., Nature 466, 714 (2010). 5. M. L. Stitzel et al., Cell Metab. 12, 443 (2010). 6. K. J. Gaulton et al., Nat. Genet. 42, 255 (2010). 7. A. W. Dodd, C. M. Syddall, J. Loughlin, Eur. J. Hum. Genet. 21, 517 (2013). 8. M. M. Pomerantz et al., Nat. Genet. 41, 882 (2009). 9. B. D. Strahl, C. D. Allis, Nature 403, 41 (2000).
Acknowledgments: We thank S. Kelada, M. Deshmukh, P. Rajasethupathy, M. Stitzel, and A. Laederach for critical reading and discussion. 10.1126/science.1246755
BIOCHEMISTRY
Have Your PIC!
Structural analyses help to elucidate a key step in DNA transcription.
Sohail Malik and Robert G. Roeder
R
NA polymerase II (Pol II), the enzyme that transcribes protein-coding genes, requires additional factors to accurately initiate transcription from promoter-directed start sites. These general transcription factors (GTFs) assemble with Pol II into a pre-initiation complex (PIC) that is a key intermediate in the transcription activation pathway (1). The main role of the GTFs is to accurately position and orient Pol II on the DNA template and to facilitate access of the catalytic site to the transcribed strand. Recent structural studies, including the cryoelectron microscopy (cryo-EM) reconstruction of the yeast PIC by Murakami et al. on page 709 of this issue (2), provide detailed views into the structural organization of this large complex. These studies will pave the way for a more comLaboratory of Biochemistry and Molecular Biology, Rockefeller University, 1230 York Avenue, New York, NY 10065, USA. E-mail: [email protected]; [email protected]
706
plete understanding of how gene expression is regulated. Biochemical studies have shown that the PIC initially forms on double-stranded promoter DNA (3). Prior to transcription of the body of the gene by Pol II, the PIC undergoes two critical transitions. First, it isomerizes into an open complex in which the promoter DNA exists in a partially melted state (“bubble”) to provide the enzyme’s active site access to the initiating nucleotide on the template. Next, in what is termed promoter escape, the melted region is extended in a discontinuous fashion to allow templated RNA chain synthesis to occur. This initially entails generation of short, abortive RNA products. Successful escape occurs when a stable RNA-DNA hybrid is established. Murakami et al.’s structure shows that the PIC has two distinct lobes, reminiscent of the ribosome with which the PIC is often compared. The P-lobe is almost exclusively formed from Pol II subunits, whereas the
G-lobe comprises the various GTFs. Higherresolution views reveal both the layout of individual subunits and the path of the template DNA, which, while sandwiched between the two lobes, is mainly associated with the G-lobe. This arrangement is surprising; one might have expected from earlier biochemical studies that the DNA, Pol II, and GTFs would be intertwined in the PIC. The GTFs include TFIIA, TFIIB, TFIID, TFIIE, TFIIF, and TFIIH (see the figure, panel A). Consistent with previous studies (4–6), Murakami et al.’s structure (see the figure, panel B) shows how these individual GTFs are arrayed with respect to the template and Pol II, poised for their roles in the various transitions that the PIC will undergo. Murakami et al. do not present a structure for the open state, but their PIC reconstruction suggests hypotheses for how Pol II engages the template strand after transfer of the DNA from the G-lobe to the Pol II cleft, at the bottom of which lies the catalytic cen-
8 NOVEMBER 2013 VOL 342 SCIENCE www.sciencemag.org Published by AAAS
If you wish to distribute this article to others, you can order high-quality copies for your colleagues, clients, or customers by clicking here. Permission to republish or repurpose articles or portions of articles can be obtained by following the guidelines here. The following resources related to this article are available online at www.sciencemag.org (this information is current as of November 9, 2013 ): Updated information and services, including high-resolution figures, can be found in the online version of this article at: http://www.sciencemag.org/content/342/6159/705.full.html A list of selected additional articles on the Science Web sites related to this article can be found at: http://www.sciencemag.org/content/342/6159/705.full.html#related This article cites 9 articles, 3 of which can be accessed free: http://www.sciencemag.org/content/342/6159/705.full.html#ref-list-1
Science (print ISSN 0036-8075; online ISSN 1095-9203) is published weekly, except the last week in December, by the American Association for the Advancement of Science, 1200 New York Avenue NW, Washington, DC 20005. Copyright 2013 by the American Association for the Advancement of Science; all rights reserved. The title Science is a registered trademark of AAAS.
Downloaded from www.sciencemag.org on November 10, 2013
This copy is for your personal, non-commercial use only.
PERSPECTIVES GENETICS
DNA sequence variants are associated with factors and epigenetic states that modify gene expression and contribute to quantitative differences in traits.
Genetics Driving Epigenetics Terrence S. Furey1,2,3,4 and Praveen Sethupathy 2,3,4
CREDIT: V. ALTOUNIAN/SCIENCE
H
umans vary according to a plethora of traits, such as height, hair color, behavior, and susceptibility to disease. Both genetics (nature) and environment (nurture) contribute to this variation. Recent large-scale genetic studies have identified thousands of specific DNA variations in the human population that are associated with different traits. However, these studies do not answer a key question: By what means do most DNA variants alter cellular behavior and contribute to differences in specific traits, such as height? A trio of papers in this issue by Kasowski et al. on page 750 (1), Kilpinen et al. on page 744 (2), and McVicker et al. on page 747 (3) provide a framework for exploring the mechanistic link between genetic and trait variation in the human population. Specifically, they find that DNA variants influence a layer of gene regulation called epigenetics through the sequence-specific activity of transcription factors. One of the most important discoveries in genetics in the last 10 years is that the vast majority of trait-associated DNA variations occur in regions of the genome that were once labeled as “junk DNA” because they do not code for proteins. We now know that these regions harbor genetic elements that control where, when, and to what extent specific genes are expressed to make functional RNA and protein products. Therefore, most trait-associated DNA variants are thought to alter not the gene itself, but rather, the regulatory elements that control the process of gene expression. In the last 3 years, several generegulatory variants have been strongly implicated in traits such as blood cholesterol concentrations (4) and diseases such as diabetes (5, 6), osteoarthritis (7), and prostate cancer (8). Despite these advances, precisely how most regulatory variants alter gene expression has been poorly understood. Epigenetic mechanisms are known to control heritable gene expression but have been generally viewed as independent of
the underlying DNA sequence. DNA is packaged in a three-dimensional structure, chromatin, whose basic repeating unit, the nucleosome, consists of ~146 nucleotides of DNA wrapped around an octamer of specialized proteins called histones. Each of the eight histone proteins has amino acid “tails” that stick out from the nucleosome. Specific amino acids in these tails are subject to a vast array of chemical modifications, such as methylation, acetylation, or phosphorylation, which are carried out by a variety of nuclear enzymes. The “histone code hypothesis” (9) proposes that specific combinations of histone tail modifications (epigenetic marks) are associated with transcription factors
1
Human trait variation. (A) Differences in a human trait (such as height) are partly due to the combined effects of genetic variants that alter the expression of multiple genes. (B) At a specific genomic position, a nucleotide [such as adenine (A)] is associated with accessible DNA, which facilitates transcription factor binding. This step leads to histone tail modifications that promote a chromatin environment favorable for the expression of neighboring genes. (C) At the same genomic position, a nucleotide variant [such as guanine (G)] has low affinity for transcription factor binding, which leads to a chromatin environment unfavorable for gene expression. Pol II, polymerase II
Department of Biology, University of North Carolina, Chapel Hill, NC 27599, USA. 2Department of Genetics, University of North Carolina, Chapel Hill, NC 27599, USA. 3Carolina Center for Genome Sciences, University of North Carolina, Chapel Hill, NC 27599, USA. 4Lineberger Comprehensive Cancer Center, University of North Carolina, Chapel Hill, NC 27599, USA. E-mail: [email protected]; [email protected]
that increase or decrease gene expression. Kasowski et al., Kilpinen et al., and McVicker et al. perform integrative analysis of diverse data types generated from lymphoblastoid cell lines across numerous individuals and family trios. They demonstrate that histone tail modifications are highly variable in the human population and that they are heritable across generations. The studies identified hundreds of DNA variants that are associated with both histone tail modification and gene expression variation, indicating that genetics coordinates epigenetic effects on gene regulation. But how does variation in the DNA sequence influence chemical modification at histone tails?
Tall
Short
A Amino acid tails Histone DNA
Gene
B
Histone-modifying enzyme Transcription factor Methyl group Pol II G
A
A
C
T
G
T
C
C
T
T
G
A
C
A
G
A
Tall
Gene expression
Acetyl group
C
Short
G
No gene expression
G
A
A
C
T
G
T
C
C
T
T
G
G
C
A
G
www.sciencemag.org SCIENCE VOL 342 8 NOVEMBER 2013 Published by AAAS
705
PERSPECTIVES The three studies point to transcription factor activity as the missing link. Most transcription factors bind directly to DNA, each with a preference for a particular DNA sequence pattern. Some DNA variants can substantially alter transcription factor binding affinity at particular genomic locations and thereby influence gene transcription. Kasowski et al., Kilpinen et al., and McVicker et al. used computationally predicted and empirically determined transcription factor binding data to identify hundreds of DNA variants that affect the strength of transcription factor binding. Many of these variants were also associated with variation in histone tail modifications. These findings suggest a mechanism in which transcription factor binding to DNA initiates the recruitment of histone-modifying enzymes that set the histone tail modification pattern. Thus, a possible model (see the figure) is that trait-associated variants, most of which are gene regulatory in nature, affect the recruitment and binding of transcription factors to DNA. Differential transcription factor binding leads to variable histone tail modifications that collectively influence gene expression. Gene expression variations can manifest as trait differences.
Among some of the distinct findings of the studies, McVicker et al. showed that a single DNA variant can influence histone modifications at multiple related regions in the genome, providing information on their functional relationships. Kasowski et al. found that an individual’s ancestry can affect what genomic regions exhibit genetically driven variability in chromatin marks. Kilpinen et al. noted that coordinated effects of DNA variation extend beyond transcription factor binding and histone tail modifications to other aspects of gene regulation, such as rate of transcription. Interestingly, all three studies found that many of the DNA variants associated with both transcription factor binding and histone tail modification variability were not associated with gene expression variability. This suggests that there is an abundance of nonconsequential regulatory variation, and/ or that there are widespread mechanisms to compensate for the effects of regulatory variation, and/or that the regulatory effects of some transcription factor binding events are evident only under specific environmental conditions that are not well captured in cell culture. The studies of Kasowski et al., Kilpinen et al., and McVicker et al. provide new insight into genetic mechanisms that affect
complex traits and disease, and also elucidate basic gene-regulatory processes, but by no means is either of these problems solved. For example, as the authors of these three studies emphasize, not every regulatory variant will lead to trait differences or even gene expression differences. Why are some regulatory variants more critical than others for trait variability or disease risk? There is much more to uncover to answer this and related questions, but these studies bring us one step closer and provide a framework for exploring this topic further. References and Notes 1. M. Kasowski et al., Science 342, 750 (2013); 10.1126/ science.1242510. 2. H. Kilpinen et al., Science 342, 744 (2013); 10.1126/ science.1242463. 3. G. McVicker et al., Science 342, 747 (2013); 10.1126/ science.1242429. 4. K. Musunuru et al., Nature 466, 714 (2010). 5. M. L. Stitzel et al., Cell Metab. 12, 443 (2010). 6. K. J. Gaulton et al., Nat. Genet. 42, 255 (2010). 7. A. W. Dodd, C. M. Syddall, J. Loughlin, Eur. J. Hum. Genet. 21, 517 (2013). 8. M. M. Pomerantz et al., Nat. Genet. 41, 882 (2009). 9. B. D. Strahl, C. D. Allis, Nature 403, 41 (2000).
Acknowledgments: We thank S. Kelada, M. Deshmukh, P. Rajasethupathy, M. Stitzel, and A. Laederach for critical reading and discussion. 10.1126/science.1246755
BIOCHEMISTRY
Have Your PIC!
Structural analyses help to elucidate a key step in DNA transcription.
Sohail Malik and Robert G. Roeder
R
NA polymerase II (Pol II), the enzyme that transcribes protein-coding genes, requires additional factors to accurately initiate transcription from promoter-directed start sites. These general transcription factors (GTFs) assemble with Pol II into a pre-initiation complex (PIC) that is a key intermediate in the transcription activation pathway (1). The main role of the GTFs is to accurately position and orient Pol II on the DNA template and to facilitate access of the catalytic site to the transcribed strand. Recent structural studies, including the cryoelectron microscopy (cryo-EM) reconstruction of the yeast PIC by Murakami et al. on page 709 of this issue (2), provide detailed views into the structural organization of this large complex. These studies will pave the way for a more comLaboratory of Biochemistry and Molecular Biology, Rockefeller University, 1230 York Avenue, New York, NY 10065, USA. E-mail: [email protected]; [email protected]
706
plete understanding of how gene expression is regulated. Biochemical studies have shown that the PIC initially forms on double-stranded promoter DNA (3). Prior to transcription of the body of the gene by Pol II, the PIC undergoes two critical transitions. First, it isomerizes into an open complex in which the promoter DNA exists in a partially melted state (“bubble”) to provide the enzyme’s active site access to the initiating nucleotide on the template. Next, in what is termed promoter escape, the melted region is extended in a discontinuous fashion to allow templated RNA chain synthesis to occur. This initially entails generation of short, abortive RNA products. Successful escape occurs when a stable RNA-DNA hybrid is established. Murakami et al.’s structure shows that the PIC has two distinct lobes, reminiscent of the ribosome with which the PIC is often compared. The P-lobe is almost exclusively formed from Pol II subunits, whereas the
G-lobe comprises the various GTFs. Higherresolution views reveal both the layout of individual subunits and the path of the template DNA, which, while sandwiched between the two lobes, is mainly associated with the G-lobe. This arrangement is surprising; one might have expected from earlier biochemical studies that the DNA, Pol II, and GTFs would be intertwined in the PIC. The GTFs include TFIIA, TFIIB, TFIID, TFIIE, TFIIF, and TFIIH (see the figure, panel A). Consistent with previous studies (4–6), Murakami et al.’s structure (see the figure, panel B) shows how these individual GTFs are arrayed with respect to the template and Pol II, poised for their roles in the various transitions that the PIC will undergo. Murakami et al. do not present a structure for the open state, but their PIC reconstruction suggests hypotheses for how Pol II engages the template strand after transfer of the DNA from the G-lobe to the Pol II cleft, at the bottom of which lies the catalytic cen-
8 NOVEMBER 2013 VOL 342 SCIENCE www.sciencemag.org Published by AAAS
