BIOPHYSICAL INTERROGATION OF
HISTONE H3/H4 TRANSITIONS FROM HISTONE CHAPERONES TO DNA
by
WALLACE HAU LIU
B.S., University of California San Diego, 2007
A thesis submitted to the Faculty of the Graduate School of the University of Colorado in partial fulfillment
of the requirements for the degree of Doctor of Philosophy
Pharmacology Program 2014
ii
This thesis for the Doctor of Philosophy degree by Wallace Hau Liu
has been approved for the Pharmacology Program
by
K. Ulrich Bayer, Chair Jeffrey S. Kieft
Heide L. Ford Paul C. Megee Aaron M. Johnson Mair E. A. Churchill, Advisor
iii Liu, Wallace Hau (Ph.D., Pharmacology)
Biophysical Interrogation of Histone H3/H4 Transitions from Histone Chaperones to DNA
Thesis directed by Professor Mair E. A. Churchill
ABSTRACT
In eukaryotes, DNA and histones are packaged into nucleosomes, which protect the genome in the form of chromatin. Chromatin is a dynamic structure that is regulated by many proteins, including modifying complexes, remodellers, and histone chaperones. Histone chaperones are a functionally-related family, defined by the ability to bind histones and mediate nucleosome assembly. During DNA replication, the chaperones Asf1 (Anti-silencing function 1) and CAF-1 (Chromatin Assembly Factor 1) cooperate to deposit histones H3/H4 onto DNA. In this Thesis, I undercover the molecular
mechanisms of H3/H4 transfer, as it transitions from Asf1 to CAF-1 to DNA.
To understand the details behind these H3/H4-bound complexes, I employed in
vitro biophysical methods using the full-length proteins from budding yeast. A
combination of electrophoretic mobility shift assays, fluorescence spectroscopy, hydrogen/deuterium exchange, and chemical crosslinking experiments facilitated a thermodynamic analysis of Asf1-H3/H4, Asf1-CAF-1, and CAF-1-H3/H4 interactions, including the strength and stoichiometry of these associations. I discovered that the C-terminal tail of Asf1 directly enhances interactions with CAF-1 and H3/H4. Despite this, Asf1, H3/H4, and CAF-1 do not form one stable complex. Instead, H3/H4 forms a tight complex with CAF-1 exclusive of Asf1, in which a monomer of the tri-subunit CAF-1 chaperone binds two H3/H4 dimers with a low nanomolar dissociation constant.
iv
Importantly, the CAF-1 complex induces an arrangement of H3/H4 that resembles the nucleosomal (H3/H4)2 tetramer form, ultimately depositing (H3/H4)2 tetramers onto DNA. Thus, the transition of H3/H4 from the Asf1-associated dimer to the DNA-associated tetramer is promoted by CAF-1-induced H3/H4 oligomerization.
The architecture of the CAF-1-H3/H4 complex is mediated through extensive interactions with the Cac1 subunit of CAF-1. Cac1 bridges interactions between the other two CAF-1 subunits, Cac2 and Cac3. The last ~170 amino acids of the Cac1 polypeptide associates with Cac2, H3, and H4, which involves core helices and N-terminal tails in both histone proteins. Notably, the Cac1 and Cac2 subunits together are sufficient to induce the high affinity binding and stoichiometry unique to CAF-1-H3/H4. Together, these findings show how nascent and parental H3/H4 is transitioned through CAF-1, revealing a mechanism of (H3/H4)2 tetramer deposition that supports a conservative model of tetramer inheritance.
The form and content of this abstract are approved. I recommend its publication. Approved: Mair E. A. Churchill
v
ACKNOWLEDGMENTS
I would like to thank my thesis advisor, Mair Churchill, for her consummate support in all matters science, from advice to funding to cheerleading.
I also would not have succeeded in my thesis work, were it not for my previous mentors who influenced me in unique ways. They include Gourisankar Ghosh, who told me I wasn’t meant to be a scientist. I had the fortune to train under Stephen Perry and Sam Hoare, who both patiently taught me the basic principles of research techniques and applications. I had the further privilege of working with Lutz Tautz and Torkel Vang, an experience that refined my critical thinking, and showed me that true scientists are independent and creative.
I would also like to thank the many friends I have been lucky enough to meet through graduate school, both in and out of the Department of Pharmacology. They include, but are not limited to: Charis and Jesse, who allowed me to trail run with them every weekend; Chrissy and Hank, who dragged me to the climbing gym and let their dogs smother me; Ryan, for our many adventures togther; Prince, and his fresh outlook on life; Angelo and Nobu, for being terrific table tennis practice partners and teammates; Jess, Ollie, and Lyn, for our laughs together; Cat, for being genuine and dependable; Pirooz, for our good times watching and celebrating sports; Christina, for being the perfect roommate; Jen and Tyler, who find humor in everything; Suzy, who knows too many secrets about me; the twins, who are all too sweet; Johnny, who is the most trustworthy soul I’ve met; and Britney, with whom every moment is beautiful.
vi
Most importantly, I would like to acknowledge my family, for their many sacrifices and unconditional love. I would like to thank my sister, who supports me through any and all situations, as well as gung gung and poh poh, who raised me with unmatched care and will always be in my heart. Finally, no one means more to me than my parents, who set perfect examples with their hard work and tireless dedication to their family.
vii
TABLE OF CONTENTS
CHAPTER
I. INTRODUCTION ... 1
Chromatin Biology ... 1
Regulation of Chromatin Assembly ... 2
Histone Chaperones... 7
The Asf1-CAF-1 Axis in DNA Replication ... 15
Thesis Statement ... 23
II. MATERIALS AND METHODS ... 25
Preparation of Histones H3/H4 ... 25
Fluorophore Labeling of H4 ... 28
Preparation of Asf1 ... 29
Fluorophore Labeling of Asf1 ... 31
Preparation of CAF-1 Proteins ... 32
Electrophoretic Shift Mobility Assays ... 34
Immunoblotting and Far Western Blotting ... 36
Fluorescence Spectroscopy ... 38
Analytical Size-exclusion Chromatography... 41
Limited Proteolysis ... 41
Bioinformatic Analyses ... 42
Chemical Crosslinking ... 43
Protein Prospector Determination of Crosslinks ... 44
viii
III. BIOPHYSICAL ANALYSES OF ASF1 INTERACTIONS WITH H3/H4 AND
CAF-1 ... 47
Introduction ... 47
Results ... 56
Asf1 and CAF-1 bind with submicromolar affinity ... 56
The C-terminus of Asf1 contributes to H3/H4 binding ... 58
Asf1 dissociates from CAF-1–H3/H4 complexes ... 62
Discussion ... 64
IV. INTERROGATION OF CAF-1-H3/H4 STOICHIOMETRY AND AFFINITY ... 66
Introduction ... 66
Results ... 74
All CAF-1 subunits can individually interact with H3/H4 ... 74
CAF-1 binds to H3/H4 with a weaker affinity than Asf1-H3/H4 ... 81
CAF-1 remains a monomer with and without H3/H4 ... 83
CAF-1 oligomerizes H3/H4 and deposits (H3/H4)2 tetramers on DNA ... 86
The Cac1 subunit is sufficient for tetramer deposition on DNA ... 92
Discussion ... 94
V. ARCHITECTURE OF THE CAF-1 AND CAF-1-H3/H4 COMPLEXES ... 101
Introduction ... 101
Results ... 108
Formation of CAF-1 and CAF-1-H3/H4 induces major structural changes in all CAF-1 subunits ... 108
The C-terminal third of Cac1 directly crosslinks to Cac2, H3, and H4 ... 125
Discussion ... 127
ix
Perspective ... 135
Conclusions and Model of Histone Transfer from Asf1 to CAF-1 ... 136
Conclusions and Model of CAF-1-H3/H4 Architecture ... 137
Future Directions ... 140
1
CHAPTER I INTRODUCTION
Chromatin Biology
The early visual inspections of eukaryotic chromatin and chromosome segregation in the middle and late 19th century led to a chromosome-based model of inheritance (Sutton, 1902, 1903). This model offered that the chromosome structure is the physical carrier of genetic transmission from mother to daughter cells. Later studies confirmed that DNA is the hereditary information (Avery et al., 1944), and its packaging in the nucleus along with histone proteins forms the major components of chromosomes. The molecular architecture of chromatin, however, was predicted to be a continuous superhelix of DNA and histones, until microscopic studies revealed an array of “beads on a string” (Olins and Olins, 1974; Woodcock, 1973). These breakthrough observations led to extensive biochemical and biophysical characterization of the beaded chromatin subunit – termed the nucleosome – and the complex regulation of DNA and histone accessibility it entails.
As the fundamental repeating unit of chromatin, nucleosomes are nucleoprotein complexes consisting of ~150 DNA base pairs wrapped around an octamer of positively charged histone proteins. The landmark crystal structure of the entire nucleosome unit achieved high resolution insight into the histone-histone and histone-DNA interactions within (Luger et al., 1997) (Figure 1.1A). Histones have small, globular cores of three alpha helices arranged in an L-shape, each core bordered with disordered N- and C-terminal stretches. Within the nucleosome, four members of the histone family form
2
specific interactions with each other via the core domains: histones H3 and H4 co-fold head-to-tail into heterodimers, while the H2A and H2B histones associate likewise into H2A/H2B dimers. These associations comprise the basis of nucleosome quaternary structure, in which two H3/H4 dimers assemble into a stable (H3/H4)2 tetramer, stabilized by a “four helix bundle” arrangement that includes two helices from two H3 molecules. The tetramer is flanked on opposite sides by separate H2A/H2B dimers. Extensive interactions with DNA are available with this octamer structure: the central ~80 base pairs of nucleosomal DNA wraps the (H3/H4)2 tetramer into a unit known as the tetrasome. Binding of an H2A/H2B dimer to each face of the tetrasome wraps the additional ~70 bp of DNA to give rise to the complete nucleosome core particle. Such a stable arrangement not only physically shields the genome from undesirable exposures, but allows for complex regulation of DNA-histone assembly.
On a broader scale, multiple nucleosomes are observed to fold onto a 30 nm fiber (Everid et al., 1970), an architecture that is thought to be facilitated by incorporation of histone H1. These linker histones are excluded from the nucleosome core particle, but bind ~10 base pairs on each side of the nucleosome (Simpson, 1978). At these positions, H1 stabilizes higher ordered chromatin structures (Thoma et al., 1979), although these structures appear polymorphic and lack a unifying model (Wong et al., 2007).
Regulation of Chromatin Assembly
All cellular processes that use DNA as the template – including transcription, replication, and repair – must involve dynamic changes in local chromatin structure. Most of our knowledge of chromatin assembly stems from studies at the level of individual
3
nucleosomes; hence, the terms “chromatin assembly” and “nucleosome assembly” are largely interchangeable. At this fundamental level, regulation involves assembly and disassembly of DNA-histone species, as well as nucleosome density at specific sites of DNA reorganization.
The assembly of a nucleosome is thought to occur in a stepwise process, whereby an (H3/H4)2 tetramer is first assembled onto a DNA template, followed by deposition of the two H2A/H2B dimers (Figure 1.1B). This model is supported best by isolation of histones from cellular extract, from which H3/H4 was separated independently of H2A/H2B upon phosphocellulose binding. The H3/H4 fraction was competent to stimulate DNA replication on the simian virus 40 (SV40) origin, which required only H3/H4, T antigen, nucleoside triphosphates, and topoisomerases. The fraction was unable to promote nucleosome assembly, however, as assessed by plasmid supercoiling in the presence of the assembly factor CAF-1 (Chromatin Assembly Factor 1). Subsequent introduction of the H2A/H2B fraction was necessary to promote supercoiling.
Furthermore, micrococcal digestion experiments demonstrated that DNA replicated by the H3/H4 fraction are protected, but have irregularly spaced histone-DNA species. Addition of H2A/H2B established more distinct nucleosome spacing. Together, the observations supported H3/H4-DNA as the intermediate species preceding H2A/H2B incorporation during nucleosome assembly (Smith and Stillman, 1991).
Many factors have been discovered that influence nucleosome positioning and DNA and histone accessibility. These include patterns in DNA sequence, such as AT-rich tracts that discourage nucleosome presence (Yuan et al., 2005). DNA may also be
4
methylated at cytosines, which are notably found at CpG sequences proximal to
transcription start sites, thereby repressing long-term gene expression (Kass et al., 1997). On a larger scale, hypermethylated DNA is observed within heterochromatin, which are regions of silenced chromatin.
Figure 1.1 Structure and stepwise assembly of the S. cerevisiae nucleosome. A.
Crystal structure of the nucleosome core particle containing S. cerevisiae histones and 146 base pairs DNA (PDB accession code 1ID3). Histone H3 is colored blue, H4 green, H2A yellow, and H2B red. B. Model of the stepwise nucleosome assembly/disassembly process, in which a stable (H3/H4)2 associates with DNA, followed by inclusion of separate H2A/H2B dimers.
5
Additionally, higher eukaryotes possess histone variants that are found at distinct chromatin regions; for example, H2A.Z is located 5’ of transcription start sites and correlates with lower nucleosome density (Raisner et al., 2005). macroH2A, on the other hand, is found in areas of transcriptional silencing (Zhang et al., 2005). The H3.1 and H3.2 variants in mammals are found exclusively during DNA replication and repair, while H3.3 histones are deposited into nucleosomes independently of replication (Hake and Allis, 2006; Tagami et al., 2004). At centromeric chromatin, the H3 variant CENP-A is incorporated, marking the centromere for kinetochore formation (Smith, 2002).
Post-translational modifications on histones often correlate well with nucleosome accessibility and occupancy. Some of these modifications affect direct histone-histone or histone-DNA binding, such as acetylation of H4 on K16 (H4K16ac). This modification disrupts an electrostatic interaction between the N-terminal tail of H4 with an acidic region on H2A/H2B of a neighboring nucleosome (Shogren-Knaak et al., 2006). H4K16ac, accordingly, is found enhanced at regions of upregulated transcription (Akhtar and Becker, 2000). Other modifications may act as docking sites for protein-histone
interactions. For example, methylation of H3 on lysine 9 recruits HP1 (Heterochromatin Protein 1), which links multiple nucleosomes together to form silenced heterochromatin (James and Elgin, 1986). Conversely, levels of tri-methylation of H3 on lysine 4 are high at promoters and transcription start sites (Pokholok et al., 2005). This particular pattern of modifications recruits the chromatin remodeller NURF (NUcleosome Remodelling Factor) via an interaction with the PHD (Plant HomeoDomain) domain of NURF subunit BPTF (Bromodomain PHD finger Transcription Factor). As a chromatin remodelling
6
complex, NURF slides nucleosomes away from active genes using payoff from ATP (Adenosine TriPhosphate) hydrolysis (Wysocka et al., 2006).
The post-translational modifications on DNA and histones are potentially epigenetic; that is, they are thought to be inherited from mother to daughter cells, affecting gene expression without changes to the DNA sequence. The best evidence of epigenetics from DNA methylation come from studies of Arabidopsis thaliana, a plant with a small genome that is ~14% methylated. Specific regions of methylation are stable for at least eight generations, and differences in methylation between breeds at these loci account for differences in plant flowering time and root length. The epigenetic character of this observation is further supported by DNA sequencing, which show nearly identical DNA sequences between the methylated and unmethylated breeds (Cortijo et al., 2014).
Similar evidence for epigenetic behavior from histone modifications, however, is lacking. Nonetheless, these modifications do influence histone and DNA accessibility. The most common mechanism involves recruitment of chromatin-modifying or
remodelling complexes, as discussed above. The recruitment mechanism is epitomized by the aforementioned H3K9 methylation, which promotes HP1 binding through an
interaction with the HP1 chromodomain. HP1 is thought to spread silenced regions of chromatin by oligomerizing with itself, thereby joining and compacting nucleosomes expressing the methylation marks. in vitro binding studies support this interpretation, showing an enhanced binding affinity of HP1 for methylated H3K9. While HP1 does not bind unmodified H3 N-terminal peptides, it binds tri-methylated H3K9 with micromolar affinity (KD = 4 µM) (Fischle et al., 2003). Most measurements of chromatin-associated proteins with histone modifications exhibit this similar, weak affinity. Likely, the
7
interactions are stabilized by additional histone-binding subunits typically included in these complexes, while the modifications provide specificity that link the complexes with appropriate events.
The localization of the appropriate enzymes that modify chromatin may be mediated by lncRNAs (long non-coding RNAs). These RNAs position at specific DNA sites and recruit enzymes through secondary structure motifs (Mattick, 2012; Tsai et al., 2010). For instance, the lncRNA HOTAIR (HOX Transcript AntIsense RNA) targets the histone methyltransferase PRC2 (Polycomb Repressive Complex 2) to chromatin, and subsequent tri-methylation of H3K27 by PCR2 recruitment and activity leads to a silenced chromatin state (Tsai et al., 2010).
As expected with any ordered process, chromatin disassembly must be favored through energetic means. To achieve this, a family of chromatin remodellers feature ATPase subunits able to catalyze ATP hydrolysis. The resulting energy use promotes sliding of nucleosomes at active genes by remodellers such as NURF. Alternatively, chromatin remodellers may promote assembly of ordered nucleosome arrays, such as the activity of ACF (ATP-utilizing Chromatin assembly and remodelling Factor) during DNA replication (Haushalter and Kadonaga, 2003). The work of ATP-dependent
remodellers are intimately involved in all active chromatin-dependent processes, such as transcription, silencing, and replication.
Histone Chaperones
In support of the above factors, histone chaperones are universal players in nucleosome assembly and disassembly. The concept of molecular “chaperones” was first
8
used to describe histone chaperones, and later expanded to include any protein that can protect other proteins from unfolding or participating in incorrect interactions. The first observed chaperone was nucleoplasmin, which was isolated from Xenopus laevis nuclear extract (Laskey et al., 1978). Nucleoplasmin capably bound histones, favoring
nucleosome formation and preventing the histone-DNA aggregates that otherwise would have formed. Since, histone chaperones have been classified as proteins with high affinity binding for histones that can deposit their cargo on DNA, independently of ATP
hydrolysis. The exceptional binding enables these chaperones to prevent the highly basic histone proteins from off-pathway histone-DNA interactions (Das et al., 2010). The DNA deposition propensity provides nucleosome assembly capability to this family. Histone chaperones also promote productive histone interactions with other chaperones in the family, histone-modifying enzymes, and ATP-dependent nucleosome remodellers (Liu and Churchill, 2012). Thus, the spatiotemporal properties of histone trafficking in the cell – and ultimately nucleosome assembly – is directed by histone chaperones.
As a result of the many pathways available to histones, which lead to specific destinations during specific signaling events, histone chaperones comprise a diverse family of 15 proteins (Table 1.1). This family can be divided into chaperones that specifically bind H3/H4 in vivo and those that bind H2A/H2B. H2A/H2B chaperones include FACT (FAcilitates Chromatin Transcription), which assists in nucleosome disassembly and re-assembly during transcription and DNA replication
(Belotserkovskaya et al., 2003; Okuhara et al., 1999). Nap1 (Nucleosome assembly protein 1) participates in H2A/H2B removal ahead of transcription and shuttles newly translated H2A/H2B into the nucleus (Ito et al., 2000; Mosammaparast et al., 2002).
9
Table 1.1 Summary of histone chaperones, including their interactions, structures, and
functions.
Chaperone Histone
Partner Structure Information Stoichiometry
Other
Interactions Function
Nucleoplasmin H2A/H2B H2A/H2B storage in X. laevis
Nap1 H1,
H2A/H2B
Homodimer; N-terminal tail; 6 α helices
in "headphone" shape' C-terminal acidic tail
2 Nap1 :
2 H3/H4 ACF
Transcription elongation; DNA replication; nuclear import
FACT H2A/H2B
2 protein subunits: Spt16 N-terminal helical pita-fold, middle
domain PH domain; Pob3 PH domain
DNA Pol I Transcription elongation; DNA replication; DNA repair
ANP32E H2A/H2B
Leucine-rich repeats; C-terminal helix-loop that
binds H2A.Z
H2A/H2B exchange for H2A.Z/H2B
NASP H1, H3/H4 HAT1,
Asf1 H3/H4
Immunoglobulin-like β sheet core that binds H3/H4; C-terminal tail 1 Asf1 : 1 H3/H4 CAF-1, HIRA, MCM2-7, RFC, Rad53, H3/H4 storage; nuclear import; transcription elongation; DNA replication; DNA repair; cell senescence
CAF-1 H3/H4 3 protein subunits:
Cac3 β Propeller
Asf1,
PCNA, DNA replication; DNA repair
HIRA H3/H4 4 protein subunits Asf1 Transcription silencing and
activation Rtt106 H3/H4 Homodimer; N-terminal dimerization domain; C-terminal double PH domains
CAF-1 DNA replication
Vps75 H3/H4
Homodimer; 6 α helices in "headphone" shape;
C-terminal acidic tail
2 Vps75 : 2 H3/H4 Rtt109 Stimulates Rtt109 acetyltransferase activity; transcription elongation FKBP H3/H4 rapamycin JDP2 H3/H4
DAXX H3/H4 6 α helices in histone-binding domain 1 DAXX : 1 H3/H4
ATRX, HDAC2,
CK2
Telomeric and pericentric silencing; chromatin assembly
on gene regulatory elements
DEK H3/H4 EcR Transcriptional co-activator of
EcR
Scm3 CenH3/H4
N-terminal α helix and loop, C-terminal α helix
bind CenH3/H4
2 Scm3 : 2 CenH3/H4
Deposition of centromeric CenH3/H4
10
H3/H4 histones are ubiquitiously chaperoned by Asf1 (Anti-silencing function 1), as it is the only chaperone known to interact with the H3.1, H3.2, and H3.3 variants. Asf1 was originally identified as a derepressor of silent genes in S. cerevisiae when
overexpressed, and later revealed to exist primarily in Drosophila melanogaster complexed with newly synthesized H3/H4 (Le et al., 1997; Tyler et al., 1999). Asf1 stimulates chromatin assembly in a synergistic manner with another H3/H4 chaperone, CAF-1, in vitro and in vivo (Tyler et al., 1999). The importance of Asf1 is exemplified by
S. pombe and S. cerevisiae mutants lacking functional Asf1. The former experience
lethality, while the latter acquires growth defects, as well as sensitivity to DNA damage and replicational stress (Tyler et al., 1999; Umehara et al., 2002).
Structures of yeast and human Asf1 bound to H3/H4 reveal the interactions of Asf1 with both H3 and H4 that promote the formation of a high-affinity complex (English et al., 2006; Natsume et al., 2007). Importantly, Asf1 binds to H3/H4 heterodimers by obstructing the H3-H3’ tetramerization interface (Figure 1.2). The structured core of Asf1 folds in an anti-parallel arrangement of β strands, which interacts extensively with H3 and H4 while sequestering large surface areas of the histones. As a result, Asf1 binds to H3/H4 more tightly than other H3/H4 histone chaperones studied to date (KD ~ 0.3 - 2 nM) (Donham et al., 2011; Zhang et al., 2012b). This is consistent with the finding that most nuclear H3/H4 that is not found in nucleosomes is bound with Asf1 (Tagami et al., 2004). Thus, Asf1 is employed as a histone “sink,” sequestering H3/H4 in storage until nucleosome assembly is necessary (Groth et al., 2005). Accordingly, even before newly co-folded H3/H4 translocate into the nucleus, Asf1 accompanies them in the cytoplasm and supports their nuclear import (Campos et al., 2010).
11
Upon nuclear entry, the Asf1-H3/H4 complex functions as an upstream coordinator for further histone processing. Histones H3.1/H4 and H3.3/H4 are
subsequently partitioned to specific nucleosome assembly pathways handled by different chaperones (Tagami et al., 2004). Downstream histone chaperone engagement with the specific histone variants H3.1 and H3.3 has been extensively characterized, leading to the segregated classification of replication-coupled and replication–independent histone chaperones. The most studied H3/H4 chaperones include CAF-1 and HIRA (HIstone Regulatory Homolog A), which funnel H3/H4 through these respective pathways (Figure 1.3).
Figure 1.2 Structure of Asf1 bound to H3/H4. A. Crystal structure of the S. cerevisiae
Asf1 Immunoglobulin-like domain complexed with a heterodimer of histones H3/H4 (PDB accession code 2HUE). Asf1 is colored purple, histone H3 blue, and H4 green. B. Asf1 physically obstructs the four helix bundle that stabilizes the (H3/H4)2 tetramer in the nucleosome. The orange arrow indicates H3-H3’ tetramerization interface.
12
Outside of S phase, Asf1 transfers H3/H4 to HIRA, which facilitates chromatin formation on transcribed regions. S. cerevisiae deleted of HIR genes experience defects in silencing (Osley and Lycan, 1987), and HIRA cooperates with Asf1 to promote
heterochromatin formation in both fission yeast and humans (Zhang et al., 2005). On the
Figure 1.3 Model of H3/H4 transfer from Asf1 to downstream histone chaperones.
Newly synthesized histones H3/H4 (blue and green respectively) are found in a complex along with sNASP (cyan), the small subunit of the CAF-1 complex (S), and HAT1
(olive). This complex promotes HAT1-mediated acetylation of Lys5 and Lys12 on the H4 N-terminal tail (green star). Association with Asf1 precedes the productive hand-off of H3/H4 to the appropriate downstream chaperones, including CAF-1 (small, medium and large subunits are indicated by the letters S, M and L in tan) and HIRA (yellow).
13
other hand, HIRA may also stimulate transcription, as budding yeast that lack HIRA experience synthetic lethality with FACT, and lose stress-induced gene expression (Chujo et al., 2012). These observations suggest that HIRA regulates both nucleosome assembly on silenced genes and disassembly on active ones.
Other replication-independent pathways exist, as H3.3/H4 in higher eukaryotes can also be handled by the DAXX (Death-associated protein) chaperone when it is associated with the ATRX (Alpha-Thalassaemia mental Retardation syndrome X-linked) chromatin remodeller (Lewis et al., 2010). This complex incorporates H3.3/H4 at
telomeric and pericentric heterochromatin, as well as transcription factor binding sites. In addition, the DEK chaperone interacts with EcR (ECdysone Receptor), a nuclear receptor and transcriptional activator. DEK co-activates transcription by targeting H3.3/H4 to regulatory gene elements (Sawatsubashi et al., 2010). These H3.3/H4 histones are found with transcriptionally active marks, such as methylation of H3K4.
To engage in DNA replication-dependent nucleosome assembly pathways, H3/H4 is transferred from Asf1 to the downstream chaperone CAF-1 (Tyler et al., 2001). Given the high affinity of Asf1 for H3/H4, how Asf1 allocates H3/H4 to CAF-1 or HIRA is an open question in understanding histone transactions. A possible mechanism centers on the B domain possessed by both CAF-1 and HIRA, which binds to Asf1 on the opposite face of the Asf1-H3/H4 binding site. The identical binding sites of CAF-1 and HIRA on the edge of the Asf1 β-sheet sandwich creates a mutually exclusive series of interactions that explain the mutually exclusive functions of CAF-1 and HIRA (Figure 1.4).
CAF-1 was originally recovered in a search for factors from human embryonic kidney (HEK) 293 cell nuclei that can stimulate DNA replication on the SV40 origin
14
(Smith and Stillman, 1989). A purified trio of proteins, supplemented with only T antigen and nucleoside triphosphates, was sufficient to assemble regularly spaced nucleosomes on SV40 DNA. This observation was attributed to the histone H3/H4 binding capability of these proteins, collectively named CAF-1. Subsequently, CAF-1 was found conserved in eukaryotes from yeast to humans as the essential H3/H4 chaperone during DNA replication (Kaufman et al., 1997). In addition, these early experiments established that
Figure 1.4 The Asf1 immunoglobulin-like core interacts with H3/H4 and B domains through opposite faces. Ribbon diagram showing the complex of Asf1 (magenta) bound
to the dimer of histones H3 (blue) and H4 (green) (PDB accession code 2HUE),
superimposed onto the Asf1 (magenta) complex with the B domain peptide from HIRA (yellow) (PDB accession code 2I32) and CAF-1 Cac2 subunit (red) (PDB accession code 2Z3F). Residues in the core domains that differ between the histone H3 variants, H3.1 and H3.3 (amino acids 31, 87, 89, 90 and 96), are shown in orange because of their involvement in replication-dependent and -independent pathways.
15
not only is replication stalled without chromatin assembly, but CAF-1 is unable to
promote assembly without DNA replication, highlighting the intimately coupled behavior between these nuclear phenomena (Hoek and Stillman, 2003).
Much of our current knowledge on histone chaperones come from studies in S.
cerevisiae, which are ideal organisms given their smaller genome, ease of genetic
manipulation, and relative simplicity of histones. Other than the centromeric variant cenH3, S. cerevisiae possess only histone H3, without having S-phase specific and nonspecific variants that the higher organisms have. Budding yeast also lack the DAXX and DEK chaperones (Campos and Reinberg, 2010; Lewis et al., 2010). Fungal H3/H4 is similarly handled by Asf1, which hands off histones to either CAF-1 or HIRA. Yeast chaperones, however, also include Rtt106, which is not found in higher eukaryotes but seems to have overlapping roles with CAF-1 during replication (Huang et al., 2005). Both bind newly synthesized H3/H4 and also associate with each other. If and how the Rtt106-CAF-1 interaction influences replication-dependent nucleosome assembly, however, is unresolved.
The Asf1-CAF-1 Axis in DNA Replication
DNA replication ensures duplication of the eukaryotic genome with the cooperation of several proteins, collectively termed the “replisome.” It involves the MCM2-7 (MiniChromosome Maintenance) complex, which possesses helicase activity necessary for unwinding duplex DNA. The separation of DNA strands results in one parental “leading” strand that serves as a template for the 5’ – 3’ fashioning of a new daughter strand, and a parental “lagging” strand. The latter can not accommodate
16
continuous 5’ – 3’ assembly of a daughter strand. On both parental strands, an RNA primer is established by primase and Pol α (DNA Polymerase α), from which daughter strand synthesis can occur. The primase-Pol α complex is dislodged when the clamp loader RFC (Replication Factor C) allows PCNA (Proliferating Cell Nuclear Antigen) to enclose around the DNA, forcing DNA replication to occur in a processive manner. Subsequently, Pol δ (DNA Polymerase δ) is recruited, and DNA elongation occurs with the concerted unwinding of DNA by MCM2-7. On the lagging strand, complementary 5’ – 3’ RNA primers are placed every couple hundred base pairs, allowing the same
machinery to replicate DNA, albeit in a discontinuous fashion (Lodish et al., 2008). The duplication of genomic DNA requires the duplication of nucleosomes
assembled on DNA. This includes both parental histones and newly synthesized histones, which are deposited by histone chaperones (Ransom et al., 2010). Moreover,
nucleosomes must be destabilized during replication, as the PCNA clamp can not slide over nucleosomes, and Pol δ can not elongate histone-bound DNA. In fact, more than 250 base pairs of naked DNA lies ahead of and behind the replication fork (Gasser et al., 1996; Sogo et al., 1986). These observations suggest that DNA replication is tightly coupled to histone chaperone activity. Indeed, the passage of the replisome is observed to stall or slow with depletion of the histone chaperones FACT, Asf1, and CAF-1 in vivo or
in vitro (Groth et al., 2007; Hoek and Stillman, 2003; Schlesinger and Formosa, 2000).
Clearly, chromatin disassembly ahead of the replication fork, along with assembly behind the fork, is necessary for replisome and S phase progression.
Several histone chaperones localize to replication forks through interactions with replisome proteins. FACT interacts with the MCM helicase complex, facilitating the
17
unwinding of DNA and replication initiation (Tan et al., 2006). Thus, FACT is thought to disassemble H2A/H2B ahead of replication. Asf1 also interacts with the MCM proteins, likely indirectly through H3/H4 (Groth et al., 2007). Importantly, these histones display post-translational modification patterns indicative of parental histones, suggesting that Asf1 disassembles nucleosomes to grant replication initiation. Asf1 is also localized to replication forks by binding to RFC (Franco et al., 2005). This interaction stabilizes the replisome, as RFC and PCNA are uncoupled from the replication fork in yeast lacking Asf1.
After replisome passage, H3/H4 histones are assembled onto DNA by CAF-1. CAF-1 loaded with H3/H4 is recruited to the replication fork through an interaction with PCNA (Shibahara and Stillman, 1999). At the scene, CAF-1 deposits the parental and newly synthesized histones onto newly replicated DNA. Since H3/H4, in contrast to H2A/H2B dimers, may exist in the nucleus as either a dimer or an (H3/H4)2 tetramer, H3/H4 may conceivably be assembled onto DNA by means of two mechanisms
(Annunziato, 2005). H3/H4 histones may be inherited as a single conservative tetramer, which originates from an “old” tetramer belonging to a nucleosome since disassembled, and subsequently transported intact by CAF-1. Alternatively, the histones may be split into two separate dimers, whereupon H3/H4 dimers from the same previous nucleosome would be spread to different daughter nucleosomes. This semi-conservative mechanism would predict a mixing of parental and newly synthesized H3/H4, ensuring that any epigenetic marks would be dispersed to different nucleosomes.
Most studies, however, support a conservative model of histone acquisition. One of the earliest experiments, in 1980, came from in vitro labeling of H3/H4 with a pyrene
18
fluorophore, positioned at the H3:H3’ tetramerization interface. Pyrene emits blue light, but the emission shifts to a longer wavelength when two pyrene molecules are within ~10 Å proximity. Thus, when two H3/H4 dimers form a tetramer, pyrene fluorescence would report the oligomerization as a wavelength shift. The labeled H3/H4 molecules were allowed to be taken up by the slime mold Physarum, which are noted for the ability to incorporate intact, exogenous proteins into normal cellular functions. After several generations, predominantly green emission was observed from the labeled histones, demonstrating that replication had favored conservative inheritance of (H3/H4)2 tetramers. Even after chasing the experiment with unlabeled H3/H4, green emission persisted for generations (Prior and Cantor, 1980). Later, tetramers were visually
observed using electron microscopy of SV40 replicating DNA crosslinked with psoralen, which connects linker but not nucleosomal DNA. The observed DNA consisted of ~90 or ~180 nucleotide bubbles indicative of occupancy by (H3/H4)2 tetramer and (H3/H4)2-2(H2A/H2B) octamers, respectively (Gruss et al., 1993). The length of nucleotides, importantly, does not correspond to H3/H4 dimer size. Since the discovery of the nucleosome, no H3/H4-DNA “disome” species have been observed ex vivo or in vivo, whereas tetrasome intermediates have been evident.
Three decades later, SILAC (Stable Isotope Labeling with Amino acids in Cell culture) experiments revisited the same discussion using more quantitative methods. The expression of N-terminally tagged human H3.1 in HeLa cells was placed under the control of an inducible promoter. H3.1 expression was allowed to proceed and then stopped, generating “old” histones. Subsequent nocodazole treatment synchronized the cells, which proceeded to S phase with nocodazole release. Importantly, the cells were
19
afforded isotopic lysine in culture, in order to distinguish newly synthesized histones from old histones. After up to 72 hours, the epitope-purified H3.1 nucleosomes had negligible amounts of isotopic lysine – detected by mass spectrometry – indicating no association between “old” (H3.1/H4)2 tetramers that were already incorporated, and newly synthesized H3.1. The H2A/H2B dimers were ~50% isotopic, suggesting that these histones are readily exchanged. Importantly, when the same setup was used for histone H3.3, almost a quarter of the affinity-purified nucleosomes were isotopic (Xu et al., 2010). Together, these experiments demonstrate that replication-dependent H3.1 assembly incorporates conservative tetramers, while replication-independent assembly mixes separate dimers. The consistent observation of conservative H3/H4 acquisition in different species support this model, but further studies are needed to decipher the underlying mechanisms.
In particular, understanding the interactions and stoichiometry of the histone chaperone CAF-1 may reveal how parental H3/H4 is acquired conservatively by
replicated DNA (Figure 1.5). In corollary, understanding this complex may prove critical to understanding the possible epigenetic inheritance of parental histone modifications. Although H3/H4 ultimately resides in the nucleosome as a stable (H3/H4)2 tetramer, H3/H4 bound to Asf1 exists in heterodimer form. This is explained by the physical interaction between Asf1 and H3/H4, which occludes the H3:H3’ tetramerization interface seen in the canonical nucleosome. The stoichiometry of the CAF-1-H3/H4 complex, however, is unknown, yet this observation may explain how H3/H4 is deposited as (H3/H4)2 tetramers, rather than semi-conservatively as two disparate H3/H4 dimers
20
Figure 1.5 Model of H3/H4 inheritance during DNA replication. Newly synthesized
H3/H4 (blue and green, respectively) are transferred by Asf1 (purple; PDB accession code 2HUE) to CAF-1 for nucleosome assembly on newly replicated DNA (red, either strand). Parental (H3/H4)2 tetramers do not split into two dimers, but remains as an intact tetramer for CAF-1-mediated assembly. For simplicity, histones H2A/H2B, nucleosomal DNA, and replisome components are not depicted.
21
onto newly replicated DNA. Post-translational modifications on H3 or H4, considered to be possible epigenetic marks, would be inherited accordingly.
The architecture of CAF-1 with or without H3/H4 is largely undetermined, even though the Asf1-H3/H4 and the nucleosome crystal structures are available. As for all eukaryotic CAF-1, yeast CAF-1 is a three-subunit complex composed of a large subunit (Cac1), a middle-sized subunit (Cac2), and a small subunit (Cac3), all of which are fairly well-conserved in function with other eukaryotic counterparts. Both Cac2 and Cac3 are strongly predicted by sequence homology to form WD β-propeller domains, whereas Cac1 is predicted to be largely disordered (Kaufman et al., 1997). Interestingly, X. laevis and H. sapiens Cac1 are reported to dimerize, an event needed to induce nucleosome assembly coupled to DNA repair in vitro (Gérard et al., 2006; Quivy et al., 2001). This may provide a means for oligomerizing two CAF-1 complexes, each loaded with an H3/H4 dimer, in order to produce an (H3/H4)2 tetramer. Indeed, other histone chaperones that bind histone tetramers – including Vps75, Nap1, and Rtt106 – dimerize to
accommodate the larger histone oligomer (Bowman et al., 2011; Su et al., 2012). The best biochemical evidence for CAF-1-H3/H4 stoichiometry come from epitope-tagged H3.1 expressed in HeLa cells, which were analyzed for proteins that remained bound to the exogenous H3 (Tagami et al., 2004). The partners included H4, human Asf1 isoforms, the chaperone NASP (Nuclear Autoantigenic Sperm Protein), importin 4, HAT1 (Histone AcetylTransferase 1), and all three CAF-1 subunits.
Unexpectedly, no endogenous H3 was detected. This suggests that tagged H3 is found in complex with H4 as a dimer, not as a tetramer, even in the presence of the endogenous H3/H4 abundantly available. The potential drawback with this strategy is that the H3
22
histone was C-terminally tagged, which may interfere with the H3-H3’ tetramerization interface present near the C-terminus of H3. Unfortunately, how the epitopes may affect H3/H4 oligometrization was not investigated. With the disconnect in intrepreting these biochemical data, more minimal and quantitative systems would be necessary to determine the CAF-1-H3/H4 stoichiometric state.
A greater understanding of interactions among this Asf1-H3/H4-CAF-1 axis would be beneficial, as it is an essential part of eukaryotic replication and potentially epigenetic inheritance. Yet the interactions among Asf1-CAF-1 and CAF-1-H3/H4 complexes, while discovered through biochemical experiments, remain poorly understood from a structural or thermodynamic perspective. Critically, protein
overexpression for human Asf1 isoforms and human CAF-1 subunits are linked to tumor development in solid cancers, correlating with poorer prognosis. The Asf1a isoform is found overexpressed in ovary, thyroid, laryngopharynx, and skin tumors (Das et al., 2009). Asf1b is overexpressed in breast cancer, with a clear positive correlation with patient survival (Corpet et al., 2011). In breast cancer, the human homologs of CAF-1 subunits Cac1 and Cac2 – p150 and p60, respectively – are also overexpressed at the protein level. Overexpression of p150 and p60 at both the mRNA and protein levels correlate well with breast tumor cell proliferation. In particular, p60 expression is a strong indicator of tumor size, proliferation index, and DNA ploidy from patient-derived tumor samples representing cervical, breast, endometrial, and renal cancers, among many other solid cancers. (Mascolo et al., 2012a, 2012b; Polo et al., 2004, 2010; de Tayrac et al., 2013). Importantly, peak p60 expression correlated with adverse outcomes, such as relapse, metastasis, and death. As these histone chaperones work in a concerted manner
23
during DNA replication, their relative overabundance may cause overactive nucleosome assembly that protects the rapidly replicating tumor DNA.
Thesis Statement
Chromatin structure as a canvas to regulate gene expression, DNA replication, and DNA repair has emerged as a complex theme of many concurrent processes. As the major escort of histones, histone chaperones are intimately involved in these schemes, providing protection, structure, and direction to their cargo. Accordingly, mutations of histone chaperones at the genetic level and overexpression at the protein level often associate with enhanced disease risk.
This thesis focuses on one particular histone H3/H4 pathway in Saccharomyces
cerevisiae that involves the chaperones Asf1 and CAF-1. The cooperation between these
proteins dictate H3/H4 incorporation during replication. In Chapter III of this thesis, I analyzed the interactions of Asf1 with H3/H4 and CAF-1 using biophysical methods that include electrophoretic mobility shift assays (EMSA), as well as fluorescence intensity and anisotropy experiments. The results delineate which domains of Asf1 contribute to the overall binding affinity for H3/H4 and CAF-1. The competitive interactions of Asf1 and CAF-1 for H3/H4 are recapitulated, predicting the relative dissociation constant needed for histone transfer. I designed and performed all of these experiments.
In Chapter IV, I dissected the affinities and stoichiometry of the CAF-1-H3/H4 complex using gel filtration chromatography, chemical crosslinking, Förster resonance energy transfer (FRET), and EMSA systems. The results lend insight into the
CAF-1-24
mediated oligomerization of H3/H4 prior to DNA loading. I also detected and quantified the dissociation constants of each CAF-1 subunit and the overall CAF-1 complex for H3/H4, using EMSA, FRET, and fluorescence anisotropy methods. I designed and performed most of these experiments. Mair Churchill (UC Denver) designed and performed initial studies of the FRET experiment.
In Chapter V, I explored the architecture of the CAF-1-H3/H4 complex using limited proteolysis, EMSA, hydrogen/deuterium exchange coupled to mass spectrometry, and chemical crosslinking coupled to mass spectrometry. The results determine sites of direct interactions between the CAF-1 subunits, and between the CAF-1 subunits and H3/H4. They also suggest conformational changes adapted by both CAF-1 and H3/H4 to accommodate each other. Mass spectrometer operation after chemical crosslinking was performed by Travis Nemkov, Chris Ebmeier, and Kirk Hansen (UC Denver).
Hydrogen/deuterium exchange experiments and the subsequent mass spectrometer operation was performed by Jeremy Balsbaugh and Natalie Ahn (UC Boulder).
Together, the results offer insight into the transition of H3/H4 as the histones transfer from different structural complexes in the Asf1-H3/H4-CAF-1-DNA axis.
25
CHAPTER II
MATERIALS AND METHODS1
Preparation of Histones H3/H4
Preparation of Xenopus laevis H3/H4 was carried out as previously detailed, with mutations at H3C110A and H4T71C (Donham et al., 2011; Dyer et al., 2004; Scorgie et al., 2012). For each purification of H3 or H4, fresh bacterial stocks were transformed. pET3A plasmids bearing the H3 or H4 gene and resistance to ampicillin were transformed into 20 µL of Rosetta 2 (DE3) pLysS Escherichia coli (Novagen). The bacterial cells were
allowed to incubate with plasmid for 30 minutes on ice, and heat shocked in a 42 °C water bath for 30 seconds. Following 60 minutes of shaking, at 225 rpm (revolutions per minute) at 37 °C in 500 µL 2XYT (16 g/L Tryptone, 10 g/L Yeast extract, 5.0 g/L NaCl), the cells were plated on agar plates containing 2XYT and 50 µg/µL ampicillin. The plates were incubated overnight at 37 °C.
For growth of Rosetta 2 (DE3) pLysS cells, five cultures of 4 mL 2XYT each – containing 50 µg/µL ampicillin, 25 µg/µL chloramphenicol and 0.1% glucose – were incubated with streaks of the overnight transformation. The cultures were incubated at 37 °C at 225 rpm until visibly cloudy (~2 hours). Four of the cultures were then transferred to 100 mL media containing the same additives. When the OD600 (Optical Density at 600 nm) read 0.4 units, 20 mL of culture were transferred to 1 L of media containing the same additives. Typically, three 0.5 L cultures were prepared. When the OD600 of the 1 L cultures reached 0.4 units, IPTG (Isopropyl β-D-1-thiogalactopyranoside) dissolved in
1 A portion of this research was originally published in Nucleic Acids Research: Liu, et
26
water was added at a final concentration of 0.4 mM for H3 cultures, and 0.2 mM for H4 cultures. After two more hours of shaking at 225 rpm and 37 °C, the cells were harvested by centrifugation at room temperature. The pelleted cells were suspended in 35 mL of buffer containing 50 mM Tris, 100 mM NaCl, 1 mM benzamide, 1 mM βME, pH 7.5. The resuspension was flash frozen with liquid nitrogen, and stored at -80 °C.
For isolation of pure H3 or H4, the frozen cell resuspensions were lysed by thawing at 37 °C. The volume was brought to 50 mL with resuspension buffer containing 50 mM Tris pH, 100 mM NaCl, 5 mM βME, 1 mM EDTA, 1 mM Benzamidine, pH 7.5. The lysate was sonicated at 20% amplitude for 2 minutes, at pulses of 0.5 seconds followed by 0.5 seconds of pause. The lysate was then centrifuged at 10,000 rpm (revolutions per minute) for 10 minutes at room temperature. The resulting supernatant was decanted, while the pellet was washed by 50 mL resuspension buffer containing 0.1% Triton X-100. The cycle of sonication, centrifugation, and washing was repeated twice more, with the final step excluding Triton X-100. The resulting pellet was broken up with 200 µL DMSO (DiMethyl SulfOxide), and incubated at room temperature for 30 minutes. The extract was solubilized with 8 mL Gsol-Tris buffer (20 mM Tris, 6 M Guanidine-HCl, 10 mM DTT, pH 7.5) for 30 minutes at room temperature. This was followed by room temperature centrifugation at 10,000 rpm for 10 minutes. The cycle of DMSO treatment, Gsol-6 solubilization, and centrifugation was repeated once more. After both cycles, the supernatant was saved for dialysis. Dialysis was performed with the supernatant in 3,500 MWCO (Molecular Weight Cut Off) tubing (Millipore), in 1 L of ion-exchanged 8 M urea. The urea solution was changed once.
27
After dialysis, the supernatant was centrifuged for 10,000 rpm for 10 minutes at room temperature. 4 mL of Q Sepharose slurry (GE Healthcare) and 25 mL of SP Sepharose slurry (GE Healthcare) were separately washed with water and equilibrated with 8 M urea in gravity flow columns (Biorad). The supernatant was run through the Q beads by gravity flow. The flow through was saved and then similarly passed through the SP beads. H3 or H4 were eluted with a salt gradient, ranging from 25 mL of 0 mM NaCl to 1.2 M NaCl (all in 8 M urea). The elutions were checked for purity using denaturing gel electrophoresis followed by Coomassie staining. The best elutions, typically in 400 – 1000 mM NaCl, were each extensively dialyzed into Milli-Q water with 3,500 MWCO tubing. The dialyzed solution was then flash frozen with liquid nitrogen, lyophilized, and stored at -80 °C.
For preparation of co-folded H3/H4, the lyophilized histones were individually dissolved in Gsol-6-HEPES (20 mM HEPES, 6 M Guanidine-HCl, 0.5 mM TCEP, pH 7.5) (Dyer et al., 2004). To check for protein concentration and amount, the sample absorption at 280 nm was evaluated using a NanoDrop 2000 spectrophotometer. The extinction coefficient for H3 and H4 are 3840 and 5120 M-1 cm-1, respectively. H3 was added to H4 at an 0.8:1 molar ratio, and Gsol-6-HEPES was added to dilute the final H3/H4 concentration to 0.5 mg/mL. H3/H4 was dialyzed at 4 °C in 3,500 MWCO tubing into 2 L 2 M FB (Folding Buffer) consisting of 10 mM Tris, 2 M NaCl, 1 mM EDTA, 0.5 mM TCEP, pH 7.5. The dialysis solution was changed three times. The dialyzed sample was then concentrated in 10,000 MWCO centrifugal concentrators (Sartorius) to a volume < 2 mL. Typical concentrations range from 10 µM to 150 µM, with no visible precipitate appearing. A 120 mL S200 size exclusion column was prepared by washing
28
with Milli-Q H2O and equilibrating with 2 M FB. The concentrated sample was resolved through this column, with a void retention volume around ~40 mL and H3/H4 eluting at ~80 mL. The fractions were checked for H3/H4 1:1 stoichiometry and high purity via denaturing gel electrophoresis and Coomassie staining. The best samples were
concentrated with 10,000 MWCO centrifugal concentrators, flash frozen, and saved at -80 °C. Typically, using ~10 mg of lyophilized H3 and H4 (total) yielded ~3 mg/mL H3/H4.
Fluorophore Labeling of H4
For labeling of histone H4 at position 71C, ~ 1-5 mg of lyophilized H4 was dissolved in Gsol-6-HEPES. The appropriate fluorophore was incubated at a 30-fold molar excess with H4 at room temperature. The reaction was protected from light, and proceeded for 3 hours with modest shaking on an orbital shaker. Excess, unreacted fluorophore was removed by G-15 Sephadex (Sigma) beads, which were prepared in a 10 mL disposable column. The H4 and fluorophore solution was added to beads equilibrated in Gsol-6-HEPES, then centrifuged at 1,300 rpm for 5 minutes. The flow through was saved, and allowed to co-fold with H3 as described above. The degree of labeling was calculated as the molar ratio of dye to protein. The concentration of dye was calculated by its maximum absorbance value divided by its molar extinction coefficient. The concentration of protein was calculated by its absorbance value at 280 nm, corrected for dye absorbance at 280 nm, and then divided by the protein molar extinction coefficient. Fluorophores were obtained from Life Sciences (Alexa Fluor 532 C5-maleimide: #A-10255, fluorescein-5-maleimide (FM): #F-150,
7-Diethylamino-3-(4'-29
Maleimidylphenyl)-4-Methylcoumarin (CPM): #D-346), or Sigma
(N-(1-Pyrenyl)maleimide): #P-7908). Each fluorophore was prepared in DMSO, with the stock concentration as follows: Alexa Fluor 532 at 10 mM; FM, CPM, and pyrene at 25 mM.
Preparation of Asf1
The Asf1 purification and labeling strategies have been previously described (Donham et al., 2011; Scorgie et al., 2012). pET60-DEST-Asf1 with an N-terminal GST tag and C-terminal His6 tag, along with a P-1C or I3C mutation for fluorophore labeling, was cloned to stop at residues 246, 210, 185, 169, or 155. The constructs were created using the Gateway Technology (Life Sciences). The forward primer were designed to incorporate a 5’AttB1 recombination site followed by the Precission protease site prior to the protein sequence. The reverse primers were designed to incorporate a 3’ AttB1 recombination site adjacent to eight codons. The primers were as follows (+ denotes forward primer, - denotes reverse primer):
Asf11+ 5’GGGGACAAGTTTGTACAAAAAAGCAGGCTTCCTGGAAGTTCTGTTCCAGGG GTGC3’ Asf1 246-5’GGGGACCACTTTGTACAAGAAAGCTGGGTCTACAGACCCCACCTCTTCTTC TCC3’
30 Asf1 210-5’GGGGACCACTTTGTACAAGAAAGCTGGGTCCTCTTCTTCCTCCTCTGCCGCT TC3’ Asf1 185-5’GGGGACCACTTTGTACAAGAAAGCTGGGTCATCTTCATCGTCATCTTCTTCT TC3’
The sequence including and between the primers were then introduced into the pDONR vector (Life Sciences) by att sequence-mediated recombinations. A second recombination reaction incorporated the sequence into the pET60-DEST vector.
Plasmids were then transformed in Rosetta 2 (DE3) pLysS E. coli (Novagen) using the same transformation steps as H3 or H4 transformed into the same cells. During expression, one colony was picked into 300 mL 2XYT, supplemented with 50 µg/mL ampicillin, 34 µg/mL chloramphenicol, and 0.1% glucose. The culture was grown overnight at 37 °C with shaking at 225 rpm. The next morning, 25 mL of overnight culture was used to inoculate 500 mL of 2XYT with 50 µg/mL ampicillin, 16 µg/mL chloramphenicol, and 0.1% glucose. Three to six 500 mL cultures were used. The
cultures were grown to OD600 = 0.6 at 37 °C, and induced with 0.8 mM IPTG at 27 °C for four hours. The cultures were pre-cooled before IPTG induction. The cells were pelleted by centrifugation at 4 °C at 5,000 rpm, then flash frozen and saved at -80 °C.
For each 1 L of culture, Asf1 pellets were resuspended in 50 mL buffer containing 20 mM Tris, 0.5 M NaCl, 10 mM EDTA, 1 mM DTT, pH 7.9. Protease inhibitors were
31
also added, including 2 µg/mL leupeptin, 2 µg/mL aprotonin, 1 µg/mL pepstatin, 1 mM benzamidine, and 1 mM PMSF. The resuspension was sonicated at 20% amplitude, with 0.5 second pulses followed by 0.5 second pauses for 2 minutes. The extract was clarified by centrifugation at 4 °C at 6,000 rpm for 30 minutes. The supernantant was incubated with Glutathione Sepharose 4B (GE Healthcare) with 1 mL slurry for every 1 L bacterial culture, for a total of 2 hours at 4 °C. The protein-bound beads were washed five times with 20 mM Tris, 1 M NaCl, 10 mM EDTA, 1 mM DTT, pH 7.9, each wash
accompanied by 800 rpm centrifugation at 4 °C. The beads were then washed three more times with Protease Cleavage Buffer (50 mM Tris, 150 mM NaCl, 1 mM DTT, pH 7.0), followed by overnight cleavage by Precission protease (GE Healthcare) targeting a cleavage site upstream of the GST tag.
The cleaved proteins were eluted five times with cleavage buffer (10 mM Tris, 0.5 M NaCl, 20 mM βME, pH 7.9). Proteins were then flowed through a Ni++-NTA column (1 mL HisTrapHP; GE Healthcare) in a low imidazole buffer (20 mM Tris, 1 M NaCl, 5 mM imidazole, 10% glycerol, 0.05% BRIJ 35, 0.5 mM TCEP, pH 7.9). Asf1 was eluted through a 5-250 mM imidazole gradient. Fractions containing Asf1 were evaluated for purity using denaturing gel electrophoresis followed by Coomassie staining.
Fluorophore Labeling of Asf1
Alexa Fluor 532 (Life Sciences) was dissolved at 10 mM with DMSO. For labeling, ~10-20 µg of Asf1 was incubated with 20-fold molar excess Alexa Fluor 532 in 25 mM HEPES, 0.5 M NaCl, 0.5 mM TCEP, 0.05% BRIJ, 10% glycerol, pH 7.25 for 2 hours at room temperature. The reaction was quenched with addition of 20 mM βME, and
32
labeled protein was separated from free dye via a Superdex 200 column in buffer
containing 10 mM Tris, 0.5 M NaCl, 0.5 mM TCEP, 10% glycerol, 0.05% BRIJ, pH 7.9.
Preparation of CAF-1 Proteins
The entire CAF-1 complex was expressed and purified from baculovirus-infected insect cells (Figure 2.2). Baculoviral transfer vectors were transfected along with the BD BaculoGold (BD Biosciences, Bright Linearized Baculovirus DNA (#51-552846)) baculoviral backbone in Sf9 insect cells to produce viral stocks. Each viral stock was purified twice and protein expression was confirmed via Western blotting. CAF-1 complex was produced by co-infecting Sf9 cells at 37°C for 48 hours using an MOI (Multiplicity of Infection) of 1 for each CAF-1 subunit. In a typical preparation, 30 mL of Sf9 pellet (equivalent to 3 L of culture) was used. The pellets were homogenized in a 1:1 ratio of pellet to buffer in 20 mM Tris, 350 mM NaCl, 1 mM EDTA, 10 µg/ml of DNase I, 2 µg/mL leupeptin, 2 µg/mL aprotonin, 1 µg/mL pepstatin, 1 mM benzamidine, 1 mM PMSF, pH 7.4. Homogenate was clarified by centrifugation at 10,000 rpm for 30 minutes at 4°C. The supernatant of the clarified lysate was bound to StrepTrap HP column (GE Healthcare) and washed extensively with 20 mM Tris, 350 mM NaCl, 1 mM EDTA, pH 7.4 until the optical density (OD280) of the flow through reached baseline. CAF-1 complex was eluted from the column with 20 mM Tris, 350 mM NaCl, 1 mM EDTA, 2.5 mM dsbiotin, pH 7.4. CAF-1 was further purified by size exclusion
chromatography using a Superdex 200 column (GE Healthcare) in 10 mM Tris, 350 mM NaCl, 1 mM EDTA, 1 mM DTT, pH 7.4. Purified CAF-1 complex (Figure 2.1) was concentrated with 10,000 MWCO centrifugal concentrators, aliquoted in small volumes,
33
and flash frozen. Cloning of vectors was performed by Sarah Roemer (UC Denver), and infection and growth of Sf9 cells were handled by the Tissue Culture Core (UC Denver). For a typical preparation of 30 mL cell pellet, ~1.5-3 mg of CAF-1 was obtained.
For Cac1 and Cac3 purifications, the same procedure was followed. Phosphatase inhibitors were used for Cac1 purification, including 1 mM Na3VO4 and 10 mM NaF. For a typical preparation of 30 mL cell pellet, ~1 mg of Cac1 or Cac3 was obtained.
For Cac2 purification, overexpression of pDEST-550-p60-His6 constructs was carried out in Rosetta pLysS cells (Novagen), which were transformed as described for histones. 5 mL cultures were inoculated with one colony each, and grown at 37 °C overnight with 225 rpm shaking. The cultures were used to inoculate 1 L 2XYT cultures, supplemented with 50 µg/mL ampicillin and 34 µg/mL chloramphenicol, at 37 °C. The cultures were induced at 0.8 OD600 with 0.4 mM IPTG and grown overnight (12-16 hours) at 18 °C. The pellets were harvested with 5,000 rpm centrifugation at 4 °C before flash freezing.
Harvested pellets were sonicated in 10 mM Tris, 350 mM NaCl, 5 mM Imidazole, 15 mM BME, 10 µg/ml DNase I, complete protease tablet (-)EDTA (Roche), pH 7.4. The lysate was clarified by centrifugation at 10,000 rpm for 30 minutes at 4 °C. The supernatant of the clarified lysate was bound to Ni-NTA resin (Qiagen) and washed extensively with 10 mM Tris, 350 mM NaCl, 5 mM Imidazole, 15 mM BME, pH 7.4 until the OD280 of the flow through reached baseline. Cac2 protein was eluted from the column with 10 mM Tris, 350 mM NaCl, 250 mM Imidazole, 15 mM BME, pH 7.4 using a stepwise gradient of 0-15% 1 column volumes (CV), 15-70% 15 CV, and 70-100% CV. Cac2 was further purified by size exclusion chromatography using a Superdex 200
34
column in 10 mM Tris, 350 mM NaCl, 1 mM EDTA, 1 mM DTT, pH 7.4. Purified fractions were evaluated with denaturing gel electrophoresis and Coomassie staining, and the purest fractions were concentrated with 10,000 MWCO centrifugal concentrators. For a typical preparation from 3 L of culture, ~ 10 mg of Cac2 was obtained.
Electrophoretic Shift Mobility Assays
For resolution of protein-protein interactions, proteins were incubated on ice at the indicated concentrations at 10 µL per lane for 30 minutes in 10 mM Tris, 1 mM DTT, pH 8.0. For interactions involving both histone chaperones and histones, histone
chaperone(s) were added to the reaction tubes before histones were added. For
interactions involving histone chaperones, histones, and DNA, chaperones were added first, followed by histones, then DNA.
0.2x TBE (1x = 89 mM Tris, 89 mM boric acid, 2 mM EDTA, pH 8.0) 59:1 acrylamide:bis-acrylamide native gel(s) were prepared fresh on 1.5 mm thick plates siliconized with SigmaCote (Sigma). The gels were then pre-run at 70 V (Volts) for at least 10 minutes on ice. The protein-protein complexes were separated at 70 V for 90 minutes on ice. Gels were scanned on a Typhoon 9400 imager (GE Healthcare) for FM fluorescence (excitation at 488 nm, emission at 526 nm), followed by a scan for Alexa Fluor 532 fluorescence (ex. 532 nm, em. 555 nm). For Western blots, antibodies against H3 (Abcam), Strep II (Novagen), FLAG (Abcam), His6 (Abgent), and Asf1 (Santa Cruz) were used. The unbound Asf1 species were identified according to the free Asf1 lane, while the bound species were slower migrating complexes. The integrated density value
35
(IDV) of each band was quantified by ImageQuant (v. 5.2). To determine binding constants for Asf1-CAF-1 or Asf1-Cac2 interactions, fraction bound (Y) was defined according to Equation 1 and the KD calculated according to Equation 2 by GraphPad Prism (v. 5.0d). Equation 2 assumes that the concentration of Asf1 is less than 10 times the KD value; otherwise, ligand depletion of the titrated (Cac2 or CAF-1) protein would occur.
Equation 1: Y = IDVBound– IDVBound Background / (IDVBound – IDVBound Background + IDVFree – IDVFree Background)
Equation 2: Y = Bmax * B / (KD + B)
Figure 2.1 Asf1, H3/H4 and CAF-1 proteins and labeling schemes used in this study. A. Left gel: SDS-PAGE of purified proteins. The same gel was excited with UV light
(right), then Coomassie blue stained (left). Right hand gels: Coomassie blue-stained gels of purified H3/H4, CAF-1, Cac1, and Cac3. B. Schematic showing positions of *Asf1 labeled at the -1 Cysteine position (residue added at position -1) with Alexa Fluor 532, and H3/H4 labeled at cysteine 71 of H4 with fluorescein.
36
where Bmax is maximum binding, and B is the concentration of Cac2 or CAF-1.
For binding of H3/H4FM to CAF-1, no unbound species was observed. The values of IDVBound– IDVBound Background were used to obtain a BMax in GraphPad Prism, then fraction bound (Y) was determined by the fraction of IDVBound– IDVBound Background to BMax.
The protocol to analyze histone deposition onto DNA was followed as previously described (Donham et al., 2011). Histones H3/H4 and chaperone(s) were allowed to interact in 10 mM Tris, 150 mM NaCl, 0.5 mM TCEP, pH 8.0 on ice for 20 minutes, followed by interaction with DNA on ice for 20 minutes. The complexes were separated on a pre-run 0.2x TBE gel for 150 minutes on ice. Gels were scanned for FM
fluorescence, then stained with 1:10,000 diluted Sybr Green I (Invitrogen) and re-scanned using the same parameters. To quantify the percentage of disomes formed, the IDV of the disome and tetrasome bands computed by ImageQuant were used to calculate Equation 3. Equation 3: % Tetrasomes = (IDVTetrasomes - IDVTetrasome Background/2) / ((IDVDisomes - IDVDisome Background) + (IDVTetrasomes - IDVTetrasome Background /2))
Immunoblotting and Far Western Blotting
For Far Western analysis, the native gels were transferred to a PVDF
37
apparatus (Biorad) in Towbin buffer (25 mM Tris, 192 mM glycine, 20% methanol, pH 8.3), then blocked with 5% BSA (Bovine Serum Albumin) for 60 minutes. 10 mL 1 µM H3/H4 in 5% BSA was then incubated with the PVDF membrane overnight at 4 °C. The membrane was washed extensively with TBST (Tris Buffered Saline Tween: 20 mM Tris, 150 mM NaCl, 0.05% Tween 20). The membrane was then immunoblotted for H3.
Figure 2.2 Primary sequences of proteins used in this Thesis. Amino acid sequences of
H3, H4, Asf1, Cac3, Cac2, and Cac1 are shown. Underlined residues indicate exogenous sequences, which are either epitope tags (as indicated below the sequence) or cloning artifacts.
38
For immunoblotting, the respective gels (either native or denaturing) were transferred to a PVDF membrane as described above. The membrane was then blocked with 5% BSA for > 30 minutes at room temperature, or overnight at 4 °C with gentle agitation on an orbital shaker. The membrane was incubated with 20 mL of a 1:10,000 volume/volume (v/v) dilution of the primary antibody indicated. Typically, 60 minutes at room temperature with gentle agitation was sufficient to achieve maximum signal. For primary antibodies not conjugated to HRP (Horse Radish Peroxidase), the membrane was washed extensively with TBST at room temperature, then incubated with 20 mL of a 1:50,000 dilution of an antibody recognizing the species of the primary antibody (α-mouse or α-rabbit: GE Healthcare). The reaction was washed extensively with TBST, and then 1 mL of chemiluminescent substrate (Millipore #WBKLS0500) was added to each membrane. The chemiluminescence signal was visualized by exposing the membrane to autoradiography film (GeneMate # F-9023-8X10) for time periods from 1 second to 10 minutes.
Antibodies were obtained from Abcam (H3: #ab1791; FLAG: #ab49763; His6: #ab1187), Millipore (Strep II: #71591), and Santa Cruz (S. cerevisiae Asf1: #sc-166482).
Fluorescence Spectroscopy
For all fluorescence scans, a Fluorolog-3 fluorometer (Horiba) equipped with a thermostat was used. 1 nM of Alexa Fluor 532-labeled Asf1 was equilibrated in 3 mL of assay buffer (10 mM Tris, 150 mM KCl, 2 mM MgCl2, 1% glycerol, 0.5 mM TCEP, 0.05% BRIJ, pH 7.5) at 20 °C in a 3.5 mL quartz cuvette with a 10 mm path length (Starna Cells #3-Q-10). Equilibration typically took ~20 minutes and was checked by
39
taking fluorescence scans. H3/H4 was titrated from 200 pM to 900 µM; at each point, the reaction was allowed to equilibrate for 10 minutes before exciting the fluorophore at 528 nm (slit width: 5 nm) and scanning the emission (slit width: 7 nm) from 538 to 558 nm.
Between each experiment, the cuvette was washed extensively with Milli-Q water, then 2 drops of detergent (Starna Cells #CC-05) in 1 mL of Milli-Q water, then extensively with Milli-Q water. This removed all protein from the cuvette, as assessed by subsequent fluorescence scans.
To quantitate binding affinity, the raw fluorescence intensity at 548 nm of each H3/H4 titration point (Fo) was subtracted from background fluorescence, then
normalized to Asf1 fluorescence (Fi) by: Equation 3: (Fi-Fo)/Fi
The resulting fraction bound were determined relative to the binding saturation constant (Bmax), and plotted in GraphPad Prism. Binding affinities were analyzed by Equation 4 for Asf1 mutants 279, 246, 210, or equation 2 for Asf1 mutants 185, 169, 1-155.
Equation 2: Y = Bmax * B / (KD + B) where B is the concentration of H3/H4 titrated.
Equation 4: Y=((Bmax)*(((Asf1+x+kd)-sqrt(((Asf1+x+kd)^2)-(4*x*Asf1)))/(2*Asf1)))
where Bmax is binding saturation constant, Asf1 is the concentration of fluorophore-labeled Asf1, x is the concentration of H3/H4 titrated, and kd is the dissociation constant.
For fluorescence anisotropy measurements of Asf1, 500 µL of 10 nM of Alexa Fluor 532-labeled Asf1 or 100 nM labeled-Asf11-169 was allowed to equilibrate in 20 mM
40
HEPES, 150 mM NaCl, pH 7.4. A 1 mL quartz fluorometer cell was used (Starna Cells #18F-Q-10). Cac2 or CAF-1 was then titrated; at each point after equilibration, Alexa Fluor 532 was excited at 528 nm (slit width: 6 nm) by vertically and horizontally polarized light, while the intensity (slit width: 13 nm) of polarized light (emission 548 nm) was also monitored in the vertical and horizontal planes. The anisotropy was calculated according to Equation 5 and the KD calculated by Equation 2. Typically, the total quench in fluorescence intensity from the interaction was ~ 20%. More significant amounts of quenching would overstate the KD value (Pope et al., 1999).
Euqation 5: r = (IVV – G*IVH) / (IVV + 2G*IVH), where G is the grating factor (G = IHV / IHH) used to correct for fluorometer sensitivity to polarization bias.
For fluorescence anisotropy measurements of H3/H4 with CAF-1 subunits, 750 µL of 25 nM pyrene-labeled H3/H4 was equilibrated at 20 °C with 10 mM Tris, 150 mM KCl, 2 mM MgCl2, 1% glycerol, 0.5 mM TCEP, 0.05% BRIJ, pH 7.5. Cac1, Cac2, Cac3, or pre-bound Cac1/Cac2 were then titrated into H3/H4Pyrene. For CAF-1 titrations, 5 nM of H3/H4Pyrene was used. With the polarizers in place, pyrene was excited at 345 nm (slit width: 5 nm) and measured at 375 nm (slit width: 7 nm). The anisotropy was calculated according to Equation 5. The dissociation constants were calculated by Equation 6 to account for ligand depletion:
Equation 6: Fi = 1+(Fmax)*(((KD+[A*]+[B]i)-sqrt(((KD+[A*]+[B]i)2)-(4*[A*]*[B]i)))/2*[A]))), where i indicates the varying concentrations of unlabeled protein B that were titrated into the labeled protein A*.
For FRET experiments, 750 µL of 2 nM of H3/H4 mix labeled with CPM or FM were equilibrated in 10 mM Tris, 150 mM KCl, 2 mM MgCl2, 1% glycerol, 0.5 mM
