RNA-MEDIATED PROGRAMMING OF ACTIVE CHROMATIN
A DISSERTATION
SUBMITTED TO THE PROGRAM IN CANCER BIOLOGY
AND THE COMMITTEE ON GRADUATE STUDIES
OF STANFORD UNIVERSITY
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS
FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
Yul Wonjun Yang
August 2012
© 2012 by Yul Wonjun Yang. All Rights Reserved.
Re-distributed by Stanford University under license with the author.
This work is licensed under a Creative Commons AttributionNoncommercial 3.0 United States License.
http://creativecommons.org/licenses/by-nc/3.0/us/
This dissertation is online at: http://purl.stanford.edu/zj169pb5921
ii
I certify that I have read this dissertation and that, in my opinion, it is fully adequate
in scope and quality as a dissertation for the degree of Doctor of Philosophy.
Howard Chang, Primary Adviser
I certify that I have read this dissertation and that, in my opinion, it is fully adequate
in scope and quality as a dissertation for the degree of Doctor of Philosophy.
Paul Khavari
I certify that I have read this dissertation and that, in my opinion, it is fully adequate
in scope and quality as a dissertation for the degree of Doctor of Philosophy.
Seung Kim
I certify that I have read this dissertation and that, in my opinion, it is fully adequate
in scope and quality as a dissertation for the degree of Doctor of Philosophy.
Joanna Wysocka
Approved for the Stanford University Committee on Graduate Studies.
Patricia J. Gumport, Vice Provost Graduate Education
This signature page was generated electronically upon submission of this dissertation in
electronic format. An original signed hard copy of the signature page is on file in
University Archives.
iii
iv
ABSTRACT
Various long noncoding RNAs (lncRNAs) have been recently identified, but
their functions remain unknown. To better understand the role of lncRNAs in gene
activation, we have characterized a 3.76 kilobase highly conserved lncRNA, called
HOTTIP, found at the 5’ distal tip of the HoxA locus. HOTTIP is required for the
expression of HoxA9 to HoxA13 in human fibroblasts and chicken embryos. This
region spans a linear 40 kilobases, but forms a compact three dimensional DNA
structure by high throughput chromosome conformation capture. Knockdown of
HOTTIP causes loss of the activating histone H3K4 trimethylation mark, as well as
lost occupancy of the MLL methylase complex by chromatin immunoprecipitation
coupled with tiling microarray analysis. To cause these changes, HOTTIP binds
WDR5, a component of the MLL complex, to alter histone H3K4 trimethylation and
gene activation. Together, the characterization of HOTTIP demonstrates an example
of lncRNA-mediated epigenetic activation.
A close examination of the binding interaction between HOTTIP and WDR5
has further revealed a novel pathway of lncRNA-regulated proteolysis. HOTTIP acts
as a “molecular switch,” causing increased WDR5 protein levels by preventing
proteasomal degradation through thermodynamic stabilization post poly-ubiquitination.
Increased WDR5 deposition then causes gene activation. One HOTTIP RNA binds a
single WDR5 protein through a direct RNA-protein interaction, and RNA-mediated
stabilization requires a specific HOTTIP RNA domain in a long RNA context. Using a
small scale alanine scanning mutagenesis screen, the HOTTIP binding interface on
WDR5 has been identified as the cleft between blades 5 and 6. WDR5 mutations that
abrogate lncRNA binding cannot be stabilized by HOTTIP, and are defective in gene
activation, maintenance of histone H3K4 trimethylation, and embryonic stem cell self
renewal. By altering protein turnover, lncRNAs may be able to regulate the temporal
landscape of proteins in cells, potentially altering epigenetic states and cellular
functions.
v
ACKNOWLEDGEMENTS
First and foremost, I would like to thank my advisor, Howard Chang, for his
mentorship, his unflappable demeanor in light of “not so good data,” and countless
pages of hand drawn figures to organize experiments. Without his guidance, none of
the experiments in this thesis could have occurred. In addition, I am deeply indebted to
his patience and understanding for my growth as an individual. Thanks to him, I leave
the Chang lab not only as a better scientist, but a more mature person.
I also thank Kevin Wang, who was gracious to teach every technique to a
rotation student with no prior experience in epigenetics or chromatin. I only truly
realized his skill in teaching when I began working with my own rotation students.
I thank my collaborators at University of Michigan, Yong Chen, Bingbing
Wan, and Ming Lei, who provided vital reagents and data fundamental to my work. I
also thank Kun Qu and Jiajing (Jenny) Zhang for their bioinformatics help. I also
thank Karla Kirkegaard, Michel Brahic, and J. Antonio Gomez.
I am also grateful to the rest of my labmates for reagents, data analysis, bench
space, and/or a friendly person to talk to. I would like to thank especially Ryan
Corces-Zimmerman, Ryan Flynn, and Angeline Protacio for their endless support and
enthusiasm for HOTTIP, as well as Ci Chu, Paul Giresi, Tiffany Hung, Lingjie Li,
Nicole Rapicavoli, Robert Spitale, Eduardo Torre, Miao Tsai, Yue Wan, Orly
Wapinski, and Grace Zheng for their various contributions and much needed advice.
This research was supported by the Stanford Medical Scientist Training
Program, who provided both funding and much needed overall support, especially
from Seung Kim, PJ Utz, Lorie Langdon, and Moira Louca.
Finally, I would like to thank my parents, Sung-Chul and Young-Ok Yang, for
encouraging me to do whatever I wanted. My brother, Gene Yang, for being always
proud of his little brother, and somehow knowing about epigenetics as a Sociology/
Urban Studies major. I also thank my in-laws, Pete and Kelly Liberda, who always
enjoy the fact that my research involves foreskin fibroblasts. Most importantly, I am
forever grateful to my wife, Kristine Yang, for her love and caring for me. Without her,
the work in this thesis would have never transpired.
vi
TABLE OF CONTENTS
ABSTRACT
iv
ACKNOWLEDGEMENTS
v
TABLE OF CONTENTS
vi
LIST OF ILLUSTRATIONS
viii
CHAPTER 1: BACKGROUND
1
1.1 Significance of long noncoding RNAs
2
1.2 Introduction to epigenetics
4
1.3 Long noncoding RNAs in epigenetic silencing
5
1.4 Long noncoding RNAs in epigenetic activation
7
1.5 References
8
CHAPTER 2: A LONG NONCODING RNA MAINTAINS ACTIVE
CHROMATIN TO COORDINATE HOMEOTIC GENE EXPRESSION
2.1 Abstract
14
2.2 Results
15
2.3 Discussion
30
2.4 Materials and Methods
31
2.5 Supplementary Figures
39
2.7 References
57
CHAPTER 3: LONG NONCODING RNA AS REGULATORY SWITCH
OF PROTEIN TURNOVER
3.1 Abstract
62
3.2 Introduction
63
3.3 Results
66
3.4 Discussion
89
3.5 Materials and Methods
93
3.6 Supplementary Figures
101
13
61
vii
3.7 References
CHAPTER 4: CONCLUDING REMARKS
106
111
4.1 Conclusion
112
4.2 Future Directions
115
4.3 References
117
viii
List of Illustrations
Figure 2.1. HOTTIP is a lincRNA transcribed in distal anatomic sites.
16
Figure 2.2. HOTTIP is required for coordinate activation of 5′ HOXA genes.
20
Figure 2.3. HOTTIP RNA is required for the active chromatin state of 5′
HOXA cluster.
24
Figure 2.4. HOTTIP RNA programs active chromatin via WDR5.
27
Figure 2S1. Molecular characterization of HOTTIP.
39
Figure 2S2. Single molecule RNA-fluorescence in situ hybridization
(FISH) confirms low copy number of HOTTIP RNA.
41
Figure 2S3. Lack of antisense HOTTIP RNA with HOTTIP knockdown.
42
Figure 2S4. Efficient retroviral infection in developing chick limb buds.
43
Figure 2S5. HOTTIP knockdown decreases expression of 5’ HoxA genes
in vivo.
44
Figure 2S6. HOTTIP depletion causes little change to higher-order
chromosome configuration.
46
Figure 2S7. ChIP-qPCR validation of HOTTIP-dependence of H3K4me3
occupancy at distal HOXA.
48
Figure 2S8. Zoom-in view of Figure 2.3 highlighting HOTTIP dependence
of MLL1 and WDR5 localization to HOXA.
49
Figure 2S9. HOTTIP binds stably expressed FLAG-WDR5 in HeLa cells.
50
Figure 2S10. HOTTIP overexpression does not affect distal HOXA
expression.
51
Figure 2S11. Ectopic HOTTIP expression does not activate 5’HOXA genes
nor rescue the effects of depleting endogenous nascent HOTTIP.
53
Figure 2S12. HOTTIPExons1-2 binds to WDR5 and acts in a dominant
negative manner to inhibit 5’ HOXA gene expression.
55
Figure 2S13. Model of HOTTIP action. HOTTIP is positioned near the
active 5’ HOXA genes via chromosomal looping.
56
Figure 3.1. HOTTIP RNA overexpression stabilizes exogenous
FLAG-WDR5.
68
ix
Figure 3.2. HOTTIP RNA thermodynamically stabilizes WDR5 and
does not affect ubiquitination.
72
Figure 3.3. Full length HOTTIP RNA stabilizes WDR5 through specific
interactions.
75
Figure 3.4. HOTTIP binds and stabilizes WDR5 through the same binding
site as RbBP5.
78
Figure 3.5. HOTTIP binding mutations prevent HOTTIP mediated
stabilization of WDR5, as well as show decreased ability to
activate target genes in 293T cells.
81
Figure 3.6. WDR5 F266A mutation decreases protein stability and causes
defects in mouse embryonic stem cell self renewal.
86
Figure 3.7. WDR5 F266A causes loss of self renewal genes and increased
expression of differentiation genes.
88
Figure 3S1.
101
Figure 3S2. Two regions of HOTTIP bind WDR5.
102
Figure 3S3. HOTTIP bases 1953-2453 (Frag D.3) demonstrate
equivalent binding affinity to bases 1953-3760 (Frag D).
104
Figure 3S4.
105
1
CHAPTER 1
Background
2
1.1 Significance of long noncoding RNAs
With the conclusion of the human genome project, researchers were surprised
to find that much of the genome did not appear to encode for proteins (Lander et al.,
2001; Venter et al., 2001). Because proteins were deemed to be the “building blocks of
life,” researchers at the time dubbed the extensive, nonttranslated regions as the “dark
matter” of the genome, or more simply as “junk DNA.”
Since then, many groups have been probing this untranslated “dark matter”
using traditional Sanger sequencing, or the more high throughput microarray and next
generation sequencing technologies (Birney et al., 2007; Cabili et al., 2011; Carninci
et al., 2005; Cawley et al., 2004; Core et al., 2008; Okazaki et al., 2002; Preker et al.,
2008). While much of the “dark matter” are repetitive elements, regulatory regions,
and small RNAs (eg. miRNAs), a surprising proportion is transcribed into long RNA
molecules with little or no coding potential. Ranging from several hundred to
kilobases in size, these long RNAs lacking coding potential have been called
large/long intergenic noncoding RNAs (lincRNAs) or large/long noncoding RNAs
(lncRNAs), somewhat interchangably, although the latter term is more inclusive.
Many of these lncRNAs are spliced and polyadenylated, and further demonstrate
evolutionary stability (Lebenthal and Unger, 2010; Ponjavic et al., 2007). By
correlating histone modifications signifying transcription with coding potential,
researchers have estimated that lncRNAs may number in the thousands (Guttman et al.,
2009; Khalil et al., 2009). In fact, recent research has suggested that most of the
genome may actually be transcribed, albeit some regions at very low levels with
unknown functionality (Mercer et al., 2012).
Over the last several years, many of these lncRNAs have been characterized,
revealing important roles in all aspects of cellular function, from gene transcription to
posttranslational control. For example, to regulate transcription, a noncoding RNA
transcribed from the minor DHFR promoter represses DHFR expression by forming a
stable triplex at the major promoter, inhibiting transcription factor binding (Martianov
et al., 2007). Similarly, the Alu RNA, transcribed from short interspersed elements
(SINEs) throughout the genome in heat shock, binds RNA polymerase II to prevent
3
genome-wide transcription (Mariner et al., 2008). Post transcriptionally, a noncoding
RNA has been described that occludes the 5’ splice site of the Zeb2 mRNA, reducing
splicing to cause retention of an internal ribosomal entry site (Beltran et al., 2008).
Likewise regulating mRNA levels, ½-sbsRNAs pair with Alu repeats in 3’UTRs, to
cause STAU1-mediated RNA decay (Gong and Maquat, 2011). At the protein level,
the noncoding RNA NRON regulates NFAT subcellular localization (Willingham et
al., 2005). Other noncoding RNAs have been found to regulate localization of splicing
factors and active/silent chromatin domains (Tripathi et al., 2010; Yang et al., 2011).
LncRNAs have also been found to be associated with disease processes. For
example, in heart disease, expression of the antisense noncoding RNA in the INK4
locus (ANRIL) has been found to be associated with increased atherosclerosis (Holdt
et al., 2010), and expression of the noncoding RNA MIAT correlates with increased
risk of myocardial infarction (Ishii et al., 2006). In another example, the stress-induced
BACE1-AS noncoding RNA is overexpressed in Alzheimer’s disease, where it forms
an RNA duplex with and stabilizes BACE1, resulting in a feed-forward progression of
the amyloid cascade (Faghihi et al., 2008). In cancer, ceRNAs (competitive
endogenous RNAs) bind to and reduce miRNAs, thus regulating tumor suppressor
genes (eg. PTEN) and proto-oncogenes (eg. KRAS) (Poliseno et al., 2010). Multiple
other long noncoding RNAs have been reported in malignancies of the colon (Pibouin
et al., 2002), prostate (Bussemakers et al., 1999; Fu et al., 2006), breast (Guffanti et al.,
2009), and liver (Lin et al., 2007), as well as in leukemia (Calin et al., 2007; Yu et al.,
2008), melanoma (Pasmant et al., 2007), and in metastasis-prone lung and breast
cancers (Gupta et al., 2010; Ji et al., 2003).
4
1.2 Introduction to epigenetics
Epigenetics refers to inheritable changes occurring through mechanisms other
than DNA sequence alterations. While potentially an expansive definition, much of
today’s epigenetics focuses on two changes in the DNA state: 1) methylation at CpG
sites on the DNA molecule itself and 2) chromatin remodeling by covalent histone
modifications.
Both DNA methylation and histone modifications are faithfully transmitted
from mother to daughter cells. Thus, the state of gene expression can be transmitted as
well. For example, CpG methylation, H3K9 methylation, and H3K27 methylation
indicate silenced genes. In contrast, histone acetylation, H3K4 methylation, and lack
of DNA methylation mark active genes. As the field progresses, more diverse histone
modifications are being described and characterized, such as propionylation,
sumoylation, crotonylation, citrullination, butyrylation, proline isomerization, ADP
ribosylation, formylation, phosphorylation, ubiquylation, and hydroxylation (Tan et al.,
2011). Together, these modifications generate a “histone code,” which denotes the
DNA state of particular loci (Campos and Reinberg, 2009). Current research is
ongoing to determine the functionality of each of these marks, and their relative
contributions to particular gene states.
5
1.3 Long noncoding RNAs in epigenetic silencing
LncRNAs have been found to play major roles in silencing genes by altering
DNA methylation and histone modifications. For example, transcription of antisense
RNA to HBA2 causes CpG methylation and gene silencing, resulting in thalassemia
(Tufarelli et al., 2003). Similarly, the pRNA (promoter-associated RNA) form DNARNA triplexes that recruit the DNA methyltransferase DNMT3b for similar
methylation-mediated silencing (Schmitz et al., 2010). As an example in regulating
histone modifications, the Air noncoding RNA recruits the histone methyltransferase
G9a to the Slc22a2 promoter to cause H3K9 methylation and transcriptional
repression in the placenta (Nagano et al., 2008). Silencing can also occur by inhibiting
the addition of histone activating marks. DNA damage-induced noncoding RNAs
activate the TLS protein to bind and inhibit the CBP/p300 histone acetyltransferase
activity on the Cyclin D1 locus, thus reducing gene expression (Wang et al., 2008).
In the best understood example of lncRNA-mediated epigenetic control, other
lncRNAs silence genes by recruiting the PRC2 methyltransferase complex to cause
H3K27 trimethylation. In females, the long noncoding RNA Xist binds the PRC2
complex and recruits it to silence one of the two copies of the X chromosome in
dosage compensation (Morey and Avner, 2011). Likewise, the Kcnq1ot1 RNA recruits
the PRC2 complex as well as the H3K9 methyltransferase G9a to silence the Kcnq1
domain in the placenta (Pandey et al., 2008), and a similar interaction has been found
to regulate p15/INK4B (Kotake et al., 2011). In addition, HOTAIR, a lncRNA
expressed only from distal and posterior anatomic sites, acts as a molecular scaffold,
simultaneously binding the H3K27 methyltransferase complex PRC2 and H3K4
demethyltransferase complex LSD1/CoREST/REST (Tsai et al., 2010). It then brings
these proteins to silence target genes in trans by two mechanisms: 1) adding the
H3K27 trimethyl silencing mark and 2) demethylating the H3K4 dimethyl activating
mark. In cancer, HOTAIR can silence metastasis suppressor genes – overexpression of
HOTAIR in epithelial cancer cells causes increased invasiveness and metastasis
(Gupta et al., 2010). Additional PRC2-binding lncRNAs have also been characterized,
such as COLDAIR (Heo and Sung, 2011), as well as other candidates genome-wide
6
(Guil et al., 2012; Khalil et al., 2009; Zhao et al., 2010), suggesting a common
mechanism of lncRNA-mediated gene silencing.
7
1.4 Long noncoding RNAs in epigenetic activation
While lncRNAs involved in gene silencing have been well characterized, their
role in gene activation is less understood. As the best studied example, the lncRNAs
roX1 and roX2 recruit the MSL complex to the single X chromosome in male
Drosophila dosage compensation, causing H4K16 acetylation and gene activation
(Deng and Meller, 2006). Also in Drosophila, noncoding RNA transcription from
trithorax response elements correlates with activation of entire regulatory domains in
segmental development, possibly by noncoding RNA-mediated recruitment of the
H3K4 methyltransferase Ash1 (Petruk et al., 2006; Sanchez-Elsner et al., 2006).
Research has also shown a role for activating lncRNAs in mammals. Recently,
lncRNAs from various cell types have been found to display enhancer-like functions
(Orom et al., 2010). siRNA-mediated knockdown of these lncRNAs causes
downregulation of nearby genes, and these enhancer-like lncRNAs can upregulate
nearby luciferase reporter genes in cis in reporter assays. As an individual example,
the lncRNA HOTAIRM1 has been found to be required to maintain expression of
nearby myelopoiesis-related HoxA genes (Zhang et al., 2009). Also in the Hox loci,
the Mistral RNA (Mira) has been described to bind and recruit MLL to activate
transcription of HoxA6-HoxA7 (Bertani et al., 2011). Finally, multiple lines of
evidence have shown correlation between lncRNA transcription and nearby coding
gene activation, with noncoding RNAs arising from enhancer elements and conserved
regions (Kim et al., 2010; Ponjavic et al., 2009; Rada-Iglesias et al., 2011).
Because of their extensive involvement in all aspects of cellular function, a
better understanding of lncRNAs would provide valuable in explaining disease and
developmental processes. The body of work in this dissertation focuses on the initial
discovery and ongoing characterization of the novel lncRNA HOTTIP, and the role of
lncRNAs in WDR5-mediated epigenetic activation.
8
1.5 References
Beltran, M., Puig, I., Pena, C., Garcia, J.M., Alvarez, A.B., Pena, R., Bonilla, F., and
de Herreros, A.G. (2008). A natural antisense transcript regulates Zeb2/Sip1
gene expression during Snail1-induced epithelial-mesenchymal transition.
Genes Dev 22, 756-769.
Bertani, S., Sauer, S., Bolotin, E., and Sauer, F. (2011). The noncoding RNA Mistral
activates Hoxa6 and Hoxa7 expression and stem cell differentiation by
recruiting MLL1 to chromatin. Mol Cell 43, 1040-1046.
Birney, E., Stamatoyannopoulos, J.A., Dutta, A., Guigo, R., Gingeras, T.R., Margulies,
E.H., Weng, Z., Snyder, M., Dermitzakis, E.T., Thurman, R.E., et al. (2007).
Identification and analysis of functional elements in 1% of the human genome
by the ENCODE pilot project. Nature 447, 799-816.
Bussemakers, M.J., van Bokhoven, A., Verhaegh, G.W., Smit, F.P., Karthaus, H.F.,
Schalken, J.A., Debruyne, F.M., Ru, N., and Isaacs, W.B. (1999). DD3: a new
prostate-specific gene, highly overexpressed in prostate cancer. Cancer Res 59,
5975-5979.
Cabili, M.N., Trapnell, C., Goff, L., Koziol, M., Tazon-Vega, B., Regev, A., and Rinn,
J.L. (2011). Integrative annotation of human large intergenic noncoding RNAs
reveals global properties and specific subclasses. Genes Dev 25, 1915-1927.
Calin, G.A., Liu, C.G., Ferracin, M., Hyslop, T., Spizzo, R., Sevignani, C., Fabbri, M.,
Cimmino, A., Lee, E.J., Wojcik, S.E., et al. (2007). Ultraconserved regions
encoding ncRNAs are altered in human leukemias and carcinomas. Cancer
Cell 12, 215-229.
Campos, E.I., and Reinberg, D. (2009). Histones: annotating chromatin. Annu Rev
Genet 43, 559-599.
Carninci, P., Kasukawa, T., Katayama, S., Gough, J., Frith, M.C., Maeda, N., Oyama,
R., Ravasi, T., Lenhard, B., Wells, C., et al. (2005). The transcriptional
landscape of the mammalian genome. Science 309, 1559-1563.
Cawley, S., Bekiranov, S., Ng, H.H., Kapranov, P., Sekinger, E.A., Kampa, D.,
Piccolboni, A., Sementchenko, V., Cheng, J., Williams, A.J., et al. (2004).
Unbiased mapping of transcription factor binding sites along human
chromosomes 21 and 22 points to widespread regulation of noncoding RNAs.
Cell 116, 499-509.
Core, L.J., Waterfall, J.J., and Lis, J.T. (2008). Nascent RNA sequencing reveals
widespread pausing and divergent initiation at human promoters. Science 322,
1845-1848.
Deng, X., and Meller, V.H. (2006). Non-coding RNA in fly dosage compensation.
Trends Biochem Sci 31, 526-532.
Faghihi, M.A., Modarresi, F., Khalil, A.M., Wood, D.E., Sahagan, B.G., Morgan, T.E.,
Finch, C.E., St Laurent, G., 3rd, Kenny, P.J., and Wahlestedt, C. (2008).
Expression of a noncoding RNA is elevated in Alzheimer's disease and drives
rapid feed-forward regulation of beta-secretase. Nat Med 14, 723-730.
Fu, X., Ravindranath, L., Tran, N., Petrovics, G., and Srivastava, S. (2006). Regulation
of apoptosis by a prostate-specific and prostate cancer-associated noncoding
gene, PCGEM1. DNA Cell Biol 25, 135-141.
9
Gong, C., and Maquat, L.E. (2011). lncRNAs transactivate STAU1-mediated mRNA
decay by duplexing with 3' UTRs via Alu elements. Nature 470, 284-288.
Guffanti, A., Iacono, M., Pelucchi, P., Kim, N., Solda, G., Croft, L.J., Taft, R.J., Rizzi,
E., Askarian-Amiri, M., Bonnal, R.J., et al. (2009). A transcriptional sketch of
a primary human breast cancer by 454 deep sequencing. BMC Genomics 10,
163.
Guil, S., Soler, M., Portela, A., Carrere, J., Fonalleras, E., Gomez, A., Villanueva, A.,
and Esteller, M. (2012). Intronic RNAs mediate EZH2 regulation of epigenetic
targets. Nat Struct Mol Biol.
Gupta, R.A., Shah, N., Wang, K.C., Kim, J., Horlings, H.M., Wong, D.J., Tsai, M.C.,
Hung, T., Argani, P., Rinn, J.L., et al. (2010). Long non-coding RNA
HOTAIR reprograms chromatin state to promote cancer metastasis. Nature 464,
1071-1076.
Guttman, M., Amit, I., Garber, M., French, C., Lin, M.F., Feldser, D., Huarte, M., Zuk,
O., Carey, B.W., Cassady, J.P., et al. (2009). Chromatin signature reveals over
a thousand highly conserved large non-coding RNAs in mammals. Nature 458,
223-227.
Heo, J.B., and Sung, S. (2011). Vernalization-mediated epigenetic silencing by a long
intronic noncoding RNA. Science 331, 76-79.
Holdt, L.M., Beutner, F., Scholz, M., Gielen, S., Gabel, G., Bergert, H., Schuler, G.,
Thiery, J., and Teupser, D. (2010). ANRIL expression is associated with
atherosclerosis risk at chromosome 9p21. Arterioscler Thromb Vasc Biol 30,
620-627.
Ishii, N., Ozaki, K., Sato, H., Mizuno, H., Saito, S., Takahashi, A., Miyamoto, Y.,
Ikegawa, S., Kamatani, N., Hori, M., et al. (2006). Identification of a novel
non-coding RNA, MIAT, that confers risk of myocardial infarction. J Hum
Genet 51, 1087-1099.
Ji, P., Diederichs, S., Wang, W., Boing, S., Metzger, R., Schneider, P.M., Tidow, N.,
Brandt, B., Buerger, H., Bulk, E., et al. (2003). MALAT-1, a novel noncoding
RNA, and thymosin beta4 predict metastasis and survival in early-stage nonsmall cell lung cancer. Oncogene 22, 8031-8041.
Khalil, A.M., Guttman, M., Huarte, M., Garber, M., Raj, A., Rivea Morales, D.,
Thomas, K., Presser, A., Bernstein, B.E., van Oudenaarden, A., et al. (2009).
Many human large intergenic noncoding RNAs associate with chromatinmodifying complexes and affect gene expression. Proc Natl Acad Sci U S A
106, 11667-11672.
Kim, T.K., Hemberg, M., Gray, J.M., Costa, A.M., Bear, D.M., Wu, J., Harmin, D.A.,
Laptewicz, M., Barbara-Haley, K., Kuersten, S., et al. (2010). Widespread
transcription at neuronal activity-regulated enhancers. Nature 465, 182-187.
Kotake, Y., Nakagawa, T., Kitagawa, K., Suzuki, S., Liu, N., Kitagawa, M., and
Xiong, Y. (2011). Long non-coding RNA ANRIL is required for the PRC2
recruitment to and silencing of p15(INK4B) tumor suppressor gene. Oncogene
30, 1956-1962.
Lander, E.S., Linton, L.M., Birren, B., Nusbaum, C., Zody, M.C., Baldwin, J., Devon,
K., Dewar, K., Doyle, M., FitzHugh, W., et al. (2001). Initial sequencing and
10
analysis of the human genome. Nature 409, 860-921.
Lebenthal, I., and Unger, R. (2010). Computational evidence for functionality of
noncoding mouse transcripts. Genomics 96, 10-16.
Lin, R., Maeda, S., Liu, C., Karin, M., and Edgington, T.S. (2007). A large noncoding
RNA is a marker for murine hepatocellular carcinomas and a spectrum of
human carcinomas. Oncogene 26, 851-858.
Mariner, P.D., Walters, R.D., Espinoza, C.A., Drullinger, L.F., Wagner, S.D., Kugel,
J.F., and Goodrich, J.A. (2008). Human Alu RNA is a modular transacting
repressor of mRNA transcription during heat shock. Mol Cell 29, 499-509.
Martianov, I., Ramadass, A., Serra Barros, A., Chow, N., and Akoulitchev, A. (2007).
Repression of the human dihydrofolate reductase gene by a non-coding
interfering transcript. Nature 445, 666-670.
Mercer, T.R., Gerhardt, D.J., Dinger, M.E., Crawford, J., Trapnell, C., Jeddeloh, J.A.,
Mattick, J.S., and Rinn, J.L. (2012). Targeted RNA sequencing reveals the
deep complexity of the human transcriptome. Nat Biotechnol 30, 99-104.
Morey, C., and Avner, P. (2011). The demoiselle of X-inactivation: 50 years old and
as trendy and mesmerising as ever. PLoS Genet 7, e1002212.
Nagano, T., Mitchell, J.A., Sanz, L.A., Pauler, F.M., Ferguson-Smith, A.C., Feil, R.,
and Fraser, P. (2008). The Air noncoding RNA epigenetically silences
transcription by targeting G9a to chromatin. Science 322, 1717-1720.
Okazaki, Y., Furuno, M., Kasukawa, T., Adachi, J., Bono, H., Kondo, S., Nikaido, I.,
Osato, N., Saito, R., Suzuki, H., et al. (2002). Analysis of the mouse
transcriptome based on functional annotation of 60,770 full-length cDNAs.
Nature 420, 563-573.
Orom, U.A., Derrien, T., Beringer, M., Gumireddy, K., Gardini, A., Bussotti, G., Lai,
F., Zytnicki, M., Notredame, C., Huang, Q., et al. (2010). Long noncoding
RNAs with enhancer-like function in human cells. Cell 143, 46-58.
Pandey, R.R., Mondal, T., Mohammad, F., Enroth, S., Redrup, L., Komorowski, J.,
Nagano, T., Mancini-Dinardo, D., and Kanduri, C. (2008). Kcnq1ot1 antisense
noncoding RNA mediates lineage-specific transcriptional silencing through
chromatin-level regulation. Mol Cell 32, 232-246.
Pasmant, E., Laurendeau, I., Heron, D., Vidaud, M., Vidaud, D., and Bieche, I. (2007).
Characterization of a germ-line deletion, including the entire INK4/ARF locus,
in a melanoma-neural system tumor family: identification of ANRIL, an
antisense noncoding RNA whose expression coclusters with ARF. Cancer Res
67, 3963-3969.
Petruk, S., Sedkov, Y., Riley, K.M., Hodgson, J., Schweisguth, F., Hirose, S., Jaynes,
J.B., Brock, H.W., and Mazo, A. (2006). Transcription of bxd noncoding
RNAs promoted by trithorax represses Ubx in cis by transcriptional
interference. Cell 127, 1209-1221.
Pibouin, L., Villaudy, J., Ferbus, D., Muleris, M., Prosperi, M.T., Remvikos, Y., and
Goubin, G. (2002). Cloning of the mRNA of overexpression in colon
carcinoma-1: a sequence overexpressed in a subset of colon carcinomas.
Cancer Genet Cytogenet 133, 55-60.
Poliseno, L., Salmena, L., Zhang, J., Carver, B., Haveman, W.J., and Pandolfi, P.P.
11
(2010). A coding-independent function of gene and pseudogene mRNAs
regulates tumour biology. Nature 465, 1033-1038.
Ponjavic, J., Oliver, P.L., Lunter, G., and Ponting, C.P. (2009). Genomic and
transcriptional co-localization of protein-coding and long non-coding RNA
pairs in the developing brain. PLoS Genet 5, e1000617.
Ponjavic, J., Ponting, C.P., and Lunter, G. (2007). Functionality or transcriptional
noise? Evidence for selection within long noncoding RNAs. Genome Res 17,
556-565.
Preker, P., Nielsen, J., Kammler, S., Lykke-Andersen, S., Christensen, M.S.,
Mapendano, C.K., Schierup, M.H., and Jensen, T.H. (2008). RNA exosome
depletion reveals transcription upstream of active human promoters. Science
322, 1851-1854.
Rada-Iglesias, A., Bajpai, R., Swigut, T., Brugmann, S.A., Flynn, R.A., and Wysocka,
J. (2011). A unique chromatin signature uncovers early developmental
enhancers in humans. Nature 470, 279-283.
Sanchez-Elsner, T., Gou, D., Kremmer, E., and Sauer, F. (2006). Noncoding RNAs of
trithorax response elements recruit Drosophila Ash1 to Ultrabithorax. Science
311, 1118-1123.
Schmitz, K.M., Mayer, C., Postepska, A., and Grummt, I. (2010). Interaction of
noncoding RNA with the rDNA promoter mediates recruitment of DNMT3b
and silencing of rRNA genes. Genes Dev 24, 2264-2269.
Tan, M., Luo, H., Lee, S., Jin, F., Yang, J.S., Montellier, E., Buchou, T., Cheng, Z.,
Rousseaux, S., Rajagopal, N., et al. (2011). Identification of 67 histone marks
and histone lysine crotonylation as a new type of histone modification. Cell
146, 1016-1028.
Tripathi, V., Ellis, J.D., Shen, Z., Song, D.Y., Pan, Q., Watt, A.T., Freier, S.M.,
Bennett, C.F., Sharma, A., Bubulya, P.A., et al. (2010). The nuclear-retained
noncoding RNA MALAT1 regulates alternative splicing by modulating SR
splicing factor phosphorylation. Mol Cell 39, 925-938.
Tsai, M.C., Manor, O., Wan, Y., Mosammaparast, N., Wang, J.K., Lan, F., Shi, Y.,
Segal, E., and Chang, H.Y. (2010). Long noncoding RNA as modular scaffold
of histone modification complexes. Science 329, 689-693.
Tufarelli, C., Stanley, J.A., Garrick, D., Sharpe, J.A., Ayyub, H., Wood, W.G., and
Higgs, D.R. (2003). Transcription of antisense RNA leading to gene silencing
and methylation as a novel cause of human genetic disease. Nat Genet 34, 157165.
Venter, J.C., Adams, M.D., Myers, E.W., Li, P.W., Mural, R.J., Sutton, G.G., Smith,
H.O., Yandell, M., Evans, C.A., Holt, R.A., et al. (2001). The sequence of the
human genome. Science 291, 1304-1351.
Wang, X., Arai, S., Song, X., Reichart, D., Du, K., Pascual, G., Tempst, P., Rosenfeld,
M.G., Glass, C.K., and Kurokawa, R. (2008). Induced ncRNAs allosterically
modify RNA-binding proteins in cis to inhibit transcription. Nature 454, 126130.
Willingham, A.T., Orth, A.P., Batalov, S., Peters, E.C., Wen, B.G., Aza-Blanc, P.,
Hogenesch, J.B., and Schultz, P.G. (2005). A strategy for probing the function
12
of noncoding RNAs finds a repressor of NFAT. Science 309, 1570-1573.
Yang, L., Lin, C., Liu, W., Zhang, J., Ohgi, K.A., Grinstein, J.D., Dorrestein, P.C., and
Rosenfeld, M.G. (2011). ncRNA- and Pc2 methylation-dependent gene
relocation between nuclear structures mediates gene activation programs. Cell
147, 773-788.
Yu, W., Gius, D., Onyango, P., Muldoon-Jacobs, K., Karp, J., Feinberg, A.P., and Cui,
H. (2008). Epigenetic silencing of tumour suppressor gene p15 by its antisense
RNA. Nature 451, 202-206.
Zhang, X., Lian, Z., Padden, C., Gerstein, M.B., Rozowsky, J., Snyder, M., Gingeras,
T.R., Kapranov, P., Weissman, S.M., and Newburger, P.E. (2009). A
myelopoiesis-associated regulatory intergenic noncoding RNA transcript
within the human HOXA cluster. Blood 113, 2526-2534.
Zhao, J., Ohsumi, T.K., Kung, J.T., Ogawa, Y., Grau, D.J., Sarma, K., Song, J.J.,
Kingston, R.E., Borowsky, M., and Lee, J.T. (2010). Genome-wide
identification of polycomb-associated RNAs by RIP-seq. Mol Cell 40, 939-953.
13
CHAPTER 2
A long noncoding RNA maintains active chromatin to coordinate
homeotic gene expression
This chapter was published in Wang, K.C., Yang, Y.W.*, Liu, B.*, Sanyal, A.,
Corces-Zimmerman, R., Chen, Y., Lajoie, B.R., Protacio, A., Flynn, R.A., Gupta,
R.A., Wysocka J., Lei M., Dekker J., Helms J.A., and Chang H.Y. (2011). A long
noncoding RNA maintains active chromatin to coordinate homeotic gene expression.
Nature 472, 120-124.
*These authors contributed equally to this work.
Y.W. Yang’s primary contributions to this publication were portions of the control
regions (Figure 2.2B), chick HOTTIP shRNA screen (later used in Figure 2.2C-F,
2S4-2S5), GST-tagged protein pull down assays (Figure 2.4A-B, 2S12A), antibody
screen (later used for Figure 2.4C), WDR5 siRNA experiments (Figure 2.4D), strand
specific qPCR (Figure 2S1B, 2S3A), titrated qRT-PCR (Figure 2S2C), tagged
WDR5 RNA-immunoprecipitation assays (Figure 2S9A-B), and generation of
HOTTIP and HOTTIP fragment overexpression cell lines (Figure 2S10, 2S11,
2S12B). Y.W. Yang also had a minor role in preparing and writing the manuscript.
14
2.1 Abstract
The genome is extensively transcribed into long intergenic noncoding RNAs
(lincRNAs), many of which are implicated in gene silencing (Mercer et al., 2009;
Ponting et al., 2009). Potential roles of lincRNAs in gene activation are much less
understood (Dinger et al., 2008; Petruk et al., 2006; Sanchez-Elsner et al., 2006).
Development and homeostasis require coordinate regulation of neighbouring genes
through a process termed locus control (Dean, 2006). Some locus control elements and
enhancers transcribe lincRNAs (Ashe et al., 1997; De Santa et al., 2010; Kim et al.,
2010; Orom et al., 2010), hinting at possible roles in long-range control. In vertebrates,
39 Hox genes, encoding homeodomain transcription factors critical for positional
identity, are clustered in four chromosomal loci; the Hox genes are expressed in nested
anterior-posterior and proximal-distal patterns colinear with their genomic position
from 3′ to 5′of the cluster (Chang, 2009). Here we identify HOTTIP, a lincRNA
transcribed from the 5′ tip of the HOXA locus that coordinates the activation of several
5′ HOXA genes in vivo. Chromosomal looping brings HOTTIP into close proximity to
its target genes. HOTTIP RNA binds the adaptor protein WDR5 directly and targets
WDR5/MLL complexes across HOXA, driving histone H3 lysine 4 trimethylation and
gene transcription. Induced proximity is necessary and sufficient for HOTTIP RNA
activation of its target genes. Thus, by serving as key intermediates that transmit
information from higher order chromosomal looping into chromatin modifications,
lincRNAs may organize chromatin domains to coordinate long-range gene activation.
15
2.2 Results
We examined chromosome structure and histone modifications in human
primary fibroblasts derived from several anatomic sites (Rinn et al., 2007), and found
distinctive differences in the HOXA locus. High throughput chromosome
conformation capture (5C) (Dostie et al., 2006) across HOXA revealed that its higher
order structure is dependent on positional identity. In anatomically distal cells (for
example, foreskin and foot fibroblasts), we detected abundant chromatin interactions
within the transcriptionally active 5′ HOXA locus (with reference to the directions of
transcription of constituent Hox genes), pointing to a compact and looped
conformation. In contrast, no long-range chromatin interactions are detected within the
transcriptionally silent 3′ HOXA which seems largely linear (Figure 2.1A). Strikingly,
anatomically proximal cells (for example, lung fibroblasts) have the diametrically
opposite pattern. The ON and OFF states of Hox and other key developmental genes
are maintained by the MLL/Trithorax (Trx) and polycomb group (PcG) proteins,
which mediate trimethylation of histone H3 lysine 4 (H3K4me3) to activate genes or
lysine 27 (H3K27me3) to repress genes (Schuettengruber et al., 2007). The portions of
HOXA in tight physical interaction are marked by broad domains of H3K4me3,
whereas H3K27me3 marks the physically extended and transcriptional silent regions
(Figure 2.1A).
16
Figure 2.1. HOTTIP is a lincRNA transcribed in distal anatomic sites. a, Chromatin
state map of distal versus proximal cells. Top panels, chromosome conformation
capture-carbon copy (5C) analysis of distal (foreskin) and proximal (lung) human
fibroblasts. Heat map representations (generated by my5C, ref. (Lajoie et al., 2009)) of
5C data (bin size 30 kb, step size 3 kb) for HOXA in foreskin and lung fibroblasts. Red
intensity of each pixel indicates relative interaction between the two points on the
17
genomic coordinates. The diagonal represents frequent cis interactions between
regions located in close proximity along the linear genome. 5C signals that are away
from the diagonal represent long-range looping interactions. Bottom panels, chromatin
occupancy across HOXA. x-axis is genomic coordinate; y-axis depicts occupancy of
the indicated histone marks or protein (ChIP/input). Box and arrows highlight
chromatin states of HOTTIP gene. b, HOTTIP RNA expression in primary human
fibroblasts from 11 anatomic sites. Means ± s.d. are shown (n = 2). c, In situ
hybridization of HOTTIP RNA in E13.5 mouse embryo.
18
On the very 5′ and 3′ edges of the two respective interaction clusters are two
lincRNA loci that exhibit distinct chromatin modifications. The 3′element has been
previously identified as the myelopoiesis-associated lincRNA HOTAIRM1 (Zhang et
al., 2009). The 5′ element, for which we suggest the name HOTTIP for ‘HOXA
transcript at the distal tip’, exhibits bivalent H3K4me3 and H3K27me3, a histone
modification pattern associated with poised regulatory sequences (Bernstein et al.,
2006). Comparison with RNA polymerase II occupancy and RNA expression showed
that the bivalent H3K4me3 and H3K27me3 modifications on HOTTIP gene do not
require HOTTIP transcription, but transcription of HOTTIP is associated with
increased H3K4me3 and decreased H3K27me3 (Figure 2.1A, left). Chromatin
immunoprecipitation (ChIP) analysis confirmed that the HOTTIP gene is occupied by
both polycomb repressive complex 2 (PRC2) and MLL complex, consistent with the
bivalent histone marks (Figure 2S1A).
HOTTIP transcription yields a 3,764-nucleotide, spliced and polyadenylated
lincRNA that initiates ~330 bases upstream of HOXA13. Only the strand antisense to
HOXA genes is transcribed (Figure 2S1B). Genes near the 5′ end of each HOX cluster
tend to be expressed in more posterior and/or distal anatomical locations. Consistent
with its genomic location 5′ to HOXA13, HOTTIP is expressed in distal and/or
posterior anatomic sites (Figure 2.1B). In situ hybridization of developing mouse and
chick embryos confirmed that HOTTIP is expressed in posterior and distal sites in vivo,
indicating a conserved expression pattern from development to adulthood (Figure
2.1C and Figure 2S1C). Even in distal cells where HOTTIP is expressed, its RNA
level is very low and estimated to be ~0.3 copies per cell (Figure 2S2).
We employed small interfering RNAs (siRNAs) to knock down HOTTIP RNA
in fibroblasts from a distal anatomic site (foreskin), and examined expression of 5′
HOXA genes by quantitative reverse transcription PCR. Notably, HOTTIP RNA
knockdown abrogated expression of distal HOXA genes across 40 kilobases with a
trend dependent on the distance to HOTTIP. The strongest blockade was observed for
HOXA13 and HOXA11, with progressively less severe effects on HOXA10, HOXA9
and HOXA7 (Figure 2.2A). The effect on gene transcription appeared to be
19
unidirectional, as there were no appreciable changes in the levels of EVX1, located
~40 kilobases 5′ of the HOXA cluster (data not shown). HOTTIP knockdown did not
affect expression of the highly homologous HOXD genes, other control genes, nor
induce antisense transcription at its own locus (Figure 2.2B, Figure 2S3A). Several
independent siRNAs targeting HOTTIP yielded similar results (Figure 2S3B). These
results indicate that HOTTIP RNA is necessary to coordinate activation of 5′ HOXA
genes.
20
Figure 2.2. HOTTIP is required for coordinate activation of 5′ HOXA genes. a, b,
Knockdown of HOTTIP RNA abrogates expression of 5′ HOXA genes in foreskin
fibroblasts (a), but not HOXD or BID genes (b). Means + s.d. are shown (n = 3). GFP,
green fluorescent protein. c, Schematic of chick RNAi experiment. d, HOTTIP RNA
is required for 5′ HoxA gene expression in vivo. RT–PCR of the indicated genes from
control or HOTTIP-depleted distal limb bud is shown; quantification and
normalization by Actin signal is shown below each band. GAG signal confirms
successful retroviral transduction in all cases. e, In situ hybridization of 5′ HoxA genes
21
in chick limb buds. Arrowheads highlight distal domains of high HoxA gene
expression that are affected by HOTTIP knockdown. f, Shortening of distal bony
elements in HOTTIP-depleted forelimbs. Alcian blue staining highlights the skeletal
elements. Red and purple lines highlight radius and 3rd digit lengths, respectively.
22
We next addressed the function of HOTTIP RNA in vivo in the developing
chick limb bud (Figure 2.2C). Whereas prior genetic studies of noncoding RNAs
(ncRNAs) involved deletion or insertion into the gene locus (Kmita et al., 2002), we
wished to distinguish the functions of HOTTIP RNA from its corresponding DNA
element. HOTTIP gene can nucleate H3K4 and H3K27 methylation independent of
transcription (Figure 2.1A), and the precise genomic distance between upstream
enhancer elements and Hox genes is critical for their proper colinear activation (Kmita
et al., 2002). Therefore, we used RNA interference (RNAi) in chick embryos, where
replication-competent retroviruses can deliver short-hairpin RNAs (shRNAs) with
high penetrance and precise spatiotemporal control (Harpavat and Cepko, 2006)
(Figure 2S4). In the limb bud, 5′ HoxA genes are transcribed in a nested pattern along
the proximal-distal axis (Nelson et al., 1996). In this tissue, HoxA function is highly
redundant with that of the HoxD locus, which allowed us to assess altered HoxA
expression patterns without major changes in anatomic landmarks (Kmita et al., 2005).
We injected retroviruses carrying shRNAs against chick HOTTIP into upper limb buds
of stage 13 chicks; RT–PCR and in situ hybridization were performed on both control
and knockdown samples after 2–4 days. Knockdown of HOTTIP RNA by two
independent shRNAs in limb buds decreased expression of HoxA13, HoxA11 and
HoxA10—again with a graded impact depending on genomic proximity to HOTTIP
gene. Vector control or an shRNA that fails to deplete HOTTIP RNA had little effect
on Hox gene expression (Figure 2.2D). In situ hybridization on whole embryos
(Figure 2.2E) and sections (Figure 2S5) revealed that HOTTIP RNA most strongly
affects HoxA gene expression at the distal edge of the developing limb bud, where the
5′ HoxA genes are most strongly expressed. By stage 36, limbs depleted of HOTTIP
RNA showed notable shortening and bending of distal bony elements, including the
radius, ulna and third digit (~20% length reduction for each compared to contralateral
and stage-matched limbs treated with control virus, P < 0.05, Student’s t-test, Figure
2.2F). This phenotype resembled some of the defects in mice lacking HoxA11 and
HoxA13 (Davis et al., 1995; Fromental-Ramain et al., 1996; Small and Potter, 1993).
23
Together, these data indicate that HOTTIP RNA controls activation of distal Hox
genes in vivo.
The broad impact of HOTTIP RNA on gene activation across the HOXA locus
is reminiscent of the broad domains of chromatin modifications demarcating active
and silent chromosomal domains (Rinn et al., 2007). 5C analysis of control and
HOTTIP-depleted cells showed little change in higher order chromosomal structure,
indicating that the chromosomal looping is pre-configured and upstream of gene
expression (Figure 2S6A). In contrast, HOTTIP RNA knockdown led to broad loss of
H3K4me3 and H3K4me2 across the HOXA locus, most prominently over 5′ HOXA
and HOTTIP gene itself (Figure 2.3A, Figure 2S6B-7). HOTTIP RNA knockdown
also increased H3K27me3 focally over HOTTIP gene, but had little impact on
H3K27me3 across HOXA. These results indicate that HOTTIP RNA is required for
maintenance of H3K4me3 across the HOXA. These findings also imply that loss of 5′
HOXA gene transcription upon HOTTIP RNA knockdown is likely to be due to loss of
H3K4me3 (or other changes) rather than ectopic spread of H3K27me3.
24
Figure 2.3. HOTTIP RNA is required for the active chromatin state of 5′ HOXA
cluster. a, Knockdown of HOTTIP RNA broadly decreases H3K4me3 across 5′ HOXA
locus but focally affects H3K27me3 at HOTTIP gene. Display is as in Figure 2.1A. b,
Knockdown of HOTTIP RNA abrogates peaks of MLL1 and WDR5 occupancy near
TSSs of 5′ HOXA genes and leads to accumulation of these proteins at HOTTIP gene
itself. Arrows highlight peaks of MLL1 and WDR5 occupancy; open arrowheads
highlight chromatin state of HOTTIP gene upon HOTTIP RNA knockdown.
25
H3K4 methylation of the HOX loci is carried out by the MLL family of
complexes (Ruthenburg et al., 2007). In mammals, at least six MLL family members
of SET-domain-containing lysine methyltransferases interact with a core complex of
WDR5, ASH2L, RBBP5, as well as with other proteins, for substrate recognition and
genomic targeting (Ruthenburg et al., 2007). Genetic analyses indicate that MLL1 and
2 are most essential for HOX gene expression in fibroblasts (Wang et al., 2009), and
MLL1 in particular is recruited to promoters of HOX genes to maintain their activation
states (Guenther et al., 2005). In distally-derived human fibroblasts, MLL1 and WDR5
densely occupied extended region of the 5′ HOXA cluster, coincident with the
H3K4me3 domain, with specific ‘peaks’ of occupancy near the transcriptional start
sites (TSS) of multiple 5′ HOXA genes (Figure 2.3B). Strikingly, HOTTIP RNA
knockdown abrogated the peaks of MLL1 and WDR5 occupancy near TSS, resulting
in diffuse and less intense binding of MLL1 and WDR5 across HOXA cluster, most
prominently over the 5′ HOXA domain. HOTTIP RNA knockdown also led to
increased accumulation of MLL1 and WDR5 on HOTTIP gene itself (Figure 2S8).
Thus, HOTTIP RNA seems critical for maintaining a specific pattern of MLL complex
occupancy across the HOXA locus to facilitate H3K4me3 and active transcription.
To define the molecular link between HOTTIP RNA and MLL complex, we
reasoned that HOTTIP RNA may physically interact with one or more subunits of the
MLL complex. Purified, in-vitro-transcribed, full-length HOTTIP RNA bound
specifically to recombinant glutathione-S-transferase-conjugated WDR5 (GST–
WDR5), but not to GST, RBBP5, ASH2L, or the telomeric protein TRF1 (also known
as TERF1; Figure 2.4A-B). The C terminus of MLL1, containing the SET domain,
bound non-specifically to all RNAs, consistent with previous studies (Krajewski et al.,
2005). Immunoprecipitation of endogenous WDR5 from two different cell lines each
specifically retrieved endogenous HOTTIP RNA (Figure 2.4C), indicating that
WDR5 and HOTTIP RNA interact in living cells. Immunoprecipitation of an epitopetagged WDR5 from a stable cell line that previously enabled stoichiometric
purification of WDR5-interacting proteins (Wysocka et al., 2005) also specifically
retrieved HOTTIP RNA (Figure 2S9). Knockdown of WDR5 broadly inhibited
26
expression of 5′ HOXA genes, and also abrogated HOTTIP transcription,
demonstrating mutual interdependence between HOTTIP RNA and WDR5 (Figure
2.4D).
27
Figure 2.4. HOTTIP RNA programs active chromatin via WDR5. a, Summary of
RNA–protein interaction studies. Each of the indicated recombinant protein was
purified and used to retrieve purified HOTTIP RNA or control histone RNA in vitro.
Only GST–WDR5 specifically retrieved HOTTIP. b, HOTTIP RNA binds directly and
28
specifically to WDR5. Left, purified GST and GST–WDR5 are visualized by SDS–
PAGE and Coomassie Blue staining. Right, retrieved RNAs are quantified by qRT–
PCR. c, HOTTIP RNA binds specifically to WDR5 in cells. Immunoprecipitation (IP)
of endogenous WDR5 protein from PC3 (prostate) and T24 (bladder) carcinoma cells
specifically retrieved HOTTIP, but not control IPs with IgG or chromatin binder
SIRT6. U1 spliceosomal RNA served as negative control. d, WDR5 is required for 5′
HOXA gene expression, including HOTTIP RNA. e, HOTTIP RNA recruitment
potentiates transcription. Left, the BoxB tethering system. BoxB–RNA specifically
binds λN fused to GAL4 DNA binding domain, recruiting the complex to a UASluciferase reporter gene. After transient transfection, IP of GAL4- λN specifically
retrieves BoxB–HOTTIP RNA. Right, luciferase activity after co-transfection of the
indicated constructs. *P < 0.05 Student’s t-test comparing BoxB–LacZ versus BoxB–
HOTTIP). Sloped triangle indicates increasing input of plasmids encoding ncRNAs.
Means ± s.d. (n = 3) are shown for all panels.
29
HOTTIP RNA seems to regulate genes in cis, due to its low copy number,
distance dependence of HOXA target gene activation on endogenous HOTTIP, and the
physical proximity of HOTTIP and its target genes as seen in 5C. Indeed, ectopic
expression of HOTTIP RNA by retroviral transduction of lung fibroblasts, which do
not express HOTTIP, failed to activate expression of distal HOXA genes, and did not
change H3K4me3 and H3K27me3 patterns across HOXA (Figure 2S10). Moreover, in
foreskin fibroblasts that express endogenous HOTTIP, ectopic HOTTIP expression did
not induce 5′ HOXA genes, nor rescue the effects of depleting endogenous nascent
HOTTIP RNA (Figure 2S11). The lack of response in foreskin fibroblasts is notable
because endogenous HOTTIP RNA is active in these cells, indicating that the protein
partners of HOTTIP are all present and target genes are receptive. Ectopically
expressed HOTTIP RNA, being transcribed from retroviral insertion sites scattered
randomly in the genome, may not be able to find 5′HOXA genes. In contrast,
endogenous HOTTIP RNA is directly positioned near the 5′ HOXA genes by
chromosomal looping, allowing interaction and control.
To test the requirement of an exogenous targeting mechanism, we engineered
an allele of HOTTIP RNA that can be artificially recruited to a reporter gene. Addition
of five copies of the BoxB RNA element (Baron-Benhamou et al., 2004) to HOTTIP
RNA allows the fusion transcript to be recruited to the λN RNA binding domain fused
to a GAL4 DNA-binding domain (Figure 2.4E). Recruitment of HOTTIP RNA to a
silent GAL4 promoter is not sufficient to initiate transcription, but can significantly
boost transcription if the promoter is also bound by WDR5 and transcriptionally active
(Figure 2.4E). By uncoupling the sites of HOTTIP transcription versus HOTTIP RNA
function, this experiment indicates that the proximity of HOTTIP RNA—rather than
the act of transcription—maintains target gene expression. To further support the
functionality of HOTTIP RNA, deletion analysis identified a ~1 kb domain in the 5′ of
HOTTIP RNA (HOTTIPExons 1–2) that retains WDR5 binding activity (Figure 2S12A).
Enforced overexpression of HOTTIPExons 1–2 in foreskin fibroblasts inhibited 5′ HOXA
gene expression in an apparently dominant negative manner (Figure 2S12B).
30
2.3 Discussion
In summary, HOTTIP RNA is a key locus control element of HOXA genes and
distal identity. Chromosomal looping brings HOTTIP RNA in close proximity to the
5′ HOXA genes. HOTTIP transcription acts as a switch to produce HOTTIP lincRNA,
which binds to and targets WDR5–MLL complexes to the 5′ HOXA locus, yielding a
broad domain of H3K4me3 and transcription activation (Figure 2S13). The mutual
interdependence between HOTTIP RNA and WDR5 creates a positive feedback loop
that maintains the ON state of the locus. These findings provide an integrated view
linking three dimensional genome organization to dynamic programming of chromatin
states, and ultimately to developmental pattern formation.
H3K4 methylation is a feature of almost all transcribed genes, and MLL family
proteins are involved in many cell fate decisions in development and disease
(Ruthenburg et al., 2007). Our findings suggest that additional lincRNAs, especially
those associated with enhancers or enhancer-like activities (De Santa et al., 2010; Kim
et al., 2010; Orom et al., 2010), may also be involved in gene activation by
programming active chromatin states, and highlight WDR5 and other WD40 repeat
proteins as candidate adaptors that link chromatin remodelling complexes to lincRNAs.
Cis-restricted lincRNAs may be ideally suited to link chromosome structure and gene
expression. Because such lincRNA can only act on its neighbours in space,
information in higher order chromosomal looping can be faithfully transmitted to
chromatin modification via RNA recruitment of enzymatic activities, and thus into
gene expression.
31
2.4 Materials and Methods
Cells
Primary human fibroblasts derived from different anatomic sites were as
described (Bernstein et al., 2005; Chang et al., 2002; Chang et al., 2004; Rinn et al.,
2006; Rinn et al., 2007; Rinn et al., 2008a; Rinn et al., 2008b). Primary human
fibroblasts in culture retain their positional identity and have been used to examine
chromatin states associated with positional memory, which have been confirmed in
vivo (Bernstein et al., 2005; Rinn et al., 2006; Soshnikova and Duboule, 2009).
Chromatin immunoprecipitation followed by microarray analysis
ChIP-chip was performed using anti-H3K27me3 (Abcam), anti-H3K4me3
(Abcam), anti-H3K4me2 (Abcam), anti-histone H3 (Abcam), anti-PolII (Abcam),
anti-MLL1 (gift of R. Roeder), and anti-WDR5 (Wysocka et al., 2005) antibodies as
previously described (Rinn et al., 2007). Chromatin from each indicated cell type or
RNAi treatment is split into multiple tubes and subject to ChIP with different
antibodies in parallel. Retrieved DNA and input chromatin were competitively
hybridized to custom tiling arrays interrogating human HOX loci at 5-bp resolution as
previously described (Rinn et al., 2007).
5C analysis of the ENm010 HoxA1 region
5C primers were designed at HindIII restriction sites using 5C primer design
tools previously developed (Dostie et al., 2006) and made available online at
http://my5C.umassmed.edu (Lajoie et al., 2009). Reverse primers were designed for
fragments overlapping a known transcription start site from GENCODE transcripts
(Harrow et al., 2006), or overlapping a start site as experimentally determined by
CAGE Tag data of the ENCODE pilot project (Birney et al., 2007). Forward primers
were designed for all other HindIII restriction fragments. Primers were excluded if
highly repetitive sequences prevented the design of a sufficiently unique 5C primer.
Primers settings were: U-BLAST: 3; S-BLAST: 130: 15-MER: 1320; MIN_FSIZE: 40;
MAX_FSIZE: 50000; OPT_TM: 65; OPT_PSIZE: 40. DNA sequence of the universal
32
tails of forward primers was CCTCTCTATGGGCAGTCGGTGAT; DNA sequence
for the universal tails of reverse primers was AGAGAATGAGGAACCCGGGGCAG.
A 6-base barcode was included between the specific part of the primers and the
universal tail. In total 17 reverse primers and 90 forward primers were designed in the
500 kb HoxA1 locus (ENm010) and hence a total of 1,530 cis interaction were
interrogated in this region. Primer sequences are available separately (Table S1).
3C was performed with HindIII as previously described (Dostie and Dekker,
2007) separately for fetal lung and foreskin fibroblasts (FB) and also for the control
and HOTTIP knockdown foreskin FBs. For the 5C reaction, a total of 107 forward and
reverse primers of HoxA1 region were mixed with either the ENCODE random region
(ENr) primer pool comprising of 2,673 forward and 523 reverse primers (covering 30
additional ENCODE regions) or the ENr313 primer pool comprising of 57 forward
and 58 reverse primers (covering 1 additional ENCODE region). 5C was then
performed in 10 reactions each containing an amount of 3C library that represents
200,000 genome equivalents and 1 fmol of each primer. The 5C analysis of HoxA1
region was carried out in two biological replicates of fetal lung and foreskin FBs. 5C
ligation products were amplified using a pair of universal primers that recognize the
common tails of the 5C forward and reverse primers described above and pooled
together. To facilitate paired-end DNA sequence analysis on the Illumina GA2
platform, paired-end adaptor oligonucleotides were ligated to the 5C library using the
Illumina PE protocol and PCR amplification of the library was carried out for 18
cycles with Illumina PCR primer PE 1.0 and 2.0. The 5C library was then sequenced
on the Illumina GA2 platform generating 36 base paired end reads. For fetal lung FBs
we obtained 7,625,276 and 10,947,424 mapped reads for two biological replicates of
which 1,339,861 and 242,301 could be specifically mapped back to interactions within
ENm010 using Novoalign (http://www.novocraft.com), respectively. For two
biological replicates of foreskin FBs we obtained 7,311,386 and 5,731,107 mapped
reads of which 2,752,789 and 66,769 could be mapped back to the ENm010 region,
respectively. In the case of the knockdown study, control green fluorescent protein
(GFP) knockdown foreskin FB 5C library yielded 4,909,482 mapped reads whereas
33
HOTTIP knockdown foreskin FB had 5,565,389 mapped reads of which 39,168 and
38,950 could be mapped back to ENm010 for control GFP and HOTTIP knockdown,
respectively. In the set with fetal lung and foreskin fibroblast samples, 5C for ENm010
was multiplexed for deep sequencing with 5C of one other region, ENr313; in the set
containing the knockdown samples, ENm010 was multiplexed with 5C of 30 other
genomic regions. The different extent of multiplexing resulted in different number of
sequencing reads mapping back to ENm010. In all instances the mappable reads were
proportional to the degree of multiplexing, indicating equivalent library quality despite
different read numbers. Table S2 outlines the library composition of each experiment.
The heat maps are scaled as follows—for Figure 2.1A, distal (foreskin) FBs: 262–
17,467, proximal (lung) FBs: 7–5,846; for Figure 2S6, siGFP: 1–100, siHOTTIP: 1–
100. Raw data from the 5C experiments used to generate the binned heat maps in
Figure 2.1A and Figure 2S6 can be found in File S1. Raw data are available by
request.
HOTTIP cloning, sequence and expression analysis
We previously identified a portion of HOTTIP as a non protein-coding
transcribed region named ncHOXA13-96 (Rinn et al., 2007). This region also overlaps
expressed sequence tag (EST) clone AK093987 that was previously observed to be
expressed in cancer cell lines derived from posterior anatomic sites (Sasaki et al.,
2007). 5′ and 3′ RACE (RLM Race kit, Applied Biosystems/Ambion) showed fulllength HOTTIP RNA to be 3,764 nucleotides, extending the known transcribed region
by more than 1,400 bases. BLAST and BLAT confirmed that portions of HOTTIP are
well conserved in mammals and even in avians but had no protein coding potential.
Full-length HOTTIP RNA sequence has been deposited at NCBI (accession number
GU724873). qRT–PCR with SYBR Green was conducted as recommended by the
manufacturer (Agilent Technologies). Primer sequences specific for HOTTIP were
CCTAAAGCCACGCTTCTTTG (HOTTIP-F) and TGCAGGCTGGAGATCCTACT
(HOTTIP-R). For Figure 2S11, endogenous nascent HOTTIP was distinguished from
34
ectopic HOTTIP expressed from cDNA using primers that spanned intron–exon
junctions.
Strand-specific RT–PCR
RNA extracted from primary foreskin fibroblasts was reverse transcribed
(SuperScript III, Invitrogen) using combinations of the previously described HOTTIPspecific primers HOTTIP-F and/or HOTTIP-R as diagrammed in Figure 2S1B.
Resulting cDNA was then PCR-amplified using both HOTTIP-F and HOTTIP-R
primers to visually determine strand specificity.
HOTTIP transcript count per cell
The level of HOTTIP transcript per cell was calculated from the level of
HOTTIP in 500,000 cells. Full-length HOTTIP in pcDNA3.1+ was assayed by qPCR
using primers HOTTIP-F and HOTTIP-R at predetermined concentrations in triplicate
to generate a linear amplification curve dependent on the moles of template DNA
(Figure 2S2). The qRT–PCR value from 500,000 foreskin fibroblasts was determined
and plotted, and the corresponding total molecules of transcript was divided by
500,000 to determine the approximate number of transcripts per cell.
Single-molecule RNA fluorescence in situ hybridization (RNA-FISH)
Single molecule RNA-FISH was performed as described in (Raj et al., 2008)
with the following modifications: the amount of hybridization solution per chamber
was doubled to allow for proper coating of the chamber and the amount of glucoseoxidase buffer was tripled to assist in image acquisition. Images were acquired using
an Olympus FV1000 confocal microscope within 2 h of the addition of the glucoseoxidase buffer.
RNA interference
Primary foreskin fibroblasts were transfected with siRNAs targeting HOTTIP
and WDR5 using Lipofectamine 2000 (Invitrogen) as per manufacturer’s instructions.
Total RNA was harvested 48–72 h later using TRIzol (Invitrogen) and RNeasy Mini
35
Kits (Qiagen) as previously described (Rinn et al., 2006). For the intronic HOTTIP
knockdown experiment in Figure 2S11, a pool of 10 siRNAs (Table S3) targeting
intronic regions in HOTTIP were transfected into foreskin fibroblasts, and RNA
isolated as above.
Generation of shRNAs against chicken HOTTIP
A reporter construct encoding a GFP–chicken HOTTIP fusion transcript was
used in a small-scale screen to identify highly effective shRNA constructs. Eleven
shRNAs targeting conserved regions of chicken HOTTIP were designed and inserted
into the pSMP system (Thermo/Open Biosystems). The reporter construct and shRNA
constructs were cotransfected into Phoenix cells, and HOTTIP transcript levels were
analysed via reduced GFP fluorescence and by qRT–PCR. Three shRNAs that were
effective in vitro were then cloned into RCAS vector for studies in chick embryos
(Harpavat and Cepko, 2006).
Chick RNAi
RCAS HOTTIP hairpin and RCAS AP viruses were made by transfecting DF1 cells with viral DNA. Transfected DF-1 cells were grown and passaged, after which
the virus-containing supernatant was collected, concentrated and titred. Fertilized
chicken eggs were incubated in a humidified rotating incubator at 37 °C until they
reached Hamilton/Hamburger stage 10. Eggs were then windowed to expose the
embryos. After gently removing the vitelline membrane, chicken embryos were
microinjected with RCAS-HOTTIP hairpins and RCAS-AP viruses at the prospective
wing and leg buds. All viral stocks have titres of 1 × 108 IU ml−1, and each limb was
injected five times. The infected embryos were allowed to incubate at 37 °C and were
harvested 2 or 4 days after injection to detect viral infection by immunohistochemistry.
Total RNA was extracted from injected forelimbs, and RT–PCR analysis was
performed 4 days after injection. Chicken embryos were harvested 9 days postinjection to carry out whole-mount Alcian blue staining. A total of 50 animals were
injected.
36
Hairpin sequences for chick HOTTIP were
TGCTGTTGACAGTGAGCGACCCGAAGATGTGTCTGATTTGTAGTGAAGCCA
CAGATGTACAAATCAGACACATCTTCGGGCTGCCTACTGCCTCGGA (2-2-1),
TGCTGTTGACAGTGAGCGCCGCTCTGCTCTCCTCTCTCTCTAGTGAAGCCA
CAGATGTAGAGAGAGAGGAGAGCAGAGCGATGCCTACTGCCTCGGA (3-21), and
TGCTGTTGACAGTGAGCGAATCCTTAATCGAATCTGATTTTAGTGAAGCCA
CAGATGTAAAATCAGATTCGATTAAGGATCTGCCTACTGCCTCGGA (4-4-1).
HOTTIP overexpression
Full-length HOTTIP and a truncated transcript consisting of exons 1 and 2
(HOTTIPExons 1–2) were cloned into the LZRS vector (gift of P. Khavari), and then
transfected into Phoenix cells (gift of G. Nolan) to generate amphotropic retroviruses.
Primary human fibroblasts were infected with either LZRS-full length HOTTIP (lung),
LZRS-truncated HOTTIP (foreskin), or LZRS-GFP (both lung and foreskin), then
passaged over 60 days, with periodic testing of HOXA and HOTTIP expression by
qRT–PCR. These cells were used in the rescue experiments depicted in Figure 2S11.
GST pull-down
Full-length HOTTIP, truncated HOTTIP containing exons 1 and 2
(HOTTIPExons 1–2), and histone H2B1 mRNA were transcribed in vitro using T7
polymerase according to manufacturer’s instructions (Promega), denatured, and
refolded in folding buffer (100 mM KCl, 10 mM MgCl2, Tris pH 7.0). GST-tagged
WDR5, C-terminal MLL1, RBBP5/Ash2L and TRF1 were expressed in Escherichia
coli and purified as described (Smith and Johnson, 1988). Each GST-fusion protein
was bound to glutathione beads (Amersham/GE Healthcare) and blocked with excess
yeast total mRNA in PB100 buffer (20 mM HEPES pH 7.6, 100 mM KCl, 0.05%
NP40, 1 mM DTT, 0.5 mM PMSF) for 1 h at room temperature. Beads were then
incubated with either in-vitro-transcribed HOTTIP or histone H2B1 mRNA for 45 min
at room temperature. After three washes in PB200 buffer (20 mM HEPES pH 7.6,
37
200 mM KCl, 0.05% NP40, 1 mM DTT, 0.5 mM PMSF), bound RNAs were extracted
and analysed by qRT–PCR, as previously described.
RNA immunoprecipitation
HeLa-WDR5-Flag cells: 48 h after Lipofectamine 2000-mediated transfection
of HOTTIP into HeLa WDR5-Flag cells (approximately 107 cells), total protein was
extracted as previously described, with modifications (Dignam et al., 1983). Briefly,
cells were resuspended in Buffer A (10 mM HEPES pH 7.5, 1.5 mM MgCl2, 10 mM
KCl, 0.5 mM DTT, 1.0 mM PMSF), lysed in 0.25% NP40, and fractionated by low
speed centrifugation. The nuclear fraction was resuspended and lysed in Buffer C
(20 mM HEPES pH 7.5, 10% glycerol, 0.42 M KCl, 4 mM MgCl2, 0.5 mM DTT,
1.0 mM PMSF). Combined nuclear and cytoplasmic fractions were
immunoprecipitated with mouse anti-Flag M2 monoclonal antibody (Sigma) or mouse
IgG affixed to agarose beads (Sigma) for 3 to 4 h at 4 °C. Beads were washed four
times with wash buffer (50 mM TrisCl pH 7.9, 10% glycerol, 100 mM KCl, 5 mM
MgCl2, 10 mM β-mercaptoethanol, 0.1% NP40). After elution using Flag peptide
(Sigma), co-immunoprecipitated RNA was extracted and analysed by qRT–PCR.
Endogenous WDR5 and SIRT6 RIP: cellular fractions were isolated as above and
incubated with the anti-WDR5 (Millipore 07-706) or anti-Sirt6 (ab62739, Abcam)
antibodies overnight at 4 °C. Samples were washed in wash buffer, and coimmunoprecipitated RNA was extracted and analysed by qRT–PCR.
RNA chromatography
Full-length in-vitro-transcribed HOTTIP RNA was conjugated to adipic acid
dehydrazide agarose beads as described (Michlewski and Caceres, 2010). The
complexed beads were incubated with whole cell lysates from Hela WDR5-Flag cells,
washed, and bound proteins visualized by western blotting.
BoxB tethering assay
293T cells were grown to about 50% confluence in 6-well plates on the day of
transfection. Using Lipofectamine 2000 (Invitrogen), a plasmid encoding a luciferase
38
gene under the control of five tandem GAL4 UAS sites were co-transfected with
plasmids encoding GAL4-WDR5, GAL4-λN (the 22 amino acid RNA-binding domain
of the lambda bacteriophage antiterminator protein N) peptide fused to a C-terminal
GFP tag, BoxB (containing five repeats of the λN-specific 19 nucleotide binding site),
BoxB fused to full-length LacZ, or BoxB fused to full-length HOTTIP. Cells were
lysed 48 h after transfection, and luciferase assay kit (Promega) was used to determine
relative levels of the luciferase gene product, following the manufacturer’s protocol.
39
2.5 Supplementary Figures
Figure 2S1. Molecular characterization of HOTTIP. (A) ChIP-chip confirmed that
Suz12, a subunit of PRC2, and MLL1 both occupy HOTTIP in human lung fibroblasts,
where HOTTIP RNA is not expressed. Chromatin occupancy map across the ~100
kilobase HOXA locus is shown. Genomic coordinate is on the X-axis; Y-axis depicts
relative occupancy of MLL1, Suz12, and RNA polymerase II (Pol II) on a linear scale.
Arrows depict MLL1 and Suz12 binding on HOTTIP DNA. Note that the peak of
MLL1 occupancy coincides with diminished Suz12 binding, which are flanked by
regions of higher Suz12 occupancy. This reciprocal binding of MLL and Polycomb
subunits, even when both are present and juxtaposed, have been previously observed
in bivalent domains21. (B) Strand-specific RT-PCR of HOTTIP RNA. Primer
combinations used for reverse transcription (1, 2) and HOTTIP PCR (F, R) are
40
diagrammed on the schematic. (C) Conserved distal expression pattern of chick
HOTTIP. In situ hybridization of stage 28 chick embryo limbs are shown. White scale
bar=0.1 mm. (D) Chromatin marks at the distal HOXA locus in human embryonic
stem cells (ESC). In ES cells, the entire HOXloci, including HOTTIP, is marked by a
broad domain of H3K27me3; the HOTTIP locus has no H3K4me3 or H3K36me3
marks, consistent with its silent status. Genome graphic data exported from
http://genome.ucsc.edu/, utilizing the ENCODE whole genome data sets22.
41
Figure 2S2. Single molecule RNA-fluorescence in situ hybridization (FISH) confirms
low copy number of HOTTIP RNA. (A) Primary human foreskin fibroblasts were
hybridized with a mixture of Cy-5-labeled probes against HOTTIP and assayed for
HOTTIP expression via confocal microscopy. Representative confocal images
showing an example of a cell with one (left) or no (right) copies of HOTTIP RNA
(green, highlighted by arrow). DAPI staining of cellular nuclei is in red. White scale
bar=15μm. (B) Quantitation of RNA-FISH experiments, demonstrating that the
majority (>97%) of the cells examined contain zero or one copy of HOTTIP RNA.
HOTTIP RNA is present on on average at 0.31 copies per cell. The counts were
pooled from 3 independent replicates (n=449 cells). (C) Titration curve used to
determine number of HOTTIP molecules per 500,000 cells. Y-axis depicts 2^-ct
values from qRT-PCR; X-axis depicts scaled total molecules of transcript in a sample.
Blue points outline the titration curve using a known quantity of HOTTIP molecules to
determine corresponding qRT-PCR 2^-ct values; error bars represent s.d. from three
technical replicates. The pink square represents the qRT-PCR value from a standard
sample of 500,000 foreskin fibroblasts.
42
Figure 2S3. Lack of antisense HOTTIP RNA with HOTTIP knockdown. (A) No
antisense transcription upon HOTTIP depletion. RNA from foreskin fibroblasts treated
with control or HOTTIP siRNA was reverse transcribed with strand-specific primers,
and then transcript abundance is analyzed by qPCR. N.D. not detectable. (B)
Independent siRNAs targeting HOTTIP abrogated 5’ HOXA gene expression in
primary fibroblasts. Mean + s.d. are shown.
43
Figure 2S4. Efficient retroviral infection in developing chick limb buds. (A) Broad
expression of the avian virus protein Gag in mesenchymal cells of chick forelimb 2
days after injection with RCAS-shHOTTIP. (B) Higher magnification of A showing
detection of Gag protein in mesenchymal and muscle cells of the injected limb. (C)
After 4 days of injection, Gag protein was detected in limb chondrocytes. (D) Alkaline
phosphatase activities were detected in hindlimb injected with control RCAS-AP. r,
radius, u, ulna.
44
Figure 2S5. HOTTIP knockdown decreases expression of 5’ HoxA genes in vivo.
Replication competent avian virus (RCAS) expressing short hairpin RNA targeting
HOTTIP was injected into chick limb buds at St. 17. After 4 days of incubation,
embryos were harvested and prepared for sections and in situ hybridization.
Contralateral limbs served as controls. Shown are forelimb zeugopod and autopod (i.e.
elbow to fingertip). (A)HOTTIP expression in the limb bud is reduced (B) following
over-expression of HOTTIP shRNA. In control limbs (C,E,G), HoxA13(C), HoxA11
(E), and HoxA10 (G) expression domains are strongest in distal, undifferentiated limb
45
mesenchyme. In treated limbs (D,F,H) the expression domains of the HoxA genes are
reduced in size and fewer transcripts are detectable. Abbreviations: r=radius (anterior
side of limb); u=ulna; d= digit, D=distal tip of limb bud; P= proximal end .
46
Figure 2S6. HOTTIP depletion causes little change to higher-order chromosome
configuration. (A) Chromosome Conformation Capture (5C) analysis of foreskin
fibroblasts treated with control siRNAs (siGFP) or siRNA targeting HOTTIP.
Heatmap representations of 5C data (bin size 30 kb, step size 3 kb) across human
HOXA are as in Figure 2.1A. The diagonal represents frequent cis interactions
between regions located in proximity to each other in the HoxA locus. Long-range
looping interactions are 5C signals that are away from the diagonal. Each pixel
represents median interactions in a 30kb region; the intensity of pixel corresponds to
the total number of reads. (B) Knockdown of HOTTIP broadly decreases H3K4me2
across 5’ HOXA locus but does not change overall occupancy of histone H3. ChIPchip data across ~100 kilobase HOXAlocus are shown for foreskin fibroblasts with
47
control (top) versus HOTTIP (bottom) knockdown. Genomic coordinate is on the Xaxis; Y-axis depicts relative occupancy of H3K4me2 and pan histone H3 on a linear
scale.
48
Figure 2S7. ChIP-qPCR validation of HOTTIP-dependence of H3K4me3 occupancy
at distal HOXA. Chromatin-immunoprecipitation with H3K4me3 was performed on
foreskin fibroblasts transfected with siRNA against GFP (siGFP) or HOTTIP
(siHOTTIP), and the subsequent enrichment of distal (HOXA13,HOXA11) and
proximal (HOXA1)HOXA genes were analyzed by quantitative PCR. Mean + s.d. are
shown.
49
Figure 2S8. Zoom-in view of Figure 2.3 highlighting HOTTIP dependence of MLL1
and WDR5 localization to HOXA. Knockdown of HOTTIP abrogates MLL1(A) and
WDR5 (B) peaks across the 5’ HOXA locus, and results in diffuse and less intense
binding of MLL1 and WDR5 across the 5’HOXA. HOTTIP knockdown also led to
increased accumulation of MLL1 and WDR5 on HOTTIP itself. Arrows highlight
peaks of MLL1 and WDR5 occupancy.
50
Figure 2S9. HOTTIP binds stably expressed FLAG-WDR5 in HeLa cells.
(A) HOTTIP expression levels in primary foreskin fibroblasts, FLAG-WDR5 HeLa
cells, and FLAG-WDR5 HeLa cells transduced with HOTTIP. Error bars represent
mean + s.d.. (B) HOTTIP binds specifically to WDR5 in cells. Top: IP of FLAGWDR5 verified by WDR5 immunoblot. Bottom: Quantitation of RIP by qRT-PCR,
normalized by the levels of RNA retrieved by IgG control. (C) RNA binding by H3K4
methylase complexes. Extract from from FLAG-WDR5 HeLa cells were passed over
beads (control) or bead coupled to full-length HOTTIP RNA, extensively washed, and
interrogated for retrieval of specific proteins by immunoblot. HOTTIP retrieved
WDR5, Menin (unique to MLL1/MLL2 complexes), and WDR82 (unique to the
Set1A/Set1B), but not PTIP (unique to MLL3/MLL4 complexes).
51
Figure 2S10. HOTTIP overexpression does not affect distal HOXA expression. (A)
Ectopic overexpression of HOTTIP in lung fibroblasts does not activate 5’HOXA
genes. (B) HOTTIP overexpression in lung fibroblasts do not change HOXA
chromatin state. ChIP-chip data across ~100 kilobase HOXA locus are shown for lung
fibroblasts overexpressing HOTTIP (top) compared those from regular lung
52
fibroblasts (bottom). Genomic coordinate is on the X-axis; Y-axis depicts relative
occupancy of H3K27me3, H3K4me3, and RNA polymerase II (Pol II) on a linear
scale.
53
Figure 2S11. Ectopic HOTTIP expression does not activate 5’HOXA genes nor rescue
the effects of depleting endogenous nascent HOTTIP. (A) HOTTIP overexpression in
foreskin fibroblasts is unable to activate 5’HOXA genes, and has apparently dominant
Wang et al., p. 6 negative effect on HOXA13 expression. GFP transduction served as
negative control, which did not change HOTTIP level or 5’ HOXA genes. (B)
Knockdown of endogenous nascent HOTTIP was achieved by siRNAs targeting
intronic regions in HOTTIP, leading to decreased expression of 5’ HOXA genes.
siRNAs targeting HOTTIP introns do not affect ectopic HOTTIP overexpression from
54
cDNA (>100 fold in both siScr vs. siHOTTIP-intronic treated cells compared to wild
type). Cells transduced with ectopic HOTTIP cDNA responded to endogenous
HOTTIP depletion in the same manner, indicating that ectopic HOTTIP cannot rescue
depletion of endogenous HOTTIP. Mean + s.d. are shown. *, p<0.05, Student’s t-test.
N.S., not significantly different between control and ectopic HOTTIP overexpression.
55
Figure 2S12. HOTTIPExons1-2 binds to WDR5 and acts in a dominant negative
manner to inhibit 5’ HOXA gene expression. (A) Schematic of HOTTIP fragments and
their relative affinities in binding to GST-WDR5. Each fragment was in vitro
transcribed and purified, and incubated with purified recombinant GST or GSTWDR5. Specific retrieval of RNA by GST-WDR5 was quantified by qRT-PCR. +
indicates higher interaction with GST-WDR5 than GST. Notably, HOTTP exons 1-2
bound GST-WDR5 ~2 fold better than full length HOTTIP on a molar basis. Mean
+s.d. is shown. (B) Overexpression of HOTTIPExons1-2 in foreskin fibroblasts exerts
a dominant negative effect, resulting in reduced 5’ HOXA gene expression.Error bars
represent s.d. from biologic duplicates.Bottom:Overexpression is confirmed by RTPCR using primers specific for HOTTIPExons1-2 and GFP; GAPDH is included as a
loading control.
56
Figure 2S13. Model of HOTTIP action. HOTTIP is positioned near the active 5’
HOXA genes via chromosomal looping. The top panel depicts the situation in distal
cells when HOTTIP is depleted: the chromosome is looped to bring 5’ HOXA genes
together, but MLL1/WDR5 is not properly localized, nor is there H3K4me3 and gene
transcription. Upon transcription (bottom panel), HOTTIP RNA binds to WDR5MLL1 complex, and drives MLL1/WDR5 occupancy and H3K4 trimethylation of 5’
HOXA genes. Because HOTTIP transcription is dependent on WDR5 and WDR5
occupancy is dependent on HOTTIP, this mutual interdependence creates a positive
feedback loop that maintains the ON state of the 5’ HOXA locus.
57
2.7 References
Ashe, H.L., Monks, J., Wijgerde, M., Fraser, P., and Proudfoot, N.J. (1997). Intergenic
transcription and transinduction of the human beta-globin locus. Genes Dev 11,
2494-2509.
Baron-Benhamou, J., Gehring, N.H., Kulozik, A.E., and Hentze, M.W. (2004). Using
the lambdaN peptide to tether proteins to RNAs. Methods Mol Biol 257, 135154.
Bernstein, B.E., Kamal, M., Lindblad-Toh, K., Bekiranov, S., Bailey, D.K., Huebert,
D.J., McMahon, S., Karlsson, E.K., Kulbokas, E.J., 3rd, Gingeras, T.R., et al.
(2005). Genomic maps and comparative analysis of histone modifications in
human and mouse. Cell 120, 169-181.
Bernstein, B.E., Mikkelsen, T.S., Xie, X., Kamal, M., Huebert, D.J., Cuff, J., Fry, B.,
Meissner, A., Wernig, M., Plath, K., et al. (2006). A bivalent chromatin
structure marks key developmental genes in embryonic stem cells. Cell 125,
315-326.
Birney, E., Stamatoyannopoulos, J.A., Dutta, A., Guigo, R., Gingeras, T.R., Margulies,
E.H., Weng, Z., Snyder, M., Dermitzakis, E.T., Thurman, R.E., et al. (2007).
Identification and analysis of functional elements in 1% of the human genome
by the ENCODE pilot project. Nature 447, 799-816.
Chang, H.Y. (2009). Anatomic demarcation of cells: genes to patterns. Science 326,
1206-1207.
Chang, H.Y., Chi, J.T., Dudoit, S., Bondre, C., van de Rijn, M., Botstein, D., and
Brown, P.O. (2002). Diversity, topographic differentiation, and positional
memory in human fibroblasts. Proc Natl Acad Sci U S A 99, 12877-12882.
Chang, H.Y., Sneddon, J.B., Alizadeh, A.A., Sood, R., West, R.B., Montgomery, K.,
Chi, J.T., van de Rijn, M., Botstein, D., and Brown, P.O. (2004). Gene
expression signature of fibroblast serum response predicts human cancer
progression: similarities between tumors and wounds. PLoS Biol 2, E7.
Davis, A.P., Witte, D.P., Hsieh-Li, H.M., Potter, S.S., and Capecchi, M.R. (1995).
Absence of radius and ulna in mice lacking hoxa-11 and hoxd-11. Nature 375,
791-795.
De Santa, F., Barozzi, I., Mietton, F., Ghisletti, S., Polletti, S., Tusi, B.K., Muller, H.,
Ragoussis, J., Wei, C.L., and Natoli, G. (2010). A large fraction of extragenic
RNA pol II transcription sites overlap enhancers. PLoS Biol 8, e1000384.
Dean, A. (2006). On a chromosome far, far away: LCRs and gene expression. Trends
Genet 22, 38-45.
Dignam, J.D., Lebovitz, R.M., and Roeder, R.G. (1983). Accurate transcription
initiation by RNA polymerase II in a soluble extract from isolated mammalian
nuclei. Nucleic Acids Res 11, 1475-1489.
Dinger, M.E., Amaral, P.P., Mercer, T.R., Pang, K.C., Bruce, S.J., Gardiner, B.B.,
Askarian-Amiri, M.E., Ru, K., Solda, G., Simons, C., et al. (2008). Long
noncoding RNAs in mouse embryonic stem cell pluripotency and
differentiation. Genome Res 18, 1433-1445.
Dostie, J., and Dekker, J. (2007). Mapping networks of physical interactions between
genomic elements using 5C technology. Nat Protoc 2, 988-1002.
58
Dostie, J., Richmond, T.A., Arnaout, R.A., Selzer, R.R., Lee, W.L., Honan, T.A.,
Rubio, E.D., Krumm, A., Lamb, J., Nusbaum, C., et al. (2006). Chromosome
Conformation Capture Carbon Copy (5C): a massively parallel solution for
mapping interactions between genomic elements. Genome Res 16, 1299-1309.
Fromental-Ramain, C., Warot, X., Messadecq, N., LeMeur, M., Dolle, P., and
Chambon, P. (1996). Hoxa-13 and Hoxd-13 play a crucial role in the
patterning of the limb autopod. Development 122, 2997-3011.
Guenther, M.G., Jenner, R.G., Chevalier, B., Nakamura, T., Croce, C.M., Canaani, E.,
and Young, R.A. (2005). Global and Hox-specific roles for the MLL1
methyltransferase. Proc Natl Acad Sci U S A 102, 8603-8608.
Harpavat, S., and Cepko, C.L. (2006). RCAS-RNAi: a loss-of-function method for the
developing chick retina. BMC Dev Biol 6, 2.
Harrow, J., Denoeud, F., Frankish, A., Reymond, A., Chen, C.K., Chrast, J., Lagarde,
J., Gilbert, J.G., Storey, R., Swarbreck, D., et al. (2006). GENCODE:
producing a reference annotation for ENCODE. Genome Biol 7 Suppl 1, S4 19.
Kim, T.K., Hemberg, M., Gray, J.M., Costa, A.M., Bear, D.M., Wu, J., Harmin, D.A.,
Laptewicz, M., Barbara-Haley, K., Kuersten, S., et al. (2010). Widespread
transcription at neuronal activity-regulated enhancers. Nature 465, 182-187.
Kmita, M., Fraudeau, N., Herault, Y., and Duboule, D. (2002). Serial deletions and
duplications suggest a mechanism for the collinearity of Hoxd genes in limbs.
Nature 420, 145-150.
Kmita, M., Tarchini, B., Zakany, J., Logan, M., Tabin, C.J., and Duboule, D. (2005).
Early developmental arrest of mammalian limbs lacking HoxA/HoxD gene
function. Nature 435, 1113-1116.
Krajewski, W.A., Nakamura, T., Mazo, A., and Canaani, E. (2005). A motif within
SET-domain proteins binds single-stranded nucleic acids and transcribed and
supercoiled DNAs and can interfere with assembly of nucleosomes. Mol Cell
Biol 25, 1891-1899.
Lajoie, B.R., van Berkum, N.L., Sanyal, A., and Dekker, J. (2009). My5C: web tools
for chromosome conformation capture studies. Nat Methods 6, 690-691.
Mercer, T.R., Dinger, M.E., and Mattick, J.S. (2009). Long non-coding RNAs:
insights into functions. Nat Rev Genet 10, 155-159.
Michlewski, G., and Caceres, J.F. (2010). RNase-assisted RNA chromatography. RNA
16, 1673-1678.
Nelson, C.E., Morgan, B.A., Burke, A.C., Laufer, E., DiMambro, E., Murtaugh, L.C.,
Gonzales, E., Tessarollo, L., Parada, L.F., and Tabin, C. (1996). Analysis of
Hox gene expression in the chick limb bud. Development 122, 1449-1466.
Orom, U.A., Derrien, T., Beringer, M., Gumireddy, K., Gardini, A., Bussotti, G., Lai,
F., Zytnicki, M., Notredame, C., Huang, Q., et al. (2010). Long noncoding
RNAs with enhancer-like function in human cells. Cell 143, 46-58.
Petruk, S., Sedkov, Y., Riley, K.M., Hodgson, J., Schweisguth, F., Hirose, S., Jaynes,
J.B., Brock, H.W., and Mazo, A. (2006). Transcription of bxd noncoding
RNAs promoted by trithorax represses Ubx in cis by transcriptional
interference. Cell 127, 1209-1221.
59
Ponting, C.P., Oliver, P.L., and Reik, W. (2009). Evolution and functions of long
noncoding RNAs. Cell 136, 629-641.
Raj, A., van den Bogaard, P., Rifkin, S.A., van Oudenaarden, A., and Tyagi, S. (2008).
Imaging individual mRNA molecules using multiple singly labeled probes. Nat
Methods 5, 877-879.
Rinn, J.L., Bondre, C., Gladstone, H.B., Brown, P.O., and Chang, H.Y. (2006).
Anatomic demarcation by positional variation in fibroblast gene expression
programs. PLoS Genet 2, e119.
Rinn, J.L., Kertesz, M., Wang, J.K., Squazzo, S.L., Xu, X., Brugmann, S.A.,
Goodnough, L.H., Helms, J.A., Farnham, P.J., Segal, E., et al. (2007).
Functional demarcation of active and silent chromatin domains in human HOX
loci by noncoding RNAs. Cell 129, 1311-1323.
Rinn, J.L., Wang, J.K., Allen, N., Brugmann, S.A., Mikels, A.J., Liu, H., Ridky, T.W.,
Stadler, H.S., Nusse, R., Helms, J.A., et al. (2008a). A dermal HOX
transcriptional program regulates site-specific epidermal fate. Genes Dev 22,
303-307.
Rinn, J.L., Wang, J.K., Liu, H., Montgomery, K., van de Rijn, M., and Chang, H.Y.
(2008b). A systems biology approach to anatomic diversity of skin. J Invest
Dermatol 128, 776-782.
Ruthenburg, A.J., Allis, C.D., and Wysocka, J. (2007). Methylation of lysine 4 on
histone H3: intricacy of writing and reading a single epigenetic mark. Mol Cell
25, 15-30.
Sanchez-Elsner, T., Gou, D., Kremmer, E., and Sauer, F. (2006). Noncoding RNAs of
trithorax response elements recruit Drosophila Ash1 to Ultrabithorax. Science
311, 1118-1123.
Sasaki, Y.T., Sano, M., Kin, T., Asai, K., and Hirose, T. (2007). Coordinated
expression of ncRNAs and HOX mRNAs in the human HOXA locus. Biochem
Biophys Res Commun 357, 724-730.
Schuettengruber, B., Chourrout, D., Vervoort, M., Leblanc, B., and Cavalli, G. (2007).
Genome regulation by polycomb and trithorax proteins. Cell 128, 735-745.
Small, K.M., and Potter, S.S. (1993). Homeotic transformations and limb defects in
Hox A11 mutant mice. Genes Dev 7, 2318-2328.
Smith, D.B., and Johnson, K.S. (1988). Single-step purification of polypeptides
expressed in Escherichia coli as fusions with glutathione S-transferase. Gene
67, 31-40.
Soshnikova, N., and Duboule, D. (2009). Epigenetic temporal control of mouse Hox
genes in vivo. Science 324, 1320-1323.
Wang, P., Lin, C., Smith, E.R., Guo, H., Sanderson, B.W., Wu, M., Gogol, M.,
Alexander, T., Seidel, C., Wiedemann, L.M., et al. (2009). Global analysis of
H3K4 methylation defines MLL family member targets and points to a role for
MLL1-mediated H3K4 methylation in the regulation of transcriptional
initiation by RNA polymerase II. Mol Cell Biol 29, 6074-6085.
Wysocka, J., Swigut, T., Milne, T.A., Dou, Y., Zhang, X., Burlingame, A.L., Roeder,
R.G., Brivanlou, A.H., and Allis, C.D. (2005). WDR5 associates with histone
H3 methylated at K4 and is essential for H3 K4 methylation and vertebrate
60
development. Cell 121, 859-872.
Zhang, X., Lian, Z., Padden, C., Gerstein, M.B., Rozowsky, J., Snyder, M., Gingeras,
T.R., Kapranov, P., Weissman, S.M., and Newburger, P.E. (2009). A
myelopoiesis-associated regulatory intergenic noncoding RNA transcript
within the human HOXA cluster. Blood 113, 2526-2534.
61
CHAPTER 3
Long noncoding RNA as regulatory switch of protein turnover
This chapter has been submitted as Yang, Y.W., Chen, Y., Wan, B., Qu, K., Lei, M.,
and Chang, H.Y. Long noncoding RNA as regulatory switch of protein turnover.
Submitted to Cell/Molecular Cell on 3/14/12
Y.W. Yang’s was the primary contributor to all aspects of this chapter, excepting
protein purification used for scanning alanine mutagenesis screen (Figure 3.4A-C,
Figure 3S4B), microarray data analysis (Figure 3.7A), and in vitro ubiquitination
assay (Figure 3S1B). Y.W. Yang and H.Y. Chang prepared and wrote the manuscript
with input from all authors.
62
3.1 Abstract
Long intervening noncoding RNAs (lincRNAs) are prevalent genes with
poorly understood functions. Here we discover a pathway of lincRNA-regulated
proteolysis. The enhancer-like lincRNA HOTTIP extends the half-life of its binding
protein WDR5, a subunit of the MLL H3K4 methylase complex, resulting in increased
chromatin occupancy and gene activation. LincRNA-mediated stabilization requires
direct RNA-protein interaction in a long RNA context, and blocks turnover at a step
after target protein poly-ubiquitination. We elucidate the lincRNA binding interface on
WDR5. A WDR5 mutant that selectively abrogates lincRNA binding becomes
unstable, and is defective in gene activation, maintenance of histone H3 lysine 4
trimethylation, and embryonic stem cell self renewal. The ability to modulate protein
turnover may allow lincRNAs to control the lifespan of molecular interactions at
chromatin and elsewhere, broadening their scope in epigenetics and cell fate control.
63
3.2 Introduction
Eukaryotic genomes are expansively transcribed, generating thousands of long
intervening noncoding RNA (lincRNA) transcripts ranging from several hundred
nucleotides to dozens of kilobases (Guttman et al., 2009; Guttman et al., 2010). These
lincRNAs are capped, spliced, and often polyadenylated in a manner similar to
canonical messenger RNAs, but do not appear to function by coding for proteins. Of
the large number of lincRNAs, only a small fraction has been functionally
characterized, revealing diverse functions in development and disease processes
(Wang and Chang, 2011). Intriguingly, some protein coding transcripts may also
possess lincRNA-like regulatory functions independent of their coding capacity
(Candeias et al., 2008; Zofall et al., 2012), suggesting a broad scope for RNA-based
regulation.
LincRNAs have been demonstrated to play key roles in epigenetic regulation
by controlling chromatin states (Wang and Chang, 2011). Specific lincRNAs can bind
repressive or activating chromatin modification complexes, and localize these
activities to specific gene loci. For example, the lincRNA XIST binds the Polycomb
Repressive Complex 2 (PRC2) to cause histone H3 lysine 27 trimethylation and
silence the X chromosome for dosage compensation in females (Morey and Avner,
2011). As another example, the lincRNA HOTAIR acts as a molecular scaffold,
binding both PRC2 and the H3K4 demethylase LSD1 complex to silence hundreds of
loci throughout the genome (Chu et al., 2011; Gupta et al., 2010; Rinn et al., 2007;
Tsai et al., 2010). In gene activation, the roX RNAs act in Drosophila dosage
compensation by binding the MSL complex to acetylate and activate the single X
chromosome in males (Deng and Meller, 2006). Recent studies have suggested that
enhancers and promoters often produce lincRNAs, and some of these lincRNAs
indeed have enhancer-like functions in the activation of nearby genes (Hung et al.,
2011; Kim et al., 2010; Orom et al., 2010; Rada-Iglesias et al., 2011). Furthermore, a
recent study of lincRNAs expressed in mouse embryonic stem cells has found that
many lincRNAs bind protein complexes involved in reading, writing, and erasing
chromatin marks, suggesting a comprehensive role of lincRNAs throughout epigenetic
64
regulation (Guttman et al., 2011).
HOTTIP is a prime example of an enhancer-like lincRNA (Wang, 2011).
HOTTIP is a 3.8 kilobase polyadenylated RNA transcribed from the 5’ distal tip of the
HOXA locus. HOTTIP coordinates expression of HOXA9 to HOXA13, which are
important for distal identity. HOTTIP RNA directly binds to and recruits the MLL
H3K4 methylase complex via the WDR5 protein to maintain H3K4 trimethylation
(H3K4me3). The MLL family of H3K4 methylases are evolutionarily conserved
(known as trithorax in Drosophila and COMPASS in yeast) (reviewed by Smith et al.,
2011), and MLL translocations are an important cause of human leukemias (Meyer et
al., 2009). Members of the MLL family encode single SET domain-containing H3K4
methylases, and are associated with WDR5, Ash2L, RbBP5, and additional proteins
for localization and enzymatic activity.
WDR5, which directly interacts with HOTTIP, is a multifunctional adaptor
protein that is present in several gene activating protein complexes. WDR5 can
discriminate posttranslational modifications on histone tails, as well as bind to MLL
and RbBP5 (Migliori et al., 2012; Wysocka et al., 2005). WDR5 is highly expressed in
embryonic stem cells (ESC), and is particularly required for ESC self renewal and
maintenance of active chromatin for pluripotency genes (Ang et al., 2011). WDR5 is
also targeted by the DDB1-CUL4B E3 ubiquitin ligase in the nucleus, and is thus
under control of the ubiquitin-proteasome system (UPS) (Nakagawa and Xiong, 2011).
WDR5 is a key substrate of CUL4B to regulate neuronal genes, and is believed to
underlie the role of CUL4B mutations in human X-linked mental retardation.
Regulated proteolysis is an essential strategy for cell fate control, ranging from
cell cycle progression to immunity. The UPS controls numerous aspects of gene
expression (reviewed by Muratani and Tansey, 2003), including transcription factor
half life and chromatin modifications. After a ubiquitin activating enzyme (E1)
adenylates ubiquitin and transfers it to a ubiquitin conjugating enzyme (E2), a
ubiquitin ligase (E3) transfers ubiquitin to the target protein. The proteasome then
recognizes, unfolds, and degrades the polyubiquitinated protein. The entire ubiquitinproteasome system is highly regulated, including at the level of ubiquitin ligase
65
substrate specificity and at the level of proteasome activity (Bader et al., 2011).
Nucleosomes are turned over in a dynamic and regulated fashion at specific genomic
loci (Deal et al., 2010; Dion et al., 2007; Mito et al., 2007) , and histone tails can be
proteolyzed during differentiation to erase prior histone modifications (Duncan et al.,
2008). These examples suggest that the half-lives of molecular interactions at
chromatin are a critical feature of epigenetic regulation, but the basis of temporal
regulation at chromatin is not well understood.
Here, we show that a lincRNA can regulate protein turnover. HOTTIP RNA
not only recruits WDR5 to specific genes, but also stabilizes WDR5 by preventing
proteasomal degradation. We define the mechanism of this novel mode of lincRNA
function, and demonstrate its importance for WDR5 function in chromatin regulation
and embryonic stem cell self renewal.
66
3.3 Results
HOTTIP lincRNA blocks WDR5 protein turnover
In the course of studying the HOTTIP-WDR5 interaction, we observed a
surprising effect of HOTTIP lincRNA on WDR5 stability. Co-expression of increasing
levels of HOTTIP lincRNA with constant levels of a FLAG-WDR5 expression
construct in 293T human embryonic kidney cells led to increased levels of FLAGWDR5 protein as judged by immunoblot, without a concordant effect on the FLAGWDR5 mRNA levels (Figure 3.1A). In contrast, co-expression with the PRC2associated lincRNA HOTAIR did not affect FLAG-WDR5 protein levels, and
expression of mRNA encoding either LacZ or eGFP (data not shown) caused a slight
decrease in FLAG-WDR5 protein. These results suggest that HOTTIP lincRNA
increases FLAG-WDR5 protein levels via a post-transcriptional mechanism.
Next, we performed pulse-chase analysis to determine the effects of HOTTIP
on WDR5 protein turnover. Addition of cycloheximide to stop protein translation after
transient transfection showed that FLAG-WDR5 is turned over with a half life of ~8 h
in control conditions, as previously described in HCT116 cells (Nakagawa and Xiong,
2011) (Figure 3.1B-D), and in concordance with our later experiments in embryonic
stem cells. WDR5 turnover is mediated by the proteasome as it is blocked by
proteasome inhibitor MG-132. Notably, co-expression of HOTTIP lincRNA
significantly extended the half life of FLAG-WDR5 to greater than 16 hours,
demonstrating that HOTTIP RNA prevents turnover of FLAG-WDR5. In contrast,
HOTAIR lincRNA or LacZ mRNA did not increase WDR5 protein half life.
Cycloheximide treatment did not affect transcription of HOTTIP lincRNA or FLAGWDR5 mRNA, as seen by qRT-PCR (Figure 3S1A). In contrast to ectopically
expressed WDR5 from transient transfection, endogenous WDR5 or stably transfected
WDR5 displayed a very long half life (>16 hours) in 293T cells, HeLa cells, and
human foreskin fibroblasts (data not shown), in concordance with previously
published data (Schwanhausser et al., 2011). In addition, endogenous WDR5 protein
levels were not affected by HOTTIP expression or exogenous WDR5 expression (data
not shown). Upon transfection in 293T cells, two separate populations of WDR5
67
appear to exist: (i) endogenous WDR5 already stabilized by endogenous HOTTIP-like
lincRNAs (more on this below); and (ii) ectopically expressed WDR5 protein that
require lincRNA stabilization.
We reasoned that the ability of HOTTIP to increase WDR5 protein half-life
may also prolong WDR5 chromatin occupancy, independently of the ability of
HOTTIP to recruit WDR5 to specific target genes. To test this idea, we utilized an
established reporter cell line, in which GAL4 binding sites driving a luciferase gene
have been stably integrated into 293T cells (Wysocka et al., 2005) (Figure 3.1E). Coexpression of HOTTIP lincRNA with WDR5 fused to GAL4 DNA binding domain
(GAL4-WDR5) substantially increased GAL4-WDR5 occupancy on reporter gene
chromatin that further amplified over time, as measured by chromatin
immunoprecipitation (ChIP) (Figure 3.1E). In contrast, vector- or HOTAIRexpressing cells showed less GAL4-WDR5 occupancy that quickly plateaued.
Analysis of luciferase activity confirmed that HOTTIP significantly increased reporter
gene output (Figure 3.1F). Because the GAL4 DNA binding domain also acts as a
nuclear localization signal (Chan et al., 1998), it is unlikely that the HOTTIP effect
occurs through altering the nuclear localization of GAL4-WDR5. These results
suggest that the ability of HOTTIP lincRNA to prolong WDR5 protein half life
extends to the chromatin-bound pool, which increases WDR5 chromatin occupancy at
a defined locus and boosts gene expression.
68
69
Figure 3.1. HOTTIP RNA overexpression stabilizes exogenous FLAG-WDR5. A.
Increased HOTTIP RNA expression caused increased levels of exogenous FLAGWDR5 at the protein level. Similar effects were not seen with control HOTAIR or
LacZ overexpression. Top: Representative western blot. Bottom: Average fold western
blot density quantitation normalized to FLAG-WDR5 RNA levels. B-D. HOTTIP
RNA overexpression extended FLAG-WDR5 half-life during cycloheximide treatment.
Similar effects were not seen with vector only, HOTAIR, or LacZ control conditions.
B. Representative western blots. Hours post cycloheximide indicated above. MG:
MG132 treated sample. C. Average WDR5 western blot density quantitation,
normalized to B actin. D. RT-PCR of samples at T=0. E-F. HOTTIP RNA expression
potentiated GAL4-WDR5-mediated activation of chromosomally integrated luciferase
by increasing protein occupancy on chromatin. E. Top: Schematic of chromosomally
integrated luciferase and ChIP qPCR primer locations. Bottom: HOTTIP RNA
increased GAL4-WDR5 occupancy on chromatin over time, as seen by GAL4 ChIPqPCR. Similar results were not seen with vector only or HOTAIR controls.F. HOTTIP
RNA expression increased GAL4-WDR5-mediated expression of luciferase. Similar
effects were not seen with control vector only or HOTAIR overexpression.
70
HOTTIP RNA stabilizes WDR5 at a step after poly-ubiquitination
We next defined the step that lincRNA impacts the ubiquitin-proteasome
pathway. A series of enzymes (E1, E2, E3) activate and conjugate ubiquitin to target
proteins, after which the poly-ubiquitin chains are then recognized by the proteasome
for protein degradation (Muratani and Tansey, 2003). We co-expressed epitope-tagged
ubiquitin, WDR5, and HOTTIP followed by high stringency immunoprecipitation to
map the effect of HOTTIP on WDR5 ubiquitination (Figure 3.2A). Surprisingly,
HOTTIP lincRNA quantitatively increased the levels of both un-ubiquitinated and
poly-ubiquitinated forms of WDR5 protein. The effect of HOTTIP lincRNA is similar
to prolonged incubation with MG132, which causes accumulation of substrate proteins
bearing poly-ubiquitin chains. We also used an in vitro ubiquitination assay with
purified E1, E2, and the E3 ligase CUL4B (which generates ubiquitin lysine 48
linkages), and found that HOTTIP does not interfere with CUL4B mediated polyubiquitination (Figure 3S1B). Both in vivo and in vitro studies show that HOTTIP
does not block WDR5 ubiquitination. Rather, the accumulation of ubiquitinated
WDR5 with HOTTIP expression suggests stabilization at a step after protein
ubiquitination.
Poly-ubiquitinated proteins targeted for degradation are unfolded and
proteolyzed by the proteasome. Thermodynamic stabilization of ubiquitinated proteins,
e.g. by strong ligand-protein interactions, can forestall proteasomal degradation,
presumably by increasing the energetic barrier for protein unfolding that is required to
thread the protein into the proteasome (Johnston et al., 1995). To test whether
HOTTIP binding thermodynamically stabilizes WDR5, we conducted differential
scanning fluorimetry (DSF) experiments (Niesen et al., 2007) (Figure 3.2B-C).
Purified WDR5 protein was incubated with HOTTIP RNA in the presence of the
hydrophobic dye SYPRO orange, and then slowly heated. Protein denaturation was
then detected by SYPRO orange binding to newly exposed core hydrophobic residues,
and the melting temperatures (Tm) were approximated by the first derivative maximum
of fluorescence. Whereas RNA alone did not cause any signal (data not shown),
WDR5 alone demonstrated a bimodal denaturation pattern (Figure 3.2B). Previous
71
analysis of DSF experiments has shown that the first peak likely defines the true
melting temperature, whereas the second peak represents protein aggregation post
denaturation (Kopec and Schneider, 2011). Incubation of HOTTIP increased the
WDR5 Tm by 5.1°C (p=0.002, Figure 3.2C). In contrast, addition of the nonbinding
lincRNA HOTAIR increased the Tm by only 1.65°C (p=0.058). These results
demonstrate that HOTTIP RNA binds WDR5 to physically stabilize the protein and
prevent unfolding. Indeed, thermodynamic stability may be a common mechanism by
which other nucleic acids can inhibit proteasomal processing of cognate protein
partners (Coppotelli et al., 2011; Sorokin et al., 2005).
72
Figure 3.2. HOTTIP RNA thermodynamically stabilizes WDR5 and does not affect
ubiquitination. A. HOTTIP RNA overexpression did not reduce in vivo ubiquition of
WDR5. Left: FLAG-WDR5 was immunoprecipitated, and ubiquitinated FLAGWDR5 was identified using anti-HA antibody. Right: FLAG-ubiquitin was
immunoprecipitated, and ubiquitinated V5-WDR5 was identified. B-C. Differential
scanning fluorimetry using purified GST-WDR5 and in vitro transcribed RNAs
revealed HOTTIP RNA thermodynamically stabilizes GST-WDR5, compared to no
RNA or HOTAIR controls. B. Representative first derivative fluorescence plots.
Arrows indicate melting temperatures, as seen as initial first derivative maximums. C.
Average melting temperature of GST-WDR5 when incubated with indicated RNAs.
73
A specific domain of HOTTIP confers WDR5 stabilization
To identify the region(s) of HOTTIP lincRNA that bind WDR5 protein, we
conducted in vivo native RNA immunoprecipitation (RIP) experiments (Figure 3.3A).
We found that while full length HOTTIP consistently bound purified WDR5 in vitro,
in vitro transcribed HOTTIP fragments behaved with substantial batch-to-batch
variation that was not seen in vivo, perhaps due to limitations of RNA folding in vitro
(data not shown). In RIP experiments, we co-expressed equal levels of successively
smaller HOTTIP RNA fragments with FLAG-WDR5 in 293T cells,
immunoprecipitated FLAG-WDR5, and quantified retrieval of RNAs using the same
qRT-PCR primer for either the 5’ or 3’ region (Figures 3S2, 3S3). To further prevent
biases, all RIP experiments were conducted in parallel with a positive control full
length HOTTIP RNA for normalization. Negative control RNAs (HOTAIR, U1), no
reverse transcription control, and control IgG immunoprecipitations revealed very
little or no enrichment (data not shown). We found two regions, Frag A (bases 1-1659)
and Frag D (1953-3760), that displayed strong binding affinity to WDR5 (Figure 3S2).
Because HOTTIP bases 1953-3760 demonstrated the strongest binding affinity to
WDR5, we used further deletion mapping to narrow the binding site to a 500 nt region
of HOTTIP, termed Frag D.3 (bases 1953-2453, Figure 3S3).
Because two distinct regions of HOTTIP bind WDR5, we also determined the
valency of protein:RNA interaction. Theoretically, each HOTTIP region may bind a
distinct WDR5 molecule, in which HOTTIP would act as an “RNA scaffold” to
multiplex proteins (Tsai et al., 2010). Alternatively, one HOTTIP molecule may bind a
single WDR5 molecule via multiple contacts. To distinguish between these
possibilities, we coexpressed FLAG- and V5- tagged WDR5 in 293T cells, and
conducted FLAG immunoprecipitation in native conditions (Figure 3S2I). HOTTIP
expression increased both V5-WDR5 and FLAG-WDR5 protein levels, demonstrating
that both forms of WDR5 were bound to and stabilized by HOTTIP RNA. Yet, FLAG
IP did not retrieve V5-WDR5 with or without HOTTIP addition. These results suggest
that each HOTTIP RNA binds to and stabilizes a single WDR5 protein by multiple
contacts.
74
Since full length HOTTIP can stabilize WDR5, we next examined whether
minimal WDR5 binding domains of HOTTIP RNA are sufficient to stabilize WDR5.
Cycloheximide chase experiments revealed that FLAG-WDR5 turnover was not
affected by any HOTTIP fragment tested, including either of the strong binding
regions (Figure 3.3B-C). Thus, strong binding by one HOTTIP fragment appears to
be insufficient to stabilize WDR5, suggesting additional RNA regions are required.
Similarly, Frag D was not sufficient to increase the Tm of WDR5 by DSF (data not
shown). One possible explanation is that a tandem combination of strong WDR5binding RNA sequences may be required for WDR5 stabilization. Alternatively, a
WDR5-binding RNA motif may need to be placed in the context of a long RNA in
order to function. To distinguish between these possibilities, we designed two fusion
RNAs (Figure 3.3A). The first was a duplication of the high affinity HOTTIP
fragment D, thus representing two tandem strong WDR5-binding regions. The second
fusion RNA was the minimal high affinity binding fragment D.3 fused to the 3’ end of
LacZ mRNA, thus representing a strong WDR5-binding region fused to a nonspecific
long RNA. Surprisingly, both chimeric RNAs were able to stabilize WDR5, albeit
with slightly less potency compared to full length HOTTIP (Figure 3.3B-C).
Moreover, the LacZ-FragD.3 fusion could also potentiate GAL4-WDR5-mediated
reporter gene expression, whereas LacZ or FragD.3 alone could not (Figure 3.3D).
Thus, lincRNA-mediated stabilization of WDR5 requires a specific WDR5-binding
RNA domain in the context of a nonspecific long RNA sequence.
75
Figure 3.3. Full length HOTTIP RNA stabilizes WDR5 through specific interactions.
A. Schematic of analyzed HOTTIP fragments and fusion RNAs. HOTTIP bases 3731659 (horizontal shading) and 1953-2453 (diagonal shading) demonstrated increased
binding affinity, as analyzed by cell-based native RNA immunoprecipitation. B-C.
76
HOTTIP fragment overexpression was insufficient to stabilize FLAG-WDR5 protein
with cycloheximide treatment. In contrast, WDR5-binding HOTTIP fragment fusion
RNAs was sufficient to stabilize WDR5. B. Representative western blots. Hours post
cycloheximide indicated above. MG: MG132 treated sample. C. Average western blot
density quantitation depicted by heat map, normalized to B actin. Full length HOTTIP
data same as Figure 3.1. D. HOTTIP region 1953-2453 (Frag D.3) fused to LacZ
moderately potentiated GAL4-WDR5-mediated expression of luciferase. Similar
effects were not seen with LacZ or HOTTIP bases 1953-2453 (Frag D.3)
overexpression. Full length HOTTIP data same as Figure 3.1.
77
Discovering the RNA binding interface on WDR5
To determine the HOTTIP RNA binding site, we next performed alanine
scanning mutagenesis of WDR5, guided by its crystal structure (Avdic et al., 2011;
Odho et al., 2010; Trievel and Shilatifard, 2009) (Figure 3.4A-B). WDR5 is a barrelshaped protein with several charged clefts on its surface, several of which are known
to mediate protein-protein interactions with MLL, histone H3 tail, or RbBP5 (Figure
3.4A, 3S4A). We generated 19 WDR5 point mutants, expressed them in E. coli as
GST-fusion proteins, and purified them to homogeneity (Figure 3S4B). Four out of
nineteen mutants significantly reduced the ability to retrieve HOTTIP lincRNA in
vitro: Y228A, L240A, K250A, and F266A. These WDR5 mutations defined a cleft
between blades 5 and 6, partially encompassing the same surface previously described
to bind RbBP5 amino acids 371-381 (Avdic et al., 2011; Odho et al., 2010). Thus, a
focal binding site defines the interaction between WDR5 and HOTTIP. To confirm
that HOTTIP binds an overlapping WDR5 protein surface as RbBP5, we preincubated wild type GST-WDR5 with an excess of RbBP5 peptide (amino acids 371381) or control H3K4me3 peptide (amino acids 1-20), and then assayed for HOTTIP
binding (Figure 3.4C). Whereas addition of H3K4me3 peptide had no effect, preincubation with RbBP5 peptide prevented HOTTIP binding to WDR5, thus
confirming the shared binding cleft.
To verify the HOTTIP binding site in living cells, we conducted RIP
experiments with select WDR5 mutants in 293T cells (Figure 3.4D). Immunoblot
analysis confirmed the expected interactions of wild type WDR5 with MLL complex
proteins (Figure 3.4D). As previously described, the D107A mutation affected MLL1
interactions (Song and Kingston, 2008), and the K250A mutation reduced RbBP5
binding (Avdic et al., 2011; Odho et al., 2010). Consistent with the direct in vitro
binding assay, both K250A and F266A mutations fully abrogated WDR5 binding to
HOTTIP in vivo. In contrast, mutations at D107A and R181A showed minimal effects
on WDR5-HOTTIP interactions. K250A and F266A did not affect the accumulation
of HOTTIP lincRNA (Figure 3S4C). Taken together, the in vitro and in vivo analyses
demonstrate that HOTTIP RNA binds WDR5 through a specific binding pocket.
78
Figure 3.4. HOTTIP binds and stabilizes WDR5 through the same binding site as
RbBP5. A. Crystal structure of WDR5 (PDB:3P4F, visualized with PyMol) revealed
that mutations abrogating HOTTIP RNA binding align along a cleft between blades 5
and 6, opposite the H3K4/MLL binding site. This binding surface is the same as that
for RbBP5. Top: Schematic of tested and HOTTIP binding mutations. Bottom:
Magnification of HOTTIP binding cleft. Yellow: MLL peptide. Blue: RbBP5 peptide.
B. qRT-PCR results of indicated GST-WDR5 mutants tested by in vitro assay for
binding to HOTTIP RNA or control Histone 1H2BG RNA.C. RbBP5 peptide
competition assay indicates that RbBP5 amino acids 371-381 can fully compete
HOTTIP binding to WDR5. Wild type data same as in Figure 3.4B. D. Average fold
pulldown of full length HOTTIP by cell-based native RNA immunoprecipitation of
select WDR5 mutants. Top: qRT-PCR results of RNA immunoprecipitation of FLAGWDR5 mutants. All values were normalized to input, then to FLAG pulldown of wild
type and positive control D107A. Negative RNA controls (U1, HOTAIR) and reaction
without reverse transcription (-RT) showed minimal enrichment (data not shown).
Bottom: Representative western blots of immunoprecipitations.
79
A selective WDR5 mutant reveals the importance of lincRNA binding on WDR5
half life and gene activation
Because HOTTIP binding blocks WDR5 turnover, we tested whether HOTTIP
can stabilize select WDR5 mutants with cycloheximide chase experiments (Figure
3.5A-C). Whereas without HOTTIP expression all WDR5 mutants displayed short
half lives, WDR5 D107A and WDR5 R181A, two mutants that do bind HOTTIP,
were protected from turnover by HOTTIP co-expression. In contrast, WDR5 K250A
and F266A, both of which cannot bind HOTTIP, could no longer be stabilized by
HOTTIP co-expression. Thus, binding to HOTTIP lincRNA is apparently required for
WDR5 stabilization.
To pinpoint the functional consequences of a selective lincRNA-binding
mutation of WDR5, we further analyzed WDR5 F266A. In contrast to the other
HOTTIP binding mutations, WDR5 F266A is solely defective in HOTTIP binding in
vitro and in vivo, but without any defects in binding MLL complex subunits RbBP5 or
MLL1 (Figure 3.4D). Isothermal calorimetry confirmed that F266A minimally
affected the affinity of WDR5 for RbBP5 peptide (WDR5 WT KD = 1.25 + 0.06 µM,
WDR5 F266A Kd= 5.13 + 0.49 µM). We reasoned that the F266A mutation offered
an experimental strategy to test the requirement of lincRNA binding for WDR5
function. Preliminary RIP-seq and CLIP-seq experiments have revealed multiple
RNAs that bind WDR5 (R. Flynn and Y.W.Y., unpublished observation); it would be
challenging to knock down all of these RNAs at once to determine the impact on
WDR5 function. However, the F266A mutation selectively abrogates the ability of
WDR5 to interact with HOTTIP (and possibly other lincRNAs), allowing examination
of the effects of lincRNA binding to WDR5.
To test the effects of the F266A mutation on gene activation, we transfected
increasing amounts of plasmids encoding GAL4-WDR5 WT or GAL4-WDR5 F266A
to activate luciferase reporter gene expression (Figure 3.5D). Even in the absence of
exogenously expressed lincRNA, WDR5 F266A showed a severe defect in luciferase
gene activation, requiring >100 fold transfected plasmid compared with wild type to
achieve similar reporter gene activity. We infer that some endogenous HOTTIP-like
80
activating lincRNAs may bind to and stabilize GAL4-WDR5, and the lack of RNA
binding severely reduces WDR5 function. Moreover, HOTTIP could no longer
potentiate activation by GAL4-WDR5 F266A (Figure 3.5E). Thus, even though the
WDR5 F266A mutation can bind the full MLL complex, the inability to bind
lincRNAs strongly compromises WDR5 function in activating gene expression.
81
82
Figure 3.5. HOTTIP binding mutations prevent HOTTIP mediated stabilization of
WDR5, as well as show decreased ability to activate target genes in 293T cells. A-C.
HOTTIP binding mutations abrogated HOTTIP-mediated stabilization of WDR5
during cycloheximide treatment. A-B. Representative western blots without (A) and
with (B) HOTTIP co-expression. Hours post cycloheximide indicated above. MG:
MG132 treated sample. C. Average western blot density quantitation visualized as
heat map, normalized to B actin. Wild type WDR5 data same as Figure 3.1. D-E.
WDR5 F266A is unable to activate luciferase expression. D. GAL4-WDR5 F266 was
defective in activating luciferase expression, as seen in luciferase titration tests.E.
HOTTIP could not potentiate GAL4-WDR5 F266A activation of luciferase gene
expression. Wild type WDR5 data same as Figure 3.1.
83
lincRNA binding to WDR5 is essential for H3K4 methylation and embryonic
stem cell self renewal
WDR5 has been recently shown to be required to maintain H3K4
trimethylation (H3K4me3) in mouse embryonic stem cell (ESC) genes for
pluripotency and self renewal (Ang et al., 2011). To test whether endogenous WDR5
requires lincRNA-mediated stabilization, we created “rescue” ESC lines that replace
endogenous mouse WDR5 with either wild type human WDR5 or F266A mutant
WDR5 (Figure 3.6A). ESCs were infected by lentiviruses containing two tandem
gene expression cassettes (Ang et al., 2011). The first cassette constitutively expresses
a highly efficient shRNA to repress endogenous mouse WDR5. WDR5 deficiency is
rescued by the second cassette, in which human WDR5 wild type (WT) or human
WDR5 F266A linked to GFP is expressed under control of a doxycycline (dox)inducible promoter. In the presence of dox, the ability of WDR5 mutant to support
ESC self renewal can be compared; upon dox withdrawal, the half-life of the mutant
protein and its regulatory impact are further revealed.
GFP+ cells were sorted and cultured for 4 days in the presence (+dox) and
absence (-dox) of doxycycline, and then analyzed by western blot for human WDR5
and global H3K4me3 levels (Figure 3.6B). Even though both WDR5 WT and WDR5
F266A ESCs were sorted using the same fluorescence parameters, slightly more
WDR5 WT protein was present when compared with WDR5 F266A, suggesting that
the WDR5 WT protein is more stable. Indeed, WDR5 mRNA levels between the WT
and F266A forms were similar (data not shown). In WDR5 WT ESCs, promoter
shutoff of WDR5 resulted in ~50% reduction of H3K4me3 (Lanes 1 and 2), as
previously reported (Ang et al., 2011). Interestingly, ESCs harboring WDR5 F266A
displayed a global 50% reduction of H3K4me3 (Lane 3), although WDR5 F266A
protein was expressed. In addition, promoter shutoff of WDR5 F266A caused a drastic
>80% reduction of H3K4me3 levels (Lanes 3 and 4). Thus, the F266A mutation
resembles a WDR5 loss of function, and exhibits a striking inability to maintain global
H3K4me3 levels over time.
We reasoned that because WDR5 F266A presumably cannot bind endogenous
84
HOTTIP-like lincRNAs in ESCs, WDR5 F266A is more unstable, thus causing both
reduced H3K4me3 levels and faster H3K4me3 decay. To test this hypothesis, we
conducted pulse chase experiments using doxycycline withdrawal to shut off FLAGhWDR5 transcription, and followed the fate of pre-existing wild type or F266A WDR5
protein in the nucleus and cytoplasm (Figure 3.6C). Different cullin ring ligases can
ubiquitinate WDR5, depending on cell type and whether the WDR5 protein is
cytoplasmic (CUL4A) or nuclear (CUL4B) (Nakagawa and Xiong, 2011). Indeed, we
found that in the cytoplasm, both WDR5 WT and F266A have equally short half lives
of less than 12 hours. In contrast, nuclear WDR5 WT persists for up to 24 hours, but
WDR5 F266A is turned over at least 6 hrs sooner (Figure 3.6C). The selective
instability of WDR5 F266A in the nucleus is apparently sufficient to deplete F266A
from the nuclear and chromatin fractions compared to WT (Figure 3.6D, E). The
alteration in cellular distribution of F266A can be explained by differential stability in
nuclear vs. cytoplasmic pools, or alternatively, by a role of F266 in directly controlling
nuclear localization. We believe that the former is the correct explanation based on
direct pulse chase data (Figure 3.6C) and also on the fact that fusion of F266A to a
strong nuclear localization signal (GAL4 DNA binding domain) did not rescue its
function (Figure 3.5D).
Loss of lincRNA binding also impacts the ability of WDR5 to promote ESC
self renewal. (Figure 3.6F-G). In WDR5 WT ESCs, promoter shut-off caused a ~50%
reduction in the number of alkaline phosphatase-positive colonies, a specific indicator
of ESC state. Notably, F266A WDR5 ESCs grown in the presence of doxycycline
displayed a similar ~50% reduction in colony number. By morphology, both WDR5
WT ESCs –dox and WDR5 F266A ESCs +dox formed similar small, partially
differentiated colonies. Doxycycline withdrawal in WDR5 F266A ESCs caused a
>90% reduction in number of colonies, with the majority of cells forming clusters of
fully differentiated alkaline phosphatase-negative cells. WDR5 directly binds to the
chromatin of key genes required for in ESC self renewal (Ang et al., 2011). ESCs
expressing WDR5 F266A have significantly reduced mRNA levels of pluripotency
regulators Oct4, Nanog, Sox2, and Esrrb (Figure 3.6H). Genome-wide expression
85
profiling further confirmed that WDR5 F266A was unable to sustain the gene
expression program associated with ESC self renewal and pluripotency, and instead
allowed ectopic expression of genes indicative of ectodermal and mesodermal lineages
(Figure 3.7A). Collectively, these data suggest that the ability to bind endogenous
lincRNAs are likely essential for WDR5 in maintaining H3K4me3, ESC gene
expression, and ESC self renewal.
86
Figure 3.6. WDR5 F266A mutation decreases protein stability and causes defects in
mouse embryonic stem cell self renewal. A. Schematic of lentiviral vectors, modified
from (Ang et al., 2011). B. Western blot 4 days after doxycycline removal. C. WDR5
F266A mutation decreased protein stability in the nucleus after doxycycline
withdrawal. Both WDR5 WT and WDR5 F266A were similarly unstable in the
cytoplasmic fraction. D. WDR5 F266A was defective in nuclear accumulation,
compared with WDR5 WT. E. WDR5 F266A reduced chromatin association, as seen
in chromatin isolation experiments. F-G. Alkaline phosphatase staining and
morphology of embryonic stem cell colonies after 6 days growth in conditions
specified. WDR5 F266A demonstrated reduced alkaline phosphatase positive colonies
87
with increased differentiation compared with wild type. F. Representative wells and
colony morphology. Scale bar represents 200 µm. G. Quantitation of colonies per well.
H. Loss of WDR5 or WDR5 F266A cause reduction of ESC pluripotency marker
expression by qRT-PCR.
88
Figure 3.7. WDR5 F266A causes loss of self renewal genes and increased expression
of differentiation genes. A. Microarray data reveal that both WDR5 F266A +dox and
WDR5 WT -dox displayed loss of self renewal genes. WDR5 F266A +dox further
demonstrated increased expression of ectodermal and mesodermal markers. B. Model
for HOTTIP-mediated switch of protein turnover to activate target genes. Expressed
lincRNAs bind the WDR5 protein, causing thermodynamic stability to allow protein
chromatinization, MLL complex assembly, and methylase activity for target gene
activation. Without HOTTIP or HOTTIP-like lincRNA binding, WDR5 cannot
associate effectively with chromatin and is rapidly degraded.
89
3.4 Discussion
RNA as protein timers
By examining the interaction between the HOTTIP lincRNA and the WDR5
protein, we have identified a novel mechanism by which a lincRNA regulates the
turnover of its cognate protein partner (Figure 3.7B). The ability of HOTTIP to
stabilize WDR5 is notable for several reasons. First, the interaction between HOTTIP
and WDR5 is strikingly specific. Cells are full of diverse species of RNAs; some, such
as ribosomal RNAs and snoRNAs, are very abundant. But the vast majority of cellular
RNAs are not capable of stabilizing WDR5, and HOTTIP co-expression is necessary
to stabilize ectopic WDR5. Second, the magnitude of protein stabilization is
significant. It is intriguing to consider that HOTTIP more than doubled the half life of
WDR5 protein such that WDR5 is not detectably degraded after 16 hours, which starts
to approach the time scale of an entire cell division cycle (especially in ESCs). These
results suggest that in addition to their roles in recruiting protein complexes to specific
genomic loci, some lincRNAs, like HOTTIP, can regulate the longevity of their
interacting protein complexes. As we showed, such an interaction can increase the
level and lifespan of chromatin occupancy, which is likely to have important impacts
on the transmission of chromatin states across cell generations, a critical aspect of
epigenetic regulation (Rando and Chang, 2009). Hundreds of lincRNAs have been
recently reported to bind to multiple readers, writers, and erasers of chromatin marks
(Guttman et al., 2009; Khalil et al., 2009; Zhao et al., 2010); one or more of these
lincRNAs may also control their partner proteins’ temporal activity by regulating
protein turnover, possibly with different gradations of half lives. More generally, as
there is no intrinsic barrier to this mechanism operating on proteins in other cellular
compartments, many long RNAs may modulate the half lives of RNA binding proteins
(RBPs), a large class of regulators that outnumbers transcription factors (Cook et al.,
2011). Coupled with the well-known capacity of microRNAs to inhibit protein
synthesis, this mechanism of lincRNA-mediated protein stabilization points to the
broad scope of noncoding RNAs in shaping the proteome.
Our experiments uncovered several unexpected features in the pathway of
90
lincRNA-mediated protein stabilization. We found that direct RNA-protein interaction
in the context of a long RNA is important for WDR5 stabilization. The requirement
for a long RNA context is correlated with the ability of RNA to increase WDR5
folding energy; while interaction with full length HOTTIP increased the Tm of WDR5
by DSF, the smaller WDR5 binding domain of HOTTIP did not. The requirement of a
specific RNA domain within a long RNA context is consistent with recent work
demonstrating the importance of short conserved RNA sequences in the conserved
functions of apparently divergent lincRNAs (Ulitsky et al., 2011). As we were able to
confer protein stabilizing activity to LacZ mRNA by simply grafting on a HOTTIP
domain, it is likely that other RNAs, including mRNAs, may similarly regulate the
stability and half-life of their cognate RNA-binding proteins.
In addition, we found that HOTTIP does not block WDR5 poly-ubiquitination,
but appears to block protein turnover at a subsequent step. Thus, the consequence of
lincRNA-mediated protein stabilization is the accumulation of poly-ubiquitinated
RNA binding protein that resists degradation. Our findings may be applicable to
understanding development and disease processes beyond epigenetic regulation. The
accumulation of ubiquitinated protein inclusions is a hallmark of many
neurodegenerative diseases, several of which are known to involve RNA binding
proteins (Lagier-Tourenne et al., 2010; Polymenidou et al., 2011; Wang et al., 2008).
While prior studies have focused on the impact of altered RBPs on RNA metabolism,
our study suggests that specific RNAs may play important initiating roles by altering
protein turnover. Protein interaction with nucleic acids, either DNA or RNA, has been
documented to block proteolysis(Coppotelli et al., 2011; Sorokin et al., 2005).
Proteasome activity is also regulated by accessory factors (Bader et al., 2011). The
precise mechanism by which long RNA interactions inhibits proteolysis should be
addressed in future studies.
Requirement of a lincRNA interface in WDR5 for gene activation and ESC self
renewal
Guided by the crystal structure and comprehensive mutagenesis, we identified
91
a lincRNA binding cleft in WDR5. WDR5 is a multifunctional adaptor protein that
can interact with multiple subunits of the MLL complex, such as MLL itself and
RbBP5, as well as with regulators and substrates of the complex, such as HOTTIP and
histone H3 arginine 2 symmetric demethylation (H3R2me2s) (Migliori et al., 2012).
The MLL and H3R2me2s share the same binding pocket, and we found that HOTTIP
and RbBP5 share a distinct binding pocket. While Oct4 has been reported to bind
WDR5, we found no difference in the lincRNA binding mutant for its interaction with
Oct4 (data not shown). Nonetheless, these multitudes of interactions suggest that
molecular arrangements in this complex may be dynamic and intricate. Our coimmunoprecipitation experiments with concomitant HOTTIP overexpression
demonstrate that RNA binding is compatible with the assembly of the intact MLLWDR5-Ash2L-RbBP5 complex.
By comprehensive structure-guided mutagenesis, we identified surface
residues of WDR5 that are specifically required for HOTTIP binding but do not
impact interactions with other protein partners, such as MLL1 or RbBP5. WDR5
protein mutant for these specific residues can no longer be stabilized by HOTTIP, and
are unable to activate transcription, suggesting that direct lincRNA-protein
interactions are important for WDR5 stabilization and function. Replacement of
endogenous WDR5 expression in mouse ESCs with an RNA binding mutant, WDR5
F266A, revealed that this interface is critical for the longevity of nuclear WDR5, as
well as accumulation in both the nucleus and on chromatin. The loss of stabilized
WDR5 leads to an inability to maintain global H3K4me3 level, loss of expression of
pluripotency regulators, and loss of ESC self renewal. Interestingly, the impact of
F266A mutation on H3K4me3 level is modest in steady state conditions, but becomes
quite striking with transcriptional shutoff of WDR5. We interpret these results to
indicate that WDR5 stability is important for the robustness of the active chromatin
state, exemplified by H3K4me3. In the absence of stabilized WDR5, H3K4me3
becomes much more sensitive to fluctuations in the transcriptional levels of WDR5
and likely other subunits of the MLL complex. Thus, these results reveal a new facet
in the connection between protein half lives and the temporal transmission of
92
epigenetic information via chromatin.
As HOTTIP is not expressed in ESCs, our results with WDR5 F266A imply
the existence of a family of ESC lincRNAs that stabilize WDR5. The recent discovery
of numerous lincRNAs important for ESC self renewal and pluripotency provide
plausible candidate for this concept (Chakraborty et al., 2012; Guttman et al., 2011). It
is also possible that one or more protein-based regulators may interact with WDR5 in
a F266 dependent fashion. Studies of other chromatin modification complexes have
shown that one complex (e.g. PRC2) can interact with hundreds of lincRNAs, each of
which regulates a subset of genes targeted by the complex (Khalil et al., 2009; Zhao et
al., 2010). A similar model likely applies to MLL and other gene activating complexes
as well. The MLL complex possesses multiple components with RNA-binding
capabilities. While WDR5 binds HOTTIP and potentially other lincRNAs (Wang et al.,
2011), MLL1 itself also has domains that bind RNAs in specific and nonspecific
fashions (Bertani et al., 2011; Krajewski et al., 2005). Cyp33, an allosteric regulator of
MLL, also contains a RNA recognition motif, and its regulation of MLL can be
controlled in an RNA-dependent manner (Hom et al., 2010; Wang et al., 2010). These
multiple RNA interaction domains may bind to different portions of lincRNAs either
specifically or non-specifically, allowing a long RNA to “shrink wrap” the entire
complex and stabilize it (Figure 3.7B). The presence of multiple RNA interaction
domains may complicate efforts toward systematic identification of specific RNA
species that regulate MLL activity (Krajewski et al., 2005). Our identification of
specific mutations that abrogate RNA-mediated WDR5 stabilization may facilitate
future efforts to discover additional examples of this new class of lincRNAs.
93
3.5 Materials and Methods
Cell lines, antibodies, and vectors:
HEK293T/17 cells were obtained from the American Type Culture Collection
(ATCC). The HEK293T 5XUAS-Luciferase cell line has been previously described
(Wysocka et al., 2005). V.6.5 mouse embryonic stem cells were grown on 1%
gelatinized plates in Knockout DMEM supplemented with 15% FBS, 1% Glutamax
(Invitrogen 35050), 1% nonessential amino acids, 1% penstrep, 0.2%
betamercaptoethanol, LIF (1:10000, Millipore ESG 1107), and 2 µg/mL doxycycline
(Sigma D9891). To maintain consistent levels of doxycycline, media was changed
every two days. Antibodies used were: anti-FLAGM2 (Sigma F1804), anti-β actin
(Abcam ab8227), anti-V5 (Sigma V8137), anti-HA (Cell Signaling 6E2 #2367), antiRbBP5 (Bethyl A300-109A), anti-MLL1-N (Bethyl A300-086A), anti-GAL4 (Santa
Cruz sc-577), anti-H3K4me3 (Abcam ab8580), anti-β tubulin (Abcam ab6046), antiH3 (Abcam ab1791). Excepting chromatin immunoprecipitation experiments, all other
immunoprecipitations were conducted using anti-FLAGM2 (Sigma A2220) or mouse
IgG (Sigma A0919) agarose beads, or anti-FLAGM2 magnetic beads (Sigma M8823).
Unless noted, all expression vectors were cloned into pcDNA3 or
pcDNA3.1+backbones.
qRT-PCR primers:
HOTTIP: (F:CCTAAAGCCACGCTTCTTTG, R:TGCAGGCTGGAGATCCTACT
5’ HOTTIP (F:TCTTTCCAGGGAAACAGTGG,
R:AACAGTGTGGACAGGGAAGG)
3’ HOTTIP (F:TATGGGTTGGGAGAGGGAGT,
R:AGCACCTGTAGTTGCCCATT)
Mid HOTTIP (F:CAAACTCCGTCCTCCAAAAC,
R:CAGTGAAGAGCGATCAGTGG)
HOTAIR (F:GGTAGAAAAAGCAACCACGAAGC,
R:ACATAAACCTCTGTCTGTGAGTGCC)
LacZ: (F:GTCGTTTGCCGTCTGAATTT, R:CCGCCACATATCCTGATCTT)
94
FLAG-WDR5: (F:GACTACAAAGACGATGACGACAA,
R:TCCCAGCTTGTGACCAGATA)
GAPDH (F:AGGTGGAGGAGTGGGTGTCGCTGTT,
R:CCGGGAAACTGTGGCGTGATGG)
U1 (F:ATACTTACCTGGCAGGGGAG, R:CAGGGGGAAAGCGCGAACGCA)
mβactin (F:GCTGTATTCCCCTCCATCGTG,
R:CACGGTTGGCCTTAGGGTTCAG)
mEsrrb (F:GGGTAGAGCCCACTTGTTCA, R:AGGTAGCCTGGGTTTTTGCT)
mNanog (F:TGGTCCCCACAGTTTGCCTAGTTC,
R:CAGGTCTTCAGAGGAAGGGCGA)
mOct4 (F:GTGGAGGAAGCCGACAACAATGA,
R:CAAGCTGATTGGCGATGTGAG)
mSox2 (F:CAGGAGAACCCCAAGATGCACAA,
R:AATCCGGGTGCTCCTTCATGTG)
ChIP qPCR primers:
GAPDH (F:AGGTGGAGGAGTGGGTGTCGCTGTT,
R:CCGGGAAACTGTGGCGTGATGG)
Luc1 (F: ACCATAGTCCCGCCCCTA, R: AACAGTACCGGAATGCCAAG)
Luc2 (F: CCGGTACTGTTGGTAAAATGG, R: AACCAGGGCGTATCTCTTCA)
RNA mediated effects on WDR5 protein levels
pcDNA3.1+FLAG-WDR5 and increasing amounts of indicated RNA
expression vectors were transfected into 293T cells using Lipofectamine 2000 per
manufacturer's instructions. pcdna3.1+ was added to normalize the total ng of
transfected plasmid. 96 hours after transfection, cells were harvested into two aliquots.
One aliquot was lysed in RIPA buffer (50 mM Tris-HCl pH 7.4, 1% NP40, 0.25% Nadeoxycholate, 450 mM NaCl, 1 mM EDTA, supplemented with cOmplete protease
inhibitor (Roche)), and analyzed by western blot for protein expression. FLAG-WDR5
protein levels were normalized to β-actin protein levels using ImageJ software (NIH).
95
RNA was extracted from the other aliquot using TRIzol (Invitrogen) and Rneasy
columns (Qiagen), treated with Turbo DNAfree (Ambion), and then analyzed by qRTPCR (Stratagene, Brilliant II SYBR Green). FLAG-WDR5 RNA levels were
normalized to GAPDH RNA. Normalized protein levels were then divided by
normalized RNA levels, and adjusted to 0 ng of RNA expression vector transfected.
WDR5 cycloheximide chase
pcdna3.1+FLAG-WDR5 and RNA expression vectors were transfected into
293T cells using Lipofectamine 2000. 24 hours after transfection, media was replaced
with media containing 4 µg/mL cycloheximide (Sigma). To one sample, MG132 was
also added to a final concentration of 10 µM (EMD Chemicals). Cells were harvested
at indicated time points, lysed with RIPA buffer, and analyzed by western blot.
Relative protein levels were visualized using Treeview software.
GAL4-WDR5 Chromatin immunoprecipitation
pCMX-GAL4-WDR5 and indicated RNA expression vectors were transfected
into 10 cm plates of HEK293T 5XUAS-Luciferase cells using Lipofectamine 2000. 48
and 96 hours after transfection, cells were fixed in 1% formaldehyde (Thermo Pierce).
Subsequent pellets were lysed, chromatin was extracted and immunoprecipitated using
1 µg antiGAL4 antibody, as previously described (Squazzo et al., 2006). Resulting
ChIP DNA was analyzed by qPCR (Roche Sybr Green).
Luciferase assay
pCMX-GAL4-WDR5, pRL Renilla Luicferase plasmid and indicated RNA
expression vectors were transfected into HEK293T 5XUAS-Luciferase cells using
Lipofectamine 2000. 40-48 hours after transfection, cells were harvested and analyzed
using the Dual-Luciferase Reporter Assay System (Promega). Luciferase readings
were normalized to Renilla readings, and then normalized to no GAL4-WDR5
background control.
96
Ubiquitination assays
For in vivo ubiquitination assays, indicated plasmids were cotransfected into
293T cells plated on 10 cm plates using Lipofectamine 2000. After 48h, cells were
treated with final concentration 10 µM MG132, harvested after 16h, and snap frozen
in liquid nitrogen. Cells were then lysed and immunoprecipitated using extremely high
stringency protocols, as previously described (Yong et al., 2010). Cell pellets were
resuspended in cold Empigen buffer (20 mM Tris pH 7.5, 500 mM NaCl, 2.5 mM
MgCl2, 1% Empigen, 2 mg/mL Heparin), supplemented with RnaseOUT (1:100), 100
mM Nethylmaleimide (1:50 EMD Chemicals), and cOmplete proteinase inhibitor
(1:50), and lysed with sonication. Lysates were cleared with centrifugation, then
incubated with anti-FLAG M2 magnetic beads for 4 hours. Beads were washed 5
times with cold Empigen buffer, once with room temperature RSB100 buffer (20 mM
pH 7.5, 100 mM NaCl, 2.5 mM MgCl2, 0.01% Triton X-100), washed twice more
with empigen buffer, then once with room temperature reverse crosslink buffer (20
mM Tris pH 8.0, 500 mM NaCl, 1 mM EDTA, 10 mM βmercaptoethanol). After
elution using FLAG peptide in reverse crosslink buffer, eluates were analyzed by
western blot.
In vitro ubiquitination assays were conducted as previously described
(Nakagawa and Xiong, 2011) with indicated RNAs.
Differential Scanning Fluorimetry (DSF)
DSF experiments were conducted as previously described, with modifications
(Niesen et al., 2007). GST-tagged WDR5 protein was expressed and purified from
Escherichia coli as previously described (Smith and Johnson, 1988). 300 ng GSTWDR5 was incubated with T7 RNA polymerase in vitro transcribed RNA (Promega, 0,
600 ng, or 1800 ng RNA) in signal buffer (10 mM HEPES pH 7.6, 150 mM NaCl,
SYPRO Orange 1:1000 (Invitrogen)), then incubated with a ramped temperature
(1C/min) in a Lightcycler 480 instrument (Roche, set to Multicolor hybprobe, Red
610). Data was analyzed by calculating the first derivative between each set of points,
graphed, and maximum peaks were noted. In case of multiple peaks, only the first
97
peak was recorded. P values were calculated using unpaired two tail Student’s T test.
Native RNA Immunoprecipitation
Native cell-based RNA immunoprecipitation was performed as previously
described with modifications (Dignam et al., 1983). 293T cells were transfected with
pcDNA3.1+FLAG-WDR5 and RNA expression plasmid. After 48-72 hours, cells
were harvested by scraping in cold PBS, spun down, and pellets were then snap frozen
in liquid nitrogen and stored at -80C. Cells were resuspended in Buffer A (10 mM
HEPES pH 7.5, 1.5 mM MgCl2, 10 mM KCl, 0.5 mM DTT, 1.0 mM PMSF,
supplemented with RnaseOUT (Invitrogen, 1:100)), and cell membranes were lysed
with 0.25% NP40. After centrifugation, resulting nuclei were further lysed in Buffer C
(20 mM HEPES pH7.5, 10% glycerol, 0.42M KCl, 4 mM MgCl2, 0.5 mM DTT, 1.0
mM PMSF supplemented with RnaseOUT (1:100)). Resulting whole cell lysates were
immunoprecipitatied with mouse anti-FLAG M2 or control mouse IgG agarose beads,
and washed four times with wash buffer (50 mM TrisCl pH 7.9, 10% glycerol,
100 mM KCl, 5 mM MgCl2, 10 mM β-mercaptoethanol, 0.1% NP40). To confirm
immunoprecipitation, FLAG peptide-eluted protein-RNA complexes were analyzed by
western blot. Coimmunoprecipitated RNA was extracted using TriZOL LS
(Invitrogen) and Qiagen RNeasy columns, treated with TURBO DNAfree, and then
analyzed by qRT-PCR (Brilliant II Sybr Green). Resulting values were defined as
fraction of input, then normalized to the positive control sample.
CoIP pulldown experiments were conducted as described above, with
modifications. pcDNA3.1+FLAG-WDR5, pcDNA3.1+V5-WDR5,
pcDNA3.1+HOTTIP and pcDNA3.1+ plasmids were transfected into 293T cells as
indicated. Cells were lysed and immunoprecipitated as described above, and resulting
pulldowns were analyzed by western blot.
98
In vitro GST pulldown and competition
GST pulldown and competition assays were performed as previously described,
with modifications (Wang, 2011). GST-tagged WDR5 wild type, mutants, and GST
control proteins were expressed and purified from Escherichia coli as previously
described (Smith and Johnson, 1988). GST proteins were bound to glutathione
sepharose 4B (Amersham/GE Healthcare), blocked with 0.2 mg/mL BSA (Ambion) in
2X binding buffer (40 mM Hepes pH 7.6, 200 mM KCl) at 4C for 1 hour. For peptide
competition assays, indicated peptides (Elim Biopharmaceuticals) were also added.
0.04 mg/mL heparin (1:1, Sigma H3149) was added to further block nonspecific RNA
binding. Protein-bound beads were then incubated with folded T7-transcribed
HOTTIP or histone 1H2BG mRNA for 45 minutes (Promega T7). Beads were washed
twice with PB200 (20 mM Hepes pH 7.6, 200 mM KCl, 0.05% NP40) and once with
DEPC-treated water to remove detergents. After resuspension in water, beads were
directly used in qRT-PCR reactions to determine RNA levels. To confirm lack of
degradation, bead-bound proteins were analyzed by Silver Stain Plus (Bio-Rad).
Establishment of FLAG-tagged WDR5 Rescue Embyronic Stem Cells
Lentiviral constructs were generated that that constitutively knock down mouse
WDR5 while allowing expression of a doxycycline-inducible human WDR5. The
rescue vector pLKO.tre (Ang et al., 2011) was modified by replacing the shRNAimmune mouse WDR5-cDNA with either wildtype or F266A mutated human WDR5.
The resulting pLKO vectors and rtTA (from Marius Wernig, Stanford University)
constructs were cotransfected with packaging plasmids (pLKO: second generation
pCMV-dR8.2 dvpr and pCMV-VSVG| rTta: third generation pMDLg/pRRE, pCMVVSVG, pRSV-Rev) into 293T cells using Lipofectamine 2000 (Invitrogen). After 16
hours, media was changed. Viruses were harvested after 48 and 72 hours, pooled, and
then clarified by centrifugation. Viruses were concentrated using Lenti-X concentrator,
incubated with 8 µg/mL polybrene (Sigma H9268) and then used to infect V6.5 ESCs.
GFP+ cells were sorted (Stanford Shared FACS Facility) and then cultured as
described above. For RNA analysis, 5000 GFP+ cells were plated in media with or
99
without doxycycline, and then harvested after 6 days. For qRT-PCR analysis, resulting
values were normalized to control β actin levels, and then normalized to WT+dox.
Embryonic stem cell fractionation
ESCs were fractionated as previously described (Tee et al., 2010), with
modifications. To isolate nuclei and cytoplasm, Pellets were lysed in nuclei isolation
buffer (NIB: 10 mM Tris pH7.5, 60 mM KCl, 15 mM NaCl, 1.5 mM MgCl2, 1 mM
CaCl2, 250 mM Sucrose, 10% glycerol, and 0.1% NP40. Lysates were centrifuged at
low speed to separate cytoplasm (supernatant) from nuclei (pellet). Cells were washed
once with NIB without NP40, resuspended in NIB without NP40, and treated with
Turbo Dnase (Ambion). EDTA was added to a final concentration of 10 mM to
neutralize the Dnase. After centrifugation, the resulting washed pellet was
resuspended in NIB, and NaCl was added to a final concentration of 500 mM to lyse
the nuclei.
For chromatin isolation, cells were lysed with cytoskeleton buffer (CSK: 10 mM
PIPES KOH pH7, 100 mM NaCl, 300 mM sucrose, 3 mM MgCl2, 0.5% Triton X100), centrifuged, and the resulting pellet was washed twice more with CSK buffer.
The washed pellet was resuspended in CSK buffer, and treated with Turbo Dnase to
release chromatin-bound proteins.
Alkaline phosphatase staining
For colony forming assays, FLAG-tagged WDR5 rescue V6.5 ESCs were
counted by flow cytometer, as well as hemocytometer, and 250 cells were plated onto
12 well plates with full media with or without doxycycline. After 6 days, cells were
fixed with 1:1 methanol:acetone, then stained using Vector Blue (Vector Laboratories)
per the manufacturer’s instructions.
100
Microarray analysis
cDNA was synthesized, labeled, and hybridized to Affymetrix Mouse 430 2.0
arrays in biologic duplicates (Stanford Protein and Nucleic Acid Facility). Arrays were
normalized by robust multi-array average (RMA), and probes that had an expression
value ≥50 in at least 1 sample were filtered out. Gene expression with multiple probes
were averaged. Heatmaps show mean-centered gene expression, based on previously
described expression patterns (Ang, 2011). Microarray data are available at Gene
Expression Omnibus
(http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?token=xtydzkwwmqymitm&acc=GS
E36513).
101
3.6 Supplementary Figures
Figure 3S1. A. Cycloheximide treatment does not affect RNA levels. qRT-PCR of
representative experiment of HOTTIP and Flag-WDR5 during cycloheximide
treatment, normalized to GAPDH. B. HOTTIP overexpression does not affect in vitro
ubiquitination of purified WDR5.
102
Figure 3S2. Two regions of HOTTIP bind WDR5. Successively smaller 5’ (A-C) and
3’ (E-F) fragments of HOTTIP RNA were analyzed by cell-based native RNA
immunoprecipitation. A,D. Schematic of fragments. *: estimated region of 5’ and 3’
qPCR primers. B,E. Average fold pulldown of RNA fragments in Flag-WDR5 or
representative IgG control immunoprecipitations by qRT-PCR, relative to input levels.
103
All values were normalized to Flag pulldown of full length HOTTIP.C,F. Average
relative levels of input HOTTIP fragment RNA by qRT-PCR, relative to GAPDH. G.
Representative western blots of immunoprecipitations. H. In vitro transcription of
schematized HOTTIP fragments. I. Immunoprecipitation of Flag-WDR5 does not
pulldown V5-WDR5 with HOTTIP overexpression, even though HOTTIP stabilizes
both proteins.
104
Figure 3S3. HOTTIP bases 1953-2453 (Frag D.3) demonstrate equivalent binding
affinity to bases 1953-3760 (Frag D). Successively smaller 5’ (A-C) and 3’ (E-F)
fragments of HOTTIP RNA bases 1953-3760 (Frag D) were analyzed by cell-based
native RNA immunoprecipitation. A,D. Schematic of fragments. Full length HOTTIP
is provided as reference. *: estimated region of mid and 3’ qPCR primers. B,E.
Average fold pulldown of RNA fragments in Flag-WDR5 or representative IgG
control immunoprecipitations by qRT-PCR, relative to input levels. All values were
normalized to Flag pulldown of HOTTIP fragment 1953-3760. C,F. Average relative
levels of input HOTTIP fragment RNA by qRT-PCR, relative to GAPDH. G.
Representative western blots of immunoprecipitations. H. In vitro transcription of
selected schematized HOTTIP fragments not shown on Figure 3S2H.
105
Figure 3S4. A. Electrostatic surface contouring of the WDR5 crystal structure
(PDB:3P4F, visualized with PyMol) reveal that mutations abrogating HOTTIP RNA
binding define an electropositively charged region. Orange: WDR5 residues important
for HOTTIP binding. Cyan: MLL peptide. Grey: RbBP5 peptide B. Silver stain of
post-assay glutathione-bound GST-WDR5 mutants indicates no protein degradation or
loading bias occurred during in vitro binding assay. C. Average levels of input
HOTTIP RNA for cell-based native RNA immunoprecipitation assays of WDR5
mutants by qRT-PCR, relative to GAPDH.
106
3.7 References
Ang, Y.S., Tsai, S.Y., Lee, D.F., Monk, J., Su, J., Ratnakumar, K., Ding, J., Ge, Y.,
Darr, H., Chang, B., et al. (2011). Wdr5 mediates self-renewal and
reprogramming via the embryonic stem cell core transcriptional network. Cell
145, 183-197.
Avdic, V., Zhang, P., Lanouette, S., Groulx, A., Tremblay, V., Brunzelle, J., and
Couture, J.F. (2011). Structural and biochemical insights into MLL1 core
complex assembly. Structure 19, 101-108.
Bader, M., Benjamin, S., Wapinski, O.L., Smith, D.M., Goldberg, A.L., and Steller, H.
(2011). A conserved F box regulatory complex controls proteasome activity in
Drosophila. Cell 145, 371-382.
Bertani, S., Sauer, S., Bolotin, E., and Sauer, F. (2011). The noncoding RNA Mistral
activates Hoxa6 and Hoxa7 expression and stem cell differentiation by
recruiting MLL1 to chromatin. Mol Cell 43, 1040-1046.
Candeias, M.M., Malbert-Colas, L., Powell, D.J., Daskalogianni, C., Maslon, M.M.,
Naski, N., Bourougaa, K., Calvo, F., and Fahraeus, R. (2008). P53 mRNA
controls p53 activity by managing Mdm2 functions. Nat Cell Biol 10, 10981105.
Chakraborty, D., Kappei, D., Theis, M., Nitzsche, A., Ding, L., Paszkowski-Rogacz,
M., Surendranath, V., Berger, N., Schulz, H., Saar, K., et al. (2012).
Combined RNAi and localization for functionally dissecting long noncoding
RNAs. Nat Methods.
Chan, C.K., Hubner, S., Hu, W., and Jans, D.A. (1998). Mutual exclusivity of DNA
binding and nuclear localization signal recognition by the yeast transcription
factor GAL4: implications for nonviral DNA delivery. Gene Ther 5, 12041212.
Chu, C., Qu, K., Zhong, F.L., Artandi, S.E., and Chang, H.Y. (2011). Genomic maps
of long noncoding RNA occupancy reveal principles of RNA-chromatin
interactions. Mol Cell 44, 667-678.
Cook, K.B., Kazan, H., Zuberi, K., Morris, Q., and Hughes, T.R. (2011). RBPDB: a
database of RNA-binding specificities. Nucleic Acids Res 39, D301-308.
Coppotelli, G., Mughal, N., Marescotti, D., and Masucci, M.G. (2011). High avidity
binding to DNA protects ubiquitylated substrates from proteasomal
degradation. J Biol Chem 286, 19565-19575.
Deal, R.B., Henikoff, J.G., and Henikoff, S. (2010). Genome-wide kinetics of
nucleosome turnover determined by metabolic labeling of histones. Science
328, 1161-1164.
Deng, X., and Meller, V.H. (2006). Non-coding RNA in fly dosage compensation.
Trends Biochem Sci 31, 526-532.
Dignam, J.D., Lebovitz, R.M., and Roeder, R.G. (1983). Accurate transcription
initiation by RNA polymerase II in a soluble extract from isolated mammalian
nuclei. Nucleic Acids Res 11, 1475-1489.
Dion, M.F., Kaplan, T., Kim, M., Buratowski, S., Friedman, N., and Rando, O.J.
(2007). Dynamics of replication-independent histone turnover in budding
yeast. Science 315, 1405-1408.
107
Duncan, E.M., Muratore-Schroeder, T.L., Cook, R.G., Garcia, B.A., Shabanowitz, J.,
Hunt, D.F., and Allis, C.D. (2008). Cathepsin L proteolytically processes
histone H3 during mouse embryonic stem cell differentiation. Cell 135, 284294.
Gupta, R.A., Shah, N., Wang, K.C., Kim, J., Horlings, H.M., Wong, D.J., Tsai, M.C.,
Hung, T., Argani, P., Rinn, J.L., et al. (2010). Long non-coding RNA
HOTAIR reprograms chromatin state to promote cancer metastasis. Nature
464, 1071-1076.
Guttman, M., Amit, I., Garber, M., French, C., Lin, M.F., Feldser, D., Huarte, M., Zuk,
O., Carey, B.W., Cassady, J.P., et al. (2009). Chromatin signature reveals over
a thousand highly conserved large non-coding RNAs in mammals. Nature 458,
223-227.
Guttman, M., Donaghey, J., Carey, B.W., Garber, M., Grenier, J.K., Munson, G.,
Young, G., Lucas, A.B., Ach, R., Bruhn, L., et al. (2011). lincRNAs act in the
circuitry controlling pluripotency and differentiation. Nature 477, 295-300.
Guttman, M., Garber, M., Levin, J.Z., Donaghey, J., Robinson, J., Adiconis, X., Fan,
L., Koziol, M.J., Gnirke, A., Nusbaum, C., et al. (2010). Ab initio
reconstruction of cell type-specific transcriptomes in mouse reveals the
conserved multi-exonic structure of lincRNAs. Nat Biotechnol 28, 503-510.
Hom, R.A., Chang, P.Y., Roy, S., Musselman, C.A., Glass, K.C., Selezneva, A.I.,
Gozani, O., Ismagilov, R.F., Cleary, M.L., and Kutateladze, T.G. (2010).
Molecular mechanism of MLL PHD3 and RNA recognition by the Cyp33
RRM domain. J Mol Biol 400, 145-154.
Hung, T., Wang, Y., Lin, M.F., Koegel, A.K., Kotake, Y., Grant, G.D., Horlings,
H.M., Shah, N., Umbricht, C., Wang, P., et al. (2011). Extensive and
coordinated transcription of noncoding RNAs within cell-cycle promoters. Nat
Genet 43, 621-629.
Johnston, J.A., Johnson, E.S., Waller, P.R., and Varshavsky, A. (1995). Methotrexate
inhibits proteolysis of dihydrofolate reductase by the N-end rule pathway. J
Biol Chem 270, 8172-8178.
Khalil, A.M., Guttman, M., Huarte, M., Garber, M., Raj, A., Rivea Morales, D.,
Thomas, K., Presser, A., Bernstein, B.E., van Oudenaarden, A., et al. (2009).
Many human large intergenic noncoding RNAs associate with chromatinmodifying complexes and affect gene expression. Proc Natl Acad Sci U S A
106, 11667-11672.
Kim, T.K., Hemberg, M., Gray, J.M., Costa, A.M., Bear, D.M., Wu, J., Harmin, D.A.,
Laptewicz, M., Barbara-Haley, K., Kuersten, S., et al. (2010). Widespread
transcription at neuronal activity-regulated enhancers. Nature 465, 182-187.
Kopec, J., and Schneider, G. (2011). Comparison of fluorescence and light scattering
based methods to assess formation and stability of protein-protein complexes.
J Struct Biol 175, 216-223.
Krajewski, W.A., Nakamura, T., Mazo, A., and Canaani, E. (2005). A motif within
SET-domain proteins binds single-stranded nucleic acids and transcribed and
supercoiled DNAs and can interfere with assembly of nucleosomes. Mol Cell
Biol 25, 1891-1899.
108
Lagier-Tourenne, C., Polymenidou, M., and Cleveland, D.W. (2010). TDP-43 and
FUS/TLS: emerging roles in RNA processing and neurodegeneration. Hum
Mol Genet 19, R46-64.
Meyer, C., Kowarz, E., Hofmann, J., Renneville, A., Zuna, J., Trka, J., Ben Abdelali,
R., Macintyre, E., De Braekeleer, E., De Braekeleer, M., et al. (2009). New
insights to the MLL recombinome of acute leukemias. Leukemia 23, 14901499.
Migliori, V., Muller, J., Phalke, S., Low, D., Bezzi, M., Mok, W.C., Sahu, S.K.,
Gunaratne, J., Capasso, P., Bassi, C., et al. (2012). Symmetric dimethylation
of H3R2 is a newly identified histone mark that supports euchromatin
maintenance. Nat Struct Mol Biol 19, 136-144.
Mito, Y., Henikoff, J.G., and Henikoff, S. (2007). Histone replacement marks the
boundaries of cis-regulatory domains. Science 315, 1408-1411.
Morey, C., and Avner, P. (2011). The demoiselle of X-inactivation: 50 years old and
as trendy and mesmerising as ever. PLoS Genet 7, e1002212.
Muratani, M., and Tansey, W.P. (2003). How the ubiquitin-proteasome system
controls transcription. Nat Rev Mol Cell Biol 4, 192-201.
Nakagawa, T., and Xiong, Y. (2011). X-linked mental retardation gene CUL4B targets
ubiquitylation of H3K4 methyltransferase component WDR5 and regulates
neuronal gene expression. Mol Cell 43, 381-391.
Niesen, F.H., Berglund, H., and Vedadi, M. (2007). The use of differential scanning
fluorimetry to detect ligand interactions that promote protein stability. Nat
Protoc 2, 2212-2221.
Odho, Z., Southall, S.M., and Wilson, J.R. (2010). Characterization of a novel WDR5binding site that recruits RbBP5 through a conserved motif to enhance
methylation of histone H3 lysine 4 by mixed lineage leukemia protein-1. J
Biol Chem 285, 32967-32976.
Orom, U.A., Derrien, T., Beringer, M., Gumireddy, K., Gardini, A., Bussotti, G., Lai,
F., Zytnicki, M., Notredame, C., Huang, Q., et al. (2010). Long noncoding
RNAs with enhancer-like function in human cells. Cell 143, 46-58.
Polymenidou, M., Lagier-Tourenne, C., Hutt, K.R., Huelga, S.C., Moran, J., Liang,
T.Y., Ling, S.C., Sun, E., Wancewicz, E., Mazur, C., et al. (2011). Long premRNA depletion and RNA missplicing contribute to neuronal vulnerability
from loss of TDP-43. Nat Neurosci 14, 459-468.
Rada-Iglesias, A., Bajpai, R., Swigut, T., Brugmann, S.A., Flynn, R.A., and Wysocka,
J. (2011). A unique chromatin signature uncovers early developmental
enhancers in humans. Nature 470, 279-283.
Rando, O.J., and Chang, H.Y. (2009). Genome-wide views of chromatin structure.
Annu Rev Biochem 78, 245-271.
Rinn, J.L., Kertesz, M., Wang, J.K., Squazzo, S.L., Xu, X., Brugmann, S.A.,
Goodnough, L.H., Helms, J.A., Farnham, P.J., Segal, E., et al. (2007).
Functional demarcation of active and silent chromatin domains in human
HOX loci by noncoding RNAs. Cell 129, 1311-1323.
Schwanhausser, B., Busse, D., Li, N., Dittmar, G., Schuchhardt, J., Wolf, J., Chen, W.,
and Selbach, M. (2011). Global quantification of mammalian gene expression
109
control. Nature 473, 337-342.
Smith, D.B., and Johnson, K.S. (1988). Single-step purification of polypeptides
expressed in Escherichia coli as fusions with glutathione S-transferase. Gene
67, 31-40.
Smith, E., Lin, C., and Shilatifard, A. (2011). The super elongation complex (SEC)
and MLL in development and disease. Genes Dev 25, 661-672.
Song, J.J., and Kingston, R.E. (2008). WDR5 interacts with mixed lineage leukemia
(MLL) protein via the histone H3-binding pocket. J Biol Chem 283, 3525835264.
Sorokin, A.V., Selyutina, A.A., Skabkin, M.A., Guryanov, S.G., Nazimov, I.V.,
Richard, C., Th'ng, J., Yau, J., Sorensen, P.H., Ovchinnikov, L.P., et al. (2005).
Proteasome-mediated cleavage of the Y-box-binding protein 1 is linked to
DNA-damage stress response. EMBO J 24, 3602-3612.
Squazzo, S.L., O'Geen, H., Komashko, V.M., Krig, S.R., Jin, V.X., Jang, S.W.,
Margueron, R., Reinberg, D., Green, R., and Farnham, P.J. (2006). Suz12
binds to silenced regions of the genome in a cell-type-specific manner.
Genome Res 16, 890-900.
Tee, W.W., Pardo, M., Theunissen, T.W., Yu, L., Choudhary, J.S., Hajkova, P., and
Surani, M.A. (2010). Prmt5 is essential for early mouse development and acts
in the cytoplasm to maintain ES cell pluripotency. Genes Dev 24, 2772-2777.
Trievel, R.C., and Shilatifard, A. (2009). WDR5, a complexed protein. Nat Struct Mol
Biol 16, 678-680.
Tsai, M.C., Manor, O., Wan, Y., Mosammaparast, N., Wang, J.K., Lan, F., Shi, Y.,
Segal, E., and Chang, H.Y. (2010). Long noncoding RNA as modular scaffold
of histone modification complexes. Science 329, 689-693.
Ulitsky, I., Shkumatava, A., Jan, C.H., Sive, H., and Bartel, D.P. (2011). Conserved
function of lincRNAs in vertebrate embryonic development despite rapid
sequence evolution. Cell 147, 1537-1550.
Wang, K.C., and Chang, H.Y. (2011). Molecular mechanisms of long noncoding
RNAs. Mol Cell 43, 904-914.
Wang, K.C., Yang, Y.W., Liu, B., Sanyal, A., Corces-Zimmerman, R., Chen, Y.,
Lajoie, B.R., Protacio, A., Flynn, R.A., Gupta, R.A., et al. (2011). A long
noncoding RNA maintains active chromatin to coordinate homeotic gene
expression. Nature 472, 120-124.
Wang, X., Arai, S., Song, X., Reichart, D., Du, K., Pascual, G., Tempst, P., Rosenfeld,
M.G., Glass, C.K., and Kurokawa, R. (2008). Induced ncRNAs allosterically
modify RNA-binding proteins in cis to inhibit transcription. Nature 454, 126130.
Wang, Z., Song, J., Milne, T.A., Wang, G.G., Li, H., Allis, C.D., and Patel, D.J.
(2010). Pro isomerization in MLL1 PHD3-bromo cassette connects H3K4me
readout to CyP33 and HDAC-mediated repression. Cell 141, 1183-1194.
Wysocka, J., Swigut, T., Milne, T.A., Dou, Y., Zhang, X., Burlingame, A.L., Roeder,
R.G., Brivanlou, A.H., and Allis, C.D. (2005). WDR5 associates with histone
H3 methylated at K4 and is essential for H3 K4 methylation and vertebrate
development. Cell 121, 859-872.
110
Yong, J., Kasim, M., Bachorik, J.L., Wan, L., and Dreyfuss, G. (2010). Gemin5
delivers snRNA precursors to the SMN complex for snRNP biogenesis. Mol
Cell 38, 551-562.
Zhao, J., Ohsumi, T.K., Kung, J.T., Ogawa, Y., Grau, D.J., Sarma, K., Song, J.J.,
Kingston, R.E., Borowsky, M., and Lee, J.T. (2010). Genome-wide
identification of polycomb-associated RNAs by RIP-seq. Mol Cell 40, 939953.
Zofall, M., Yamanaka, S., Reyes-Turcu, F.E., Zhang, K., Rubin, C., and Grewal, S.I.
(2012). RNA elimination machinery targeting meiotic mRNAs promotes
facultative heterochromatin formation. Science 335, 96-100.
111
CHAPTER 4
Concluding remarks
112
4.1 Conclusion
The examination of the noncoding “dark matter” of the genome has revealed
multiple surprises, most notably, the presence of long noncoding RNA transcripts.
Even more surprising, whereas the small noncoding miRNA class simply silences
target mRNAs, lncRNAs do not share a single isolated mechanism. Most lncRNAs
function through protein interactions, although a few notable exceptions primarily
modulate DNA, RNA, or miRNA directly (Rinn and Chang, 2012). lncRNAs can
recruit proteins in cis or in trans, modulate protein catalytic activity, or act as decoys
to prevent proper protein interaction with their targets (Wang and Chang, 2011). In
this manner, lncRNAs have been shown to regulate the spatial and/or functional
properties of their corresponding binding proteins.
Because many lncRNAs have been found to be important in epigenetic
silencing (Rinn and Chang, 2012), we wanted to characterize a lncRNA involved in
the opposing epigenetic activation. To this end, we cloned a novel lncRNA, HOTTIP,
which is transcribed from the distal end of the HoxA locus. HOTTIP is important in
maintaining H3K4 trimethylation and gene expression of the distal HoxA genes – a 40
kilobase region that forms a highly interacting DNA cluster in three dimensional space,
as seen by high throughput chromosome conformation capture. Overexpression of
HOTTIP does not induce upregulation of distal HoxA gene expression in fibroblasts
that do not normally express HOTTIP; thus, we believe that HOTTIP acts in cis,
similar to the previously well characterized model mechanism of Xist lncRNA in X
chromosome inactivation (Morey and Avner, 2011). HOTTIP binds specifically to
WDR5, a component of the MLL methyltransferase complex. Thus, in our initial
model, HOTTIP RNA, once transcribed, recruits WDR5 and the MLL complex to the
nearby distal HoxA locus, thus creating a localized domain of H3K4 trimethylation to
maintain gene expression for positional identity.
Since then, our understanding of the HOTTIP-WDR5 function and interaction
has greatly increased, revealing that HOTTIP and potential HOTTIP-like lncRNAs
may play a larger role than previously appreciated. Using an overexpression system in
cells not expressing HOTTIP, exogenous HOTTIP is able to prevent exogenous
113
WDR5 degradation by cycloheximide chase. HOTTIP-mediated stabilization allows
increased WDR5 deposition at a specific locus. This HOTTIP-mediated stabilization
occurs after ubiquitination, likely by increasing the energy threshold required to
unfold WDR5 for proteasomal degradation.
To better characterize the physical interaction between HOTTIP and WDR5,
we conducted mutagenesis assays to examine both the RNA and protein binding sites.
First, by deletion mapping, we found that HOTTIP bases 1953-2453 bind most
strongly to WDR5, and that RNA-mediated stability requires that the HOTTIP bases
1953-2453 be fused to an arbitrary long RNA sequence. To find the corresponding
binding site on WDR5, we conducted a small scale alanine-scanning mutagenesis
screen, and isolated the RNA binding site to the WDR5 protein cleft between blades 5
and 6, encompassing the same binding surface region as the RbBP5 peptide 371-381.
Surprisingly, HOTTIP RNA and the RbBP5 peptide 371-381 compete for the same
binding pocket in purified component binding assays, further confirming this binding
site.
One mutation, WDR5 F266A, causes reduced binding to HOTTIP, but does
not have any affects on RbBP5 binding. This unique attribute allowed us to separate
the effects of lncRNA binding on WDR5 from the protein-mediated effects. As would
be expected, WDR5 F266A is unable to be stabilized by HOTTIP, and is ineffective in
activating luciferase gene expression.
To then test the functional consequences of inability to bind lncRNAs, we then
utilized mouse embryonic stem cells, which have been previously shown to be reliant
on WDR5 to maintain stem cell identity. By shRNA knockdown of endogenous
WDR5 with inducible expression of either WDR5 WT or WDR5 F266A, we could
ascertain the importance of lncRNA binding to WDR5 on cell function. Compared
with WDR5 WT, WDR5 F266A cells are unable to maintain H3K4me3, show more
rapid WDR5 turnover, and accumulate less in the nucleus and chromatin. Furthermore,
embryonic stem cells with WDR5 F266A are defective in self renewal, as seen by
colony morphology, alkaline phosphatase staining, and gene expression. Thus, binding
to lncRNAs via the HOTTIP binding pocket is crucial for WDR5 function in ES cells.
114
Altogether, this body of work reveals a novel function of an activating lncRNA.
HOTTIP, the 3.76 kilobase lncRNA transcribed from the distal tip of the HoxA locus,
binds WDR5, and also thermodynamically stabilizes the protein to prevent
proteasomal degradation. This lncRNA-mediated protein stabilization is crucial for
WDR5 and MLL methyltransferase function, revealing a novel ability for lncRNAs to
act as “molecular switches.”
115
4.2 Future Directions
However, more questions remain regarding both the mechanism and functional
role of HOTTIP and HOTTIP-like lncRNAs. First and foremost, the true identity of
the prominent HOTTIP RNA isoform still remains a mystery. The full length 3.76
kilobase HOTTIP cDNA was RACE cloned, and found to be both spliced and
polyadenylated. Yet, because HOTTIP is present at only ~0.3 copies per cell, Northern
analysis was unsuccessful. The exact sequence of the predominant, functional form of
HOTTIP may provide useful in understanding its endogenous role in distal fibroblasts
and maintaining distal identity. One possible method to address the difficulties in
working with low copy transcripts would be to enrich for the transcript using
complementary base pairing technology, as previously described (Mercer et al., 2012).
This may then allow sufficient enrichment to determine exact isoforms and abundance
by Northern blot.
As another problem posed by the copy number, if HOTTIP is present at only
~0.3 copies per cell, how can it find, interact, and stabilize ~20,000 copies of WDR5
per cell (Schwanhausser et al., 2011)? A low copy RNA both lowers the likelihood of
random interactions with a protein, and ability to thermodynamically stabilize multiple
copies of that protein. One possibility is that HOTTIP simply acts in a localized
domain or few domains (eg. primarily distal HoxA locus), and regulates WDR5
stability there, similar to the previously described Pc2-binding lncRNAs (Yang et al.,
2011). Thus, to maintain WDR5 across the genome, an entire class of HOTTIP-like
lncRNAs must interact with WDR5, similar to the recently described PRC2-binding
lncRNAs (Guil et al., 2012; Khalil et al., 2009; Zhao et al., 2010). Another possibility
is that HOTTIP stabilizes a small population of WDR5 (eg. nascent WDR5 protein).
Enzymatic recruitment may also be involved. In this manner, perhaps temporary
lncRNA-mediated stabilization is sufficient to allow other mechanisms of stabilization,
such as MLL complex assembly.
From our mutagenesis work characterizing the RNA-protein interaction, we
surprisingly found that HOTTIP RNA and RbBP5 protein compete for the same
binding surface on WDR5. Because binding to RbBP5 is required for full catalytic
116
activity, HOTTIP binding to WDR5 appears likely to be inhibitory. How can the low
copy HOTTIP (0.3 copies per cell), compete with RbBP5 (7,000 copies per cell
(Schwanhausser et al., 2011)) for binding with WDR5 in the first place? Because
RbBP5 is found in multiple complexes, the amount of free RbBP5 may be far less in
abundance. As another explanation, HOTTIP may act for temporal control, to regulate
WDR5 stability until RbBP5 can bind and replace it. Indeed, RNAs and proteins
appear to commonly compete for the same binding site, as can be seen in the example
of the Cyp33-MLL1 interaction (Hom et al., 2010). Further investigation of the effects
of protein-RNA competition on enzyme kinetics and localization may prove fruitful in
linking biochemical observations with cellular functions.
Ultimately, the HOTTIP RNA presents both a novel example for lncRNAmediated epigenetic activation, as well as for lncRNA-mediated protein stability. Yet,
cells that do not express HOTTIP RNA (eg. mouse ESCs, lung fibroblasts) still
maintain WDR5 expression and protein levels. Furthermore, in mouse ESCs, the
F266A mutation renders WDR5 unstable and ineffective to maintain the ESC self
renewal state, in a manner very similar to WDR5 knockdown. Thus, we conclude that
other endogenous HOTTIP-like RNAs likely exist in every cell type, comprising a
novel class of lncRNAs involved in WDR5 stabilization and gene activation.
Discovery and characterization of the various HOTTIP-like lncRNAs may prove
insightful in understanding human health and disease processes.
117
4.3 References
Guil, S., Soler, M., Portela, A., Carrere, J., Fonalleras, E., Gomez, A., Villanueva, A.,
and Esteller, M. (2012). Intronic RNAs mediate EZH2 regulation of epigenetic
targets. Nat Struct Mol Biol.
Hom, R.A., Chang, P.Y., Roy, S., Musselman, C.A., Glass, K.C., Selezneva, A.I.,
Gozani, O., Ismagilov, R.F., Cleary, M.L., and Kutateladze, T.G. (2010).
Molecular mechanism of MLL PHD3 and RNA recognition by the Cyp33
RRM domain. J Mol Biol 400, 145-154.
Khalil, A.M., Guttman, M., Huarte, M., Garber, M., Raj, A., Rivea Morales, D.,
Thomas, K., Presser, A., Bernstein, B.E., van Oudenaarden, A., et al. (2009).
Many human large intergenic noncoding RNAs associate with chromatinmodifying complexes and affect gene expression. Proc Natl Acad Sci U S A
106, 11667-11672.
Mercer, T.R., Gerhardt, D.J., Dinger, M.E., Crawford, J., Trapnell, C., Jeddeloh, J.A.,
Mattick, J.S., and Rinn, J.L. (2012). Targeted RNA sequencing reveals the
deep complexity of the human transcriptome. Nat Biotechnol 30, 99-104.
Morey, C., and Avner, P. (2011). The demoiselle of X-inactivation: 50 years old and
as trendy and mesmerising as ever. PLoS Genet 7, e1002212.
Rinn, J.L., and Chang, H.Y. (2012). Genome Regulation by Long Noncoding RNAs.
Annu Rev Biochem 81, 145-166.
Schwanhausser, B., Busse, D., Li, N., Dittmar, G., Schuchhardt, J., Wolf, J., Chen, W.,
and Selbach, M. (2011). Global quantification of mammalian gene expression
control. Nature 473, 337-342.
Wang, K.C., and Chang, H.Y. (2011). Molecular mechanisms of long noncoding
RNAs. Mol Cell 43, 904-914.
Yang, L., Lin, C., Liu, W., Zhang, J., Ohgi, K.A., Grinstein, J.D., Dorrestein, P.C., and
Rosenfeld, M.G. (2011). ncRNA- and Pc2 methylation-dependent gene
relocation between nuclear structures mediates gene activation programs. Cell
147, 773-788.
Zhao, J., Ohsumi, T.K., Kung, J.T., Ogawa, Y., Grau, D.J., Sarma, K., Song, J.J.,
Kingston, R.E., Borowsky, M., and Lee, J.T. (2010). Genome-wide
identification of polycomb-associated RNAs by RIP-seq. Mol Cell 40, 939-953.