Article

Structural Basis of Poxvirus Transcription: Vaccinia

RNA Polymerase Complexes
Graphical Abstract Authors
Clemens Grimm, Hauke S. Hillen,
Kristina Bedenk, ..., Aladar A. Szalay,
Patrick Cramer, Utz Fischer

Correspondence
pcramer@mpibpc.mpg.de (P.C.),
utz.fischer@biozentrum.uni-wuerzburg.
de (U.F.)

In Brief
Structural analyses of poxvirus RNA
polymerase complexes reveal how the
components collaborate to effect
transcription with host cell cytoplasm.

Highlights
d Isolation and characterization of two native Vaccinia virus
transcription complexes

d 2.8 Å cryo-EM structures of core and complete Vaccinia RNA

polymerase (vRNAP)

d The multisubunit core vRNAP resembles host Pol II with

Vaccinia-specific features

d Complete vRNAP integrates activities required for Vaccinia

early transcription

Grimm et al., 2019, Cell 179, 1537–1550

December 12, 2019 ª 2019 Elsevier Inc.
https://doi.org/10.1016/j.cell.2019.11.024
Article

Structural Basis of Poxvirus Transcription:

Vaccinia RNA Polymerase Complexes
Clemens Grimm,1,9 Hauke S. Hillen,2,9 Kristina Bedenk,1 Julia Bartuli,1 Simon Neyer,2 Qian Zhang,5
Alexander Hüttenhofer,6 Matthias Erlacher,6 Christian Dienemann,2 Andreas Schlosser,7 Henning Urlaub,3,4
Bettina Böttcher,1,7 Aladar A. Szalay,1,5 Patrick Cramer,2,* and Utz Fischer1,5,8,10,*
1Department of Biochemistry and Cancer Therapy Research Center (CTRC), Theodor Boveri-Institute, University of Würzburg, Am Hubland,
97074 Würzburg, Germany
2Department of Molecular Biology, Max Planck Institute for Biophysical Chemistry, Am Fassberg 11, 37077 Göttingen, Germany
3Bioanalytical Mass Spectrometry Group, Max Planck Institute for Biophysical Chemistry, Am Fassberg 11, 37077 Göttingen, Germany
4Bioanalytics, Institute for Clinical Chemistry, University Medical Center Göttingen, Robert Koch Straße 40, 37075 Göttingen, Germany
5Genelux Corporation, 3030 Bunker Hill Street, San Diego, CA 92109, USA
6Division of Genomics and RNomics, Biocenter, Medical University of Innsbruck, 6020 Innsbruck, Austria
7Rudolf-Virchow-Center, University of Würzburg, Josef-Schneider-Straße 4, 97080 Würzburg, Germany
8Helmholtz Institute for RNA-based Infection Research (HIRI) and Helmholtz Centre for Infection Research (HZI), 97080 Würzburg, Germany
9These authors contributed equally
10Lead Contact

*Correspondence: pcramer@mpibpc.mpg.de (P.C.), utz.fischer@biozentrum.uni-wuerzburg.de (U.F.)

https://doi.org/10.1016/j.cell.2019.11.024

SUMMARY expression events were extensively studied for Vaccinia virus,

a non-pathogenic prototype of the Poxviridae family. These
Poxviruses encode a multisubunit DNA-dependent studies uncovered a virus-encoded multisubunit RNA polymer-
RNA polymerase (vRNAP) that carries out viral gene ase and an array of associated factors that ensure the expres-
expression in the host cytoplasm. We report cryo- sion of the viral genome (Broyles, 2003; Kates and McAuslan,
EM structures of core and complete vRNAP enzymes 1967; Munyon et al., 1967).
from Vaccinia virus at 2.8 Å resolution. The vRNAP Upon infection, Vaccinia virus enters the cell via micropinocy-
tosis and becomes uncoated (Chi and Liu, 2012; Moss, 2012).
core enzyme resembles eukaryotic RNA polymerase
Whereas the viral genome is silent in these initial events, all sub-
II (Pol II) but also reveals many virus-specific fea-
sequent steps of the replication cycle are dependent on viral
tures, including the transcription factor Rap94. The transcription and translation processes. Poxviruses coordinate
complete enzyme additionally contains the tran- the different processes of DNA replication and virion formation
scription factor VETF, the mRNA processing factors through timing of expression of individual genes grouped into
VTF/CE and NPH-I, the viral core protein E11, and early, intermediate, and late classes (Baldick and Moss, 1993).
host tRNAGln. This complex can carry out the entire Accordingly, early genes encode factors involved in events that
early transcription cycle. The structures show that shortly follow infection, such as viral DNA replication and inter-
Rap94 partially resembles the Pol II initiation factor mediate gene expression, whereas later processes of the infec-
TFIIB, that the vRNAP subunit Rpo30 resembles the tion cycle, such as virion assembly, require the expression of
Pol II elongation factor TFIIS, and that NPH-I resem- intermediate and late class gene products.
Vaccinia RNA polymerase (vRNAP) consists of eight subunits
bles chromatin remodeling enzymes. Together with
encoded by early viral genes and termed according to their
the accompanying paper (Hillen et al., 2019), these
apparent molecular masses Rpo147, Rpo132, Rpo35, Rpo30,
results provide the basis for unraveling the mecha- Rpo22, Rpo19, Rpo18, and Rpo7 (Liu et al., 2010). These subunits
nisms of poxvirus transcription and RNA processing. show varying degrees of homology to subunits of RNA polymer-
ase II (Pol II), suggesting an evolutionary relationship with the
INTRODUCTION host transcription apparatus (Table S1) (Zimmermann et al.,
2018). At the level of amino acid residues, the two largest subunits
The eukaryotic nucleus contains the machineries for DNA repli- (i.e., Rpo147 and Rpo132) have been reported to be approxi-
cation and gene transcription. Many viruses rely on factors of mately 20% identical to RPB1 and RPB2 of Pol II, respectively
the host cell for their replication and transcription and therefore (Ahn et al., 1990; Ahn and Moss, 1992; Amegadzie et al., 1991;
require at least a transient nuclear phase to ensure viral propaga- Broyles and Moss, 1986; Knutson and Broyles, 2008). To date,
tion. A remarkable exception among eukaryotic DNA viruses are there is no structural information on vRNAPs and their complexes.
the members of the Poxviridae family, whose replication and vRNAP has the catalytic potential to synthesize RNA in a DNA-
transcription are confined to the cytoplasm (Moss, 2013). These dependent manner. However, in vivo it requires additional
processes require virus-encoded factors for the production of factors in order to become specifically directed to viral early, in-
mature mRNAs from the viral genome. Such cytosolic gene termediate, and late class genes. Early transcription has been

Cell 179, 1537–1550, December 12, 2019 ª 2019 Elsevier Inc. 1537
Figure 1. Purification and Characterization of Vaccinia Virus RNA Polymerase Complexes
(A) Purification of Rpo132 and its associated proteins from GLV-1h439-infected cells using anti-FLAG affinity chromatography. Mock purification was performed
from cells infected with untagged GLV-1h68. Specific proteins from the GLV-1h493 elution were resolved on SDS gels and identified by mass spectrometry.
(B) Anti-FLAG eluate from cell extracts infected with GLV-1h439 was separated on a 10%–30% sucrose gradient and proteins visualized by silver staining on
SDS-PAGE.
(C) RNA extension assay with a nucleic acid scaffold mimicking an elongation complex transcription bubble.
(D) Transcription assay with a linearized pSB24 template containing a Vaccinia virus early promoter and early gene termination signal.
See also Figure S1.

studied most extensively and shown to require the heterodimeric poxvirus gene transcription and RNA processing and enabled
Vaccinia early transcription factor (VETF), which interacts with structure determination of functional vRNAP complexes re-
early promoters upstream and downstream of the initiation site ported in the accompanying paper (Hillen et al., 2019).
(Barnes et al., 2015). Together with Rap94, VETF mediates the
recruitment of vRNAP to its promotors and its transition into RESULTS
active elongation (Liu et al., 2010). Rap94 has also been pro-
posed to connect vRNAP with VETF and NPH-I to facilitate termi- Purification of Vaccinia vRNAP Complexes
nation (Zimmermann et al., 2018). Other virus-encoded proteins We developed a purification strategy for the isolation of vRNAP
are used to add a 50 -terminal m7G-cap and a 30 -terminal poly(A)- complexes based on the recombinant Vaccinia virus strain
tail to viral RNAs. They include the heterodimeric Vaccinia termi- GLV-1h439. This virus is derived from the Vaccinia Lister strain
nation factor/capping enzyme (VTF/CE), consisting of subunits GLV-1h68 and expresses a C-terminally hemagglutinin (HA)/
D1 and D12, and the termination factor NPH-I, which acts FLAG tagged vRNAP subunit Rpo132 (Figure S1A). GLV-1h439
together with a poly(A) polymerase to form polyadenylated 30 multiplied at similar rates to the untagged parental GLV-1h68
ends. Whether these factors are part of defined functional strain upon infection of HeLa cells, suggesting that the tag on
vRNAP complexes is unknown. Rpo132 does not interfere with the transcriptional activity and
Here, we describe the isolation of two distinct vRNAP com- replication of the virus (Figure S1B).
plexes from human cells infected by Vaccinia virus: the For affinity purification of vRNAP, HeLaS3 cells were infected
500 kDa vRNAP core enzyme and the 900 kDa complete with GLV-1h439. Extract from infected cells was then subjected
enzyme with six additional viral proteins plus tRNA-GlnTTG to purification on an anti-FLAG column and tagged Rpo132,
(chr17.trna16-GlnTTG, termed tRNAGln) from the host. We deter- along with its interacting partners, was eluted with FLAG peptide
mine the structures of these two complexes by cryoelectron mi- (Figure S1C). The eluate was separated by gel electrophoresis
croscopy (cryo-EM). Whereas the core complex represents the (Figure 1A) and analyzed by mass spectrometry (Table S2). All
active core RNA polymerase, the complete enzyme apparently known subunits of the vRNAP core enzyme (Table S1), as well
represents the packaged machinery containing the factors for as the transcription factor Rap94, the capping enzyme VTF/CE
early gene transcription. The structures reveal similarities and (D1/D12), the termination factor NPH-I, and the early transcrip-
differences between the viral cytoplasmic transcription appa- tion factor subunits VETF-l and VETF-s (A7/D11), were enriched
ratus and the nuclear RNA polymerase machinery. Our results in the GLV-1h439 elution. None of these factors were enriched in
form the basis for unraveling the molecular mechanisms of a control purification performed with extracts from cells infected

1538 Cell 179, 1537–1550, December 12, 2019

 (legend on next page)

Cell 179, 1537–1550, December 12, 2019 1539

with the untagged virus (Figure 1A). This purification also identi- S4, and S5) (Cramer et al., 2001). The two large subunits,
fied the viral core protein E11L and the host tRNAGln (Table S3) as Rpo147 and Rpo132, form two sides of a central cleft that holds
new factors associated with the Vaccinia virus transcription the active center, giving vRNAP the typical bilobal appearance of
apparatus. multisubunit RNA polymerases found in all three domains of life
(Figure 2B) (Armache et al., 2005). Subunits Rpo35 and Rpo7
Vaccinia RNAP Complexes Are Functional form a subassembly on the back of the polymerase body that
When the eluate was analyzed by sucrose gradient centrifuga- contacts both large subunits (Figure 2C).
tion and mass spectrometry, two major complexes became The entry path for the DNA duplex to the cleft is lined by two
apparent. The lighter complex contained all subunits of the ‘‘jaws’’ formed by Rpo147 and subunit Rpo22 (Figure 2C).
vRNAP core enzyme including sub-stoichiometric amounts of Rpo22 assembles with subunits Rpo19 and Rpo18 on the pe-
Rap94 (Figure 1B). Biochemical characterization showed that riphery of the polymerase (Figure 2C). Rpo18 protrudes slightly
this complex represents the catalytically active RNA polymerase from the polymerase body, forming a stalk. At its base, Rpo18
core enzyme, because it is capable of elongating an RNA primer is anchored to the polymerase body and Rpo19, which in turn
in vitro (Figure 1C). However, no transcriptional activity was de- bridges to Rpo22. Rpo30 is only partially visible in the structure
tected on an artificial gene under the control of a fully double- and binds with its N-terminal domain on the outside of the
stranded viral promoter (Figures 1D and S1D), confirming that enzyme near the ‘‘funnel’’ domain of Rpo147 (Figure 2B). The
the core enzyme requires additional factors for initiation. Vaccinia-specific transcription factor Rap94 is likewise only
The second, heavier complex contained all subunits of the partially visible in the core vRNAP structure, with two of its do-
core enzyme and additionally VTF/CE, NPH-I, VETF-l, VETF-s, mains (domain 2 [D2] and C-terminal domain) binding to the pe-
E11L, and tRNAGln (Figure 1B). This complex was capable of riphery of the polymerase on opposite sides of the cleft
early promoter-dependent transcription initiation, elongation, (Figure 2B).
and termination at a viral termination signal in vitro (Figures 1C
and 1D). Taken together, the first complex represents the cata- vRNAP Contains a Conserved Core
lytically active core vRNAP enzyme, whereas the second com- Seven of the eight core vRNAP subunits show structural homol-
plex represents a complete enzyme that comprises core vRNAP ogy to subunits found in Pol II, albeit their degree of similarity
and viral transcription and RNA processing factors and is differs (Figure 3A; Table S1). Therefore, we carried out a struc-
capable of carrying out all steps of the early Vaccinia transcrip- ture-based comparison between vRNAP and S. cerevisiae Pol II
tion cycle. that provides insight into the functional roles for the individual
subunits of vRNAP (Figures 3, S3, S4, and S5) (Armache
Structure of Vaccinia Core vRNAP et al., 2003). The two large subunits Rpo147 and Rpo132,
We analyzed the core vRNAP by single-particle cryo-EM and ob- which form the body of the polymerase, are highly similar to
tained a reconstruction at 2.8 Å resolution (Figures S2A–S2G; their Pol II counterparts Rpb1 and Rpb2, respectively (Figures
Table S4). The high resolution allowed for placement and adjust- 3B, S3, and S4). In particular, the active center and nucleic
ment of homology models or de novo modeling of all eight sub- acid-binding regions are structurally conserved. The active
units. The reconstruction showed additional densities, which site is formed by an invariant DxDxD motif in Rpo147 that binds
were found to stem from Rap94 according to chemical cross- the catalytic metal ion A (Figures 2 and S3; Video S1) and is
linking (Figure S2H; Table S6). Focused classification and flanked by the bridge helix of Rpo147 that traverses the cleft
refinement yielded improved maps that allowed us to model (Figure 2B). However, both Rpo147 and Rpo132 lack several
two domains of Rap94 on opposite sides of the polymerase. regions and are smaller compared to the yeast counterparts
The resulting structure of the vRNAP core enzyme has good ste- (Figures 3B, S3, and S4).
reochemical quality (Table S4) and contains all eight core vRNAP In all known multisubunit RNA polymerases, the two large sub-
subunits, four structural zinc ions, the catalytic magnesium ion A, units are anchored to a dimeric platform at the back of the
and two domains of Rap94. enzyme, formed by Rpb3 and Rpb11 in the case of Pol II (Cramer
The structure shows that core vRNAP resembles multisubunit et al., 2000, 2001; Engel et al., 2013; Fernández-Tornero et al.,
RNA polymerases in eukaryotic cells and, in particular, Pol II (Fig- 2013; Hirata et al., 2008; Hoffmann et al., 2015; Zhang et al.,
ure 2; Video S1). Based on structural and sequence homology, 1999). The vRNAP subunit Rpo35 combines features of both
we annotated domains in all subunits in accordance to their Rpb3 and Rpb11 in one polypeptide (Figures 2C, 3A, and S5).
counterparts in S.cerevisiae Pol II, which serves as a paradigm It contains a Rpb3-like N-terminal part and a C-terminal part
for eukaryotic multisubunit RNA polymerases (Figures 2, S3, that are similar to Rpb11. However, it lacks the zinc-binding motif

Figure 2. Structure of Core Vaccinia RNAP

(A) Schematic depiction of vRNAP subunits. Functional domains are annotated based on structure-based sequence alignment with S.cerevisiae RNA Pol II
(Armache et al., 2005; Cramer et al., 2001). Regions not visible in the core vRNAP structure are shown transparently.
(B) Structure of the core Vaccinia RNA polymerase enzyme. The protein is shown in cartoon depiction, with helices depicted as cylinders. Subunits are colored as
in (A). The active site metal A and bound structural zinc ions are shown as spheres. See also Figure S2.
(C) Cartoon depiction of Vaccinia RNAP subunits with structural details shown. Rpo147 and Rpo132 domains are colored as indicated in (A). The location of the
subunits in the enzyme is indicated schematically.
See also Figures S2, S3, S4, and S5 and Video S1.

1540 Cell 179, 1537–1550, December 12, 2019

 Figure 3. Comparison of Vaccinia RNA Poly-
merase to S.cerevisiae Pol II
(A) Comparison of subunit composition between
core vRNAP and S.cerevisiae Pol II (PDB: 1WCM)
(Armache et al., 2005). The enzymes are depicated
in schematic surface representation. Homologous
subunits are indicated in the table and colored
accordingly.
(B) Detailed comparison of core vRNAP (left) and
S.cerevisiae Pol II (right) (PDB: 1WCM) (Armache
et al., 2005). The largely conserved core is depicted
as schematic surface in gray, and the differing re-
gions are depicted as cartoon. Regions specific to
vRNAP are shown in green and regions specific to
Pol II in red. Regions located at the back of the
enzyme are labeled transparently.
See also Figures S3, S4, and S5.

Vaccinia-Specific Polymerase
Periphery
The structure-based comparison also
demonstrates that the enzyme surface
deviates substantially from that of other
multisubunit RNA polymerases (Fig-
ure 3B). In particular, vRNAP does not
contain counterparts to the Pol II surface
subunits Rpb4, Rpb8, Rpb9, and Rpb12
(Figure 3A). Moreover, differences in
related subunits of vRNAP and Pol II also
map to the surface of the enzymes
(Figure 3B). For example, the clamp core
and the regions responsible for interactions with Rpb12 and domain in the largest subunit is smaller in vRNAP but larger
Rpb10 in Pol II (Figure S5A), consistent with the absence of and involved in transcription factor interactions in Pol II (Ber-
a Rpb12-like subunit in vRNAP. The corresponding location of necky et al., 2017; Martinez-Rucobo et al., 2011; Plaschka
Rpb12 on vRNAP is instead occupied by a helical insertion in et al., 2016). Likewise, the jaw and foot domains in the largest
Rpo35. Rpo7 interacts with Rpo35 and closely resembles the subunit Rpo147 are also smaller. Rpo147 also does not possess
Pol II subunit Rpb10, in both structure and location in the enzyme the long and repetitive C-terminal domain (CTD) found in its Pol II
complex (Figures 2C and S5A). The C-terminal tail of Rpo7, how- counterpart Rpb1. Instead, it contains a short C-terminal tail
ever, extends farther, forming additional interactions with Rpo35 (C-tail) (res. 1,259–1,286) (Figure S3), which is mobile in the
and Rpo132. The Rpo35/Rpo7 subassembly therefore repre- core vRNAP structure and hence not visible. The second largest
sents the viral equivalent to the Rpb3/10/11/12 subassembly in subunit, Rpo132, lacks several small regions and contains a few
Pol II and the a2 homodimer in bacterial RNA polymerases insertions compared to its Pol II counterpart Rpb2. It has an
(Zhang et al., 1999). extended C-tailthat emerges from the clamp and wraps arounds
Rpo22 structurally resembles Rpb5 and is located at a similar the polymerase, traversing across subunit Rpo19 and toward the
position (Figures 2B and 3A), as predicted previously (Knutson foot domain of Rpo147 (Figures 2C and S4).
and Broyles, 2008). Rpo19 is a structural and functional homolog The jaws of vRNAP, formed by Rpo147 and Rpo22, also
of the Pol II subunit Rpb6. As for the latter, the N-terminal tail of show unique features. Whereas the C-terminal assembly
Rpo19 is mobile and hence invisible in our structure (Figure 2A). domain of Rpo22 is highly conserved, its jaw domain adopts
The regions flanking the conserved assembly domain of Rpo19 a unique fold (Figures 2C and 3B) and lacks the TPSA motif
(a1a and a3) are unique to the viral enzyme. Also, helix a1a forms found in its Pol II counterpart Rpb5 that interacts with down-
a contact to Rpo22 that is not observed between the corre- stream DNA (Figure S5B) (Bernecky et al., 2016). The opposite
sponding Pol II subunits Rpb5 and Rpb6 (Figures 2C and S5B). side of the jaw, formed by Rpo147, is smaller and adopts a
The foot domain of Rpo147 lacks some regions found in its Pol different orientation than in Pol II. Near this domain, the unique
II counterpart, and this space is partially occupied by the viral subunit Rpo30 binds at the rim of the cleft (Figures 2B and
Rpo19 a1a helical insertion (Figure S3). In summary, this detailed 3B). Rpo30 does not have a counterpart in Pol II, but its N-ter-
comparison of vRNAP to Pol II shows that the enzyme core is minal domain (NTD) is located in a similar position on the
largely conserved between vRNAP and other multisubunit polymerase as the dissociable Pol II elongation factor TFIIS
polymerases. (Figure 3A), which Rpo30 has been suggested to functionally

Cell 179, 1537–1550, December 12, 2019 1541

resemble based on sequence analysis (Ahn et al., 1990; Hagler nor adjustments. We also placed a newly determined crystal
and Shuman, 1993). structure of the E11 core protein (Figure S7B; Table S5) into
A prominent unique feature of vRNAP is its one-subunit stalk, the density. Then we docked the crystal structure of VTF/CE
formed by Rpo18, which is homologous to the Pol II subunit (Kyrieleis et al., 2014). The location of the bound tRNAGln could
Rpb7. Eukaryotic nuclear RNA polymerases I, II, and III and archeal also be identified. The remaining density regions were traced
RNA polymerase all contain a heterodimeric stalk (Armache et al., de novo and included NPH-I, the Rap94 NTD and central region,
2005; Engel et al., 2013; Fernández-Tornero et al., 2013; Hirata the Rpo30 C-terminal region, a compact domain of VETF-l
et al., 2008; Hoffmann et al., 2015). In Pol II, the stalk is comprised comprising residues 365–436 (VETF-l365–436; Figure 4; Video
of subunits Rpb4 and Rpb7 (Armache et al., 2003) and is involved S2), and several linker regions. The refined atomic model dis-
in multiple protein interactions with transcription factors during plays excellent stereochemistry (Table S4). The complete
different stages of the transcription cycle (Bernecky et al., 2017; vRNAP structure comprises 15 polypeptides and tRNAGln. It
Plaschka et al., 2016; Vos et al., 2018). The overall fold of Rpo18 adopts an oval-shaped, bilobal structure with overall dimensions
is virtually identical to Rpb7, except for a smaller C-terminal region of 220 Å 3 150 Å 3 130 Å (Figure 4B; Video S2). Whereas one
(Figures 2C and S5B). Rpo18 uses its tip domain to bind the poly- lobe is formed by the core vRNAP enzyme, the other lobe con-
merase core with conserved structural elements (Figure S5B). The tains the additional factors E11, VTF/CE, NPH-I, VETF, and
Rpo18 tip domain may restrict movement of the clamp, as pro- Rap94 regions that are not resolved in the core vRNAP structure.
posed for Rpb7 (Armache et al., 2003). In comparison to the
Rpb4-Rpb7 stalk, the C-terminal domain of Rpo18 appears tilted Rap94 Forms a Bridge between vRNAP Core and
toward the polymerase as it protrudes from the enzyme surface Additional Factors
(Figure 3A). In summary, these comparisons suggest that the sur- The complete vRNAP structure shows well-defined density for all
face of vRNAP has evolved specialized features, likely to facilitate parts of Rap94, which interacts with bound factors. In addition to
interactions with virus-specific transcription factors. the two domains observed in the core vRNAP structure, the NTD
(res. 1–94) and the central region (res. 325–580) of Rap94 are well
Transcription Factor Rap94 Spans the vRNAP Cleft defined. The Rap94 domains are distributed over the entire com-
The core vRNAP structure contains the poxvirus-specific tran- plex and are connected by extended linker regions (Figures 4A
scription factor Rap94 bound to the enzyme periphery (Fig- and 5A). Linker 1 (L1; res. 94–107) connects the NTD to D2.
ure 2B). Rap94 may be involved in the recognition of early viral Linker 2 (L2; res. 292–325) emerges from D2 next to the Rpo18
promoters (Ahn et al., 1994) and transcription termination stalk and extends toward Rpo19, passing the C-tail of Rpo147
(Christen et al., 2008). However, no structural information is (Figure 5B). It then continues along the polymerase dock domain
available for Rap94, and sequence-based homology searches to the back of vRNAP. On the other side of the cleft, linker 3 (L3;
do not detect substantial homology to any known proteins. res. 581–637) extends near the wall and protrusion domains of
The two Rap94 domains resolved in the core vRNAP structure Rpo132, where it traverses the binding site of Rpb12 in Pol II.
occupy distant locations on the polymerase surface on opposite L3 then extends through a groove formed by the wall and
sides of the cleft. One of these Rap94 domains, which we refer to external domains of Rpo132 and to the funnel helices of
as D2, comprises residues 107–292 and binds to the top of the Rpo147 and the Rap94 CTD (Figure 5C).
vRNAP clamp, interacting with both Rpo147 and Rpo132 (Fig- The N-terminal region of Rap94 interacts with the C-terminal
ure 2B; Video S1). It is located close to Rpo18 and may stabilize region of NPH-I. Together, they fold into a domain-like module
the stalk in the observed orientation. It consists of a b sheet that contacts VTF/CE and that we termed the ‘‘CE connector’’
flanked by helical regions on either side and shows no structural (CEC). The CEC forms a wedge between the RNA triphospha-
similarity to factors known to interact with the clamp of Pol II. The tase/guanyltransferase and the methyltransferase domains of
CTD of Rap94 comprises residues 637–795 and is located at the VTF/CE, keeping the two domains apart by 10 Å compared to
lobe of Rpo132 (Figure 2B). The CTD contacts the protrusion the VTF/CE crystal structure (Figure 5D). A further contact be-
domain with a b sheet (res. 661–686). The fold of the Rap94 tween Rap94 and NPH-I is buttressed by the dimeric E11 core
CTD does not resemble known Pol II transcription factors. The protein (Figures 5E). D2 of Rap94 adapts tRNAGln to the core
two Rap94 domains are connected via extended linkers that vRNAP (Figure 5F). In contrast to the situation in the core vRNAP
wrap around the polymerase like a belt (Figures 2A and 2B). structure, the C-tail of Rpo147 is ordered in the complete vRNAP
These linkers traverse the binding sites of Pol II subunits that and adopts an extended structure that tethers VTF/CE (Fig-
are absent in vRNAP, including the C-ribbon domain of Rpb9 ure 5B). Thus, Rap94 is highly modular and serves as a scaffold
and the zinc-binding motif of Rpb12. The central region of to assemble the complete vRNAP complex.
Rap94 (res. 317–587) is not visible in the core vRNAP structure.
The Rap94 Central Region Resembles Pol II Initiation
Structure of Complete Vaccinia vRNAP Factor TFIIB
We next determined the structure of the complete vRNAP that The central region of Rap94 in the complete vRNAP (res. 325–
contains additional transcription and RNA processing factors. 580) is reminiscent of a large portion of the Pol II initiation factor
We collected a cryo-EM dataset from the pooled fractions TFIIB (Figure 5G), and we therefore term it ‘‘B-homology region.’’
15–17 of the gradient shown in Figure 1B, which yielded a recon- It comprises a B-ribbon element (res. 325–371), a B-reader
struction at 2.8 Å resolution (Figure S6). The core vRNAP model hairpin (res. 372–385), a B-linker (res. 386–396), and a B-cyclin
could be unambiguously docked into the reconstruction with mi- domain (res. 397–580). In particular, the zinc ribbon fold and

1542 Cell 179, 1537–1550, December 12, 2019

Figure 4. Structure of the Complete vRNAP Complex
(A) Schematic depiction of the additional Vaccinia transcription factors VTF/CE, VETF-I, E11, and NPH-I contained in the complete vRNAP complex with domains
indicated. Rpo30 and Rap94 are also present in the core vRNAP complex.
(B) Overview of the complete vRNAP model, color coding as in (A). vRNAP is shown in gray. The orientation of the view in the left panel is related to the view in in the
left panel of Figure 2B by an approximately 30 rotation counter-clockwise around the viewing axis followed by an approximately 30 rotation counter-clockwise
vertical rotation. The protein is shown in cartoon depiction, with helices depicted as cylinders.
See also Figure S6.

the zinc binding site in the B-ribbon are well conserved between (res. 50–100) that wraps around the base of the jaw domain and
Rap94 and TFIIB. However, the N-terminal part of the B-ribbon is meanders into the cleft toward the trigger loop, a mobile element
formed by two unique helices in Rap94 that participate in zinc of the active center (Figure 6A, inset). The NTD of Rpo30 is con-
coordination via H328 instead of a cysteine. The B-linker and nected to a linker region that extends to the Rpo147 funnel heli-
B-reader appear smaller compared to their TFIIB counterparts ces, forming a short single-turn helical segment (Figure 6A).
but occupy comparable locations between the dock and clamp The CTD of Rpo30 (res. 131–259) shows sequence similarity to
domains of the polymerase (Sainsbury et al., 2013). The B-cyclin domain III of TFIIS, a zinc ribbon that inserts into the polymerase
domain of Rap94 corresponds to the N-terminal cyclin domain of pore to reach the active site of the enzyme (Figure S7A) (Ketten-
TFIIB with respect to its fold and location. Thus, the B-homology berger et al., 2003). This domain is mobile in both of our struc-
region in Rap94 occupies a similar location as TFIIB in Pol II tran- tures, but it can likely insert into the polymerase pore and reach
scription initiation complexes (Plaschka et al., 2016; Sainsbury the vRNAP active site, as observed for domain III of TFIIS (Fig-
et al., 2013), suggesting that Rap94 may function like TFIIB dur- ure 6A) (Kettenberger et al., 2003, 2004). This domain can trigger
ing transcription initiation. nucleolytic RNA cleavage at the Pol II active site, and Vaccinia
vRNAP has been shown to harbor nucleolytic activity; this has
Subunit Rpo30 Distantly Resembles the Pol II Elongation been suggested to be conferred by Rpo30 (Hagler and Shuman,
Factor TFIIS 1993). Thus, Rpo30 contains an NTD that binds to the polymer-
The structures show that the core vRNAP subunit Rpo30 shares ase in a manner reminiscent of domain II of TFIIS and a mobile
similarities with eukaryotic TFIIS, as suggested based on CTD that likely uses a TFIIS-like mechanism to trigger RNA
sequence analysis (Ahn et al., 1990; Hagler and Shuman, cleavage at the vRNAP active site.
1993). The Rpo30 N-terminal domain (res. 23–130) binds to the
rim of the polymerase funnel (Figure 6A) at the location occupied Rpo30 Places Its Phosphorylated C-Tail in the Active
by TFIIS domain II on Pol II (Kettenberger et al., 2003, 2004). Center
Despite their similar location, these domains differ in sequence Rpo30 additionally contains a C-tail (res. 200–259) that is not
and structure. In particular, the Rpo30 NTD contains an insertion resolved in the core vRNAP structure but is clearly visible in

Cell 179, 1537–1550, December 12, 2019 1543

Figure 5. Rap94 and Its Role in the Complete vRNAP Complex
(A) Location of Rap94 in the complete vRNAP structure. The whole model is shown as transparent gray solvent accessible surface with Rap94 shown as solid
cartoon. The active site metal A is shown as sphere.
(B) Details of the Rpo147 C-tail and the Rap94 linker 2 (L2). These two elements are shown in worm mode and the rest of the model as solvent accessible surface.
The Rpo147 C-tail was visible as a diffuse corridor in the cryo-EM density and was manually modeled as Ca trace for this figure. The quality of the density for this
element did not allow assignment of side chains; therefore, this stretch is omitted in the deposited model.
(C) The extended Rap94 linker 3 (L3, shown as worm) connects the B-cyclin domain to the CTD and binds into a cleft on the cRNAP core. The model except for
Rap94-L3 and the Rpo147 C-tail is shown as solvent accessible surface.
(D) Close-up view of the CEC and its interactions with VTF/CE and the NPH-I helicase module. Proteins are shown as cartoon with coloring as in Figure 4.
(E) Details of the E11-Rap94 interactions.
(F) Details of the Rap94 domain 2 interactions.
(G) Comparison of the Rap94 B-homology region (top) to the corresponding elements of yeast TFIIB (PDB: 4BBR) (Sainsbury et al., 2013) (bottom).

1544 Cell 179, 1537–1550, December 12, 2019

 Figure 6. Structure and Interactions of Sub-
unit Rpo30
(A) Comparison of Vaccinia Rpo30 and
S. cerevisiae TFIIS. The proteins are depicted
schematically with domains indicated. The position
of Rpo30 on the core vRNAP complex is shown on
the left, with the rest of the enzyme shown as
transparent surface representation with coloring as
in Figure 2A. The position of TFIIS in the Pol II re-
activation intermediate complex (PDB: 3PO3)
(Cheung and Cramer, 2011) is shown on the right,
with the rest of the enzyme shown as transparent
surface representation.
(B) Cross section through the solvent-accessible
surface of the complete vRNAP complex model in
the area of the active center cleft. The phosphory-
lated C-tail of Rpo30 is shown in orange as sticks
and the phosphate-moieties shown as purple
spheres. The Rap94 B-reader is shown as green
worm.
See also Figure S7A.

confirming previous predictions (Henikoff,

1993). SNF2 family proteins are ATP-
driven motors with two lobes that are con-
nected by one (INO80, Figure S7C, middle
panel) or two (SNF2, Figure S7C, right
panel) extended ‘‘brace’’ helices and two
protrusions that facilitate DNA interac-
tions. The lobes of NPH-I are connected
by a single brace helix, and the C-lobe
contains the ‘‘protrusion II’’ found in mem-
bers of the SNF2 family (Figure S7C, left
panel). An additional common feature is
the surface at the inside of the brace
the complete vRNAP structure (Figure S7A). This tail inserts into formed by the two helicase domains, which is lined by stretches
the pore of the polymerase, running past the active site and of conserved amino acid motifs denoted as motif I-VI (Figure 7B,
into the region that is predicted to interact with the DNA-RNA left panel). The motif II (Walker B) sequence qualifies NPH-I as a
hybrid at the floor of the active center cleft (Figure 6B). The inter- DExH helicase and is strictly conserved over all members of the
actions that hold the C-tail in place are centered around three poxviridae family (Deng and Shuman, 1998). NPH-I additionally
phosphorylated SP sequence motifs for which we find clear den- contains a unique C-terminal region (res. 561–639) that contacts
sity peaks that allowed us to obtain an atomic model for this the NTD of Rap94 as part of the CEC through multiple interac-
Rpo30 region. Although the function of the Rpo30 C-tail remains tions, including an inter-protein b sheet. NPH-I may therefore
unknown, structural superposition with a Pol II elongation com- have evolved from a common ancestor of the SNF2 family and
plex (Gnatt et al., 2001) shows that it would interfere with binding has adapted to its virus-specific function by the acquisition of
of the DNA-RNA hybrid and thus impair formation of a tran- its CTD.
scribing complex. In the accompanying paper (Hillen et al.,
2019), we show that the DNA-RNA hybrid indeed binds at the ex- Host tRNAGln Is an Integral Component of the
pected position and would clash with the Rpo30 C-tail. This indi- Complete vRNAP
cates that the Rpo30 C-tail must be displaced for transcription. A peculiar feature of the complete vRNAP complex is the pres-
ence of the host tRNAGln. RNA sequencing identified the isoac-
Termination Factor NPH-I Resembles Chromatin ceptor tRNAs GlnTTG and GlnCTG as the predominant species
Remodelers (Table S3). We therefore modeled the tRNA as tRNAGln. The
The complete vRNAP structure also contains the Vaccinia termi- binding site of this tRNA molecule is located on the periphery
nation factor NPH-I, consisting of NTDs and CTDs (N-lobe and and the acceptor arm points away from the center of the com-
C-lobe, respectively). NPH-I is located with its N-lobe near the plex (Figure 4B). Only weak density could be detected for the
RNA exit pore of vRNAP (Figures 4B and 7A). A structural homol- acceptor arm of the tRNA, as it is not supported by any protein
ogy search shows a striking similarity to the chromatin remodeler contacts and hence partially mobile. tRNAGln contacts D2 of
INO80 of the SNF2 family (Eustermann et al., 2018) (Figure 7B), Rap94, which forms a broad interface with the anticodon- and

Cell 179, 1537–1550, December 12, 2019 1545

 Figure 7. Interactions of NPH-I and VETF in
the Complete vRNAP Complex
(A) Location of VETF, NPH-I, E11, and tRNAGln in
complete vRNAP. The whole model is shown as
transparent gray solvent accessible surface with the
factors shown as solid cartoon models. Color cod-
ing as in Figure 4.
(B) Details of the NPH-I fold and location of its heli-
case motifs (left). Comparison to INO80 (right) (PDB:
6FHS) (Eustermann et al., 2018). Corresponding re-
gions are colored identically.
(C) Details if the NPH-I interactions with the tRNA
anticodon loop.
(D) Details of the VETF-l fold and its tRNA in-
teractions. Disulfide bridges are shown as sticks.
See also Figure S7B.

Consistent with this, VETF has been

described as a stable heterodimer of
VETF-l and VETF-s (Broyles and Moss,
1988). It is likely that during promoter
recognition, there are major rearrange-
ments in the complete vRNAP that lead to
a positioning of mobile VETF regions onto
the promoter DNA.

DISCUSSION

Here, we present a purification procedure

for endogenous Vaccinia vRNAP com-
plexes from infected cells and report
the first structures of core and com-
plete vRNAP complexes. A comparison to
cellular enzymes, in particular eukaryotic
D-arm (Figure 5F). This interaction displays no prominent con- Pol II, confirms the common evolutionary origin of multisubunit
tacts to particular bases in this area and hence does not confer RNA polymerases and suggests functions of various vRNAP
binding specificity. However, the anticodon loop of tRNAGln is subunits during transcription. Whereas the two large subunits
oriented in a manner in which it can be specifically read out by and the active center cleft are generally conserved, peripheral
the NPH-I N-lobe (Figure 7C) and VETF-l365–436 (Figure 7D), domains, subunits, and factors display virus-specific features.
which may confer specificity for tRNAGln. Due to the many In particular, the viral factor Rap94 associates with vRNAP and
observed interactions of tRNAGln, it is likely important for the sta- contains a central region that resembles the Pol II initiation factor
bility of the complete vRNAP complex. TFIIB and is thus likely involved in transcription initiation. Further-
more, the subunit Rpo30 distantly resembles the Pol II elonga-
The Initiation Factor VETF Is Anchored to tion factor TFIIS and likely confers RNA cleavage activity to
Complete vRNAP vRNAP. Such nucleolytic activity appears conserved among
The Vaccinia initiation factor VETF is known to bind promoter multisubunit RNA polymerases and allows for rescue of the tran-
DNA upstream and downstream of the transcription start site scription machinery in case of backtracking or misincorporation
during initiation of early transcription (Broyles et al., 1991). In (Fish and Kane, 2002). Whereas the protein that facilitates tran-
the complete vRNAP structure, we observe a central domain of script cleavage is stably associated with polymerase I (Pol I)
the large VETF subunit (VETF-l365–436). This domain has a novel and polymerase III (Pol III) (Engel et al., 2013; Fernández-Tornero
fold that is stabilized by three disulfide bonds and provides the et al., 2013; Hoffmann et al., 2015; Neyer et al., 2016), Pol II re-
connection between tRNAGln, the TPase module of VTF/CE, quires the auxiliary factor TFIIS (Kettenberger et al., 2003). A
and the Rpo18 stalk of the vRNAP core enzyme (Figures 7A similar function is fulfilled by the transcript cleavage factors
and 7D). Although only this domain of the 710-amino-acid GreA and GreB in bacterial transcription (Borukhov et al., 1993;
VETF-l peptide chain is visible in our density, it is likely that the Opalka et al., 2003; Polyakov et al., 1998; Stebbins et al.,
entire heterodimeric protein is anchored to the complex by this 1995). Rpo30 also contains a C-tail that is specific to poxviridae
means, as we detect VETF-l and VETF-s in stoichiometric and not found in other large DNA viruses (Mirzakhanyan and Ger-
amounts in the sucrose gradient peak fraction (Figure 1B). shon, 2017). Phosphorylation of this tail region occurs in

1546 Cell 179, 1537–1550, December 12, 2019

packaged virions (Ngo et al., 2016), and it can occupy the vRNAP involved in the recognition of the termination signal. Finally, our
active site, raising the possibility that this is a regulatory modifi- finding that NPH-I structurally resembles chromatin remodeling
cation. A comparable observation has been made in the apo ATPases supports the forward translocation model of Vaccinia
form of Pol I, in which a peptide region of the largest subunit oc- transcription termination.
cupies the active center cleft (Engel et al., 2013; Fernández-Tor- We also identified the homodimeric viral core protein E11 as a
nero et al., 2013). stoichiometric component of the complete vRNAP. The structure
A striking feature of vRNAP is the C-tail located on the largest suggests that E11 is a major contributor to the stability of the com-
subunit Rpo147. Whereas this tail is flexible in the core vRNAP plete vRNAP. E11 is a late viral product, and two temperature-
complex, it binds to the capping enzyme in the complete vRNAP sensitive mutants have been previously identified to map to its
structure. Although structurally not related, the vRNAP C-tail gene (Kato et al., 2008; Wang and Shuman, 1996). One of these,
may thus resemble the Pol II CTD with respect to its function in G66R, does not affect the virus morphogenesis but rather leads to
capping enzyme recruitment, but the Pol II CTD more generally the formation of non-infectious viral particles under non-permis-
acts as an integration hub for transcription-coupled processes sive conditions (Wang and Shuman, 1996). According to the crys-
(Harlen and Churchman, 2017; Jasnovidova and Stefl, 2013). tal structure of E11, this G66R mutant maps to a tight beta-hairpin
The CTD recruits various factors during different phases of tran- and is likely to be a structural mutant. Of note, temperature sensi-
scription in a phosphorylation-dependent manner (Buratowski, tive mutations in VETF-s and Rap94 have been reported to lead to
2009; Hsin and Manley, 2012) and is also involved in recruitment a defect in protein packaging into mature virions (Kane and Shu-
of the capping enzyme (Cho et al., 1997; Fabrega et al., 2003; man, 1992; Li et al., 1994). These findings are consistent with
McCracken et al., 1997; Noe Gonzalez et al., 2018). In the the idea that the complete vRNAP is the unit that is incorporated
accompanying paper, we show that the Rpo147 C-tail acts as into viral progenies and initiates early transcription immediately af-
a tether and alters structure upon rearrangements in the com- ter virus internalization during the infection cycle.
plete vRNAP complex that accompany the formation of an active The incorporation of an uncharged host tRNAGln molecule into
co-transcriptional capping complex (Hillen et al., 2019). a transcription complex is so far unprecedented. The tRNAGln
The additional factors observed in the complete vRNAP struc- forms an integral part of the complete vRNAP particle, and a pre-
ture are unique to the viral machinery. Rap94 acts as an integral sumed loss of tRNAGln is therefore likely to destabilize the com-
building block of the complete vRNAP, because it bridges the plete vRNAP complex. These observations suggest that it might
interaction between the polymerase and the associated factors. be part of a regulatory mechanism to synchronize the Vaccinia
Consistent with this, a loss of this factor leads to generation of vi- replication cycle to the metabolic status of the host cell. It is inter-
rions that lack vRNAP (Zhang et al., 1994). Rap94 binds NPH-I and esting in this regard to note that viral replication critically de-
locks VTF/CE away from the vRNAP core. The structural similarity pends on the amino acid glutamine as the primary energy source
and location of the Rap94 central region to TFIIB hint at a functional (Fontaine et al., 2014). It is hence a possibility that the complete
role during transcription initiation. Consistent with this, Rap94 D2 vRNAP forms in the late phase of viral infection when glutamine
occupies a position that resembles the location of the initiation fac- becomes limiting and uncharged tRNAGln accumulates.
tor TFIIE in the Pol II preinitiation complex (PIC) (Plaschka et al., Vaccinia virus transcription serves as a paradigm for the mo-
2016), and the Rap94 CTD is found at a location that is congruent lecular biology of nucleo-cytoplasmic large DNA viruses, which
with that of TFIIF in Pol II initiation complexes (He et al., 2016; include poxviruses and the African swine fever virus. Unlike
Plaschka et al., 2016). Based on its biochemical composition most other viruses, which rely on the host transcription machin-
and activity, it is likely that the complete vRNAP complex repre- ery, they utilize a virus-encoded multisubunit RNA polymerase,
sents a unit that is packaged into viral progenies and used for early which contains a conserved core in different virus taxa (Koonin
viral transcription upon virus entry into a host cell. and Yutin, 2001; Mirzakhanyan and Gershon, 2017). The vRNAP
Our structures also rationalize known functional data. Anti- structures presented here provide the first structural insight into
bodies directed against an epitope within the CEC of Rap94 the transcription machinery of poxviridiae. This provides a frame-
inhibit the formation of the PIC in vitro (Mohamed et al., 2002), work for future studies aimed at a mechanistic characterization of
underlining the importance of Rap94 for transcription initiation. the viral transcription cycle. In particular, snapshots of vRNAP
Likewise, mutations and deletions within the NPH-I portion of initiation, elongation, and termination will shed light on the transi-
the CEC inhibit termination without affecting its ATPase activity tions that occur during these processes and decipher the mech-
(Mohamed and Niles, 2000; Piacente et al., 2003). For early tran- anisms by which the virus-specific factors mediate transcription.
scription termination, a sequence motif in the transcribed mRNA As a first step in this direction, we provide structures of tran-
triggers the ATPase activity of the single-stranded DNA helicase scribing and co-transcriptional capping complexes of Vaccinia
NPH-I (Broyles, 2003). It has previously been demonstrated that vRNAP in the accompanying paper (Hillen et al., 2019).
both Rap94 and VTF/CE are involved in the recognition of the
termination motif, which may pause the elongating polymerase STAR+METHODS
(Christen et al., 2008; Luo et al., 1995; Tate and Gollnick,
2015). NPH-I may then cause transcript extrusion from the active Detailed methods are provided in the online version of this paper
site by a 50 to 30 translocase activity on the non-template strand and include the following:
(Hindman and Gollnick, 2016; Tate and Gollnick, 2011). Provided
the observed location of the CEC near the putative RNA exit tun- d KEY RESOURCES TABLE
nel is relevant for a termination intermediate, CEC may be d LEAD CONTACT AND MATERIALS AVAILABILITY

Cell 179, 1537–1550, December 12, 2019 1547

 d EXPERIMENTAL MODEL AND SUBJECT DETAILS Received: April 26, 2019
d METHOD DETAILS Revised: August 20, 2019
B Generation of recombinant Vaccinia virus GLV-1h439 Accepted: November 14, 2019
Published: December 12, 2019
B Viral replication analysis
B vRNAP purification REFERENCES
B Initiation assay
B Mass spectrometry analysis Adams, P.D., Afonine, P.V., Bunkóczi, G., Chen, V.B., Davis, I.W., Echols, N.,
B Cross-linking mass spectrometry (XLMS) Headd, J.J., Hung, L.-W., Kapral, G.J., Grosse-Kunstleve, R.W., et al. (2010).
B RNaseq analysis PHENIX: a comprehensive Python-based system for macromolecular struc-
ture solution. Acta Crystallogr. D Biol. Crystallogr. 66, 213–221.
B Structure determination of core vRNAP
B Structure determination of complete vRNAP Ahn, B.Y., Gershon, P.D., Jones, E.V., and Moss, B. (1990). Identification of
rpo30, a vaccinia virus RNA polymerase gene with structural similarity to a eu-
B X-Ray structure determination of E11
caryotic transcription elongation factor. Mol. Cell. Biol. 10, 5433–5441.
d QUANTIFICATION AND STATISTICAL ANALYSIS
Ahn, B.Y., Gershon, P.D., and Moss, B. (1994). RNA polymerase-associated
d DATA AND CODE AVAILABILITY protein Rap94 confers promoter specificity for initiating transcription of
vaccinia virus early stage genes. J. Biol. Chem. 269, 7552–7557.
SUPPLEMENTAL INFORMATION Ahn, B.Y., and Moss, B. (1992). RNA polymerase-associated transcription
specificity factor encoded by vaccinia virus. Proc. Natl. Acad. Sci. USA 89,
Supplemental Information can be found online at https://doi.org/10.1016/j. 3536–3540.
cell.2019.11.024. Amegadzie, B.Y., Holmes, M.H., Cole, N.B., Jones, E.V., Earl, P.L., and Moss,
B. (1991). Identification, sequence, and expression of the gene encoding the
ACKNOWLEDGMENTS second-largest subunit of the vaccinia virus DNA-dependent RNA polymer-
ase. Virology 180, 88–98.
We would like to thank members of the Cramer and Fischer laboratories for Armache, K.-J., Kettenberger, H., and Cramer, P. (2003). Architecture of initi-
helpful advice and discussion. We thank Christoph Wigge and Jyotishman ation-competent 12-subunit RNA polymerase II. Proc. Natl. Acad. Sci. USA
Veepaschit for support in cryo-EM sample preparation. Cryo-EM of the com- 100, 6964–6968.
plete vRNAP was carried out in the cryo-EM facility of the Julius-Maximilians Armache, K.-J., Mitterweger, S., Meinhart, A., and Cramer, P. (2005). Struc-
University Würzburg (DFG - INST 93/903-1 FUGG). We further thank Chung- tures of complete RNA polymerase II and its subcomplex, Rpb4/7. J. Biol.
Tine Lee (deceased) for XLMS. This work was supported by the Deutsche Chem. 280, 7131–7134.
Forschungsgemeinschaft within SFB860 and SPP1935 to P.C. and H.U. and
Baldick, C.J., Jr., and Moss, B. (1993). Characterization and temporal regula-
under Germany’s Excellence Strategy (EXC 2067/1- 390729940) to P.C., the
tion of mRNAs encoded by vaccinia virus intermediate-stage genes. J. Virol.
European Research Council Advanced Investigator Grant TRANSREGULON
67, 3515–3527.
(grant agreement no. 693023) to P.C., and the Volkswagen Foundation to
P.C. U.F. and A.S. were supported by the Deutsche Forschungsgemeinschaft Barnes, C.O., Calero, M., Malik, I., Graham, B.W., Spåhr, H., Lin, G., Cohen,
(Fi 573 7-2 and 18-1) and the Genelux Corporation. A.H. and M.E. obtained A.E., Brown, I.S., Zhang, Q., Pullara, F., et al. (2015). Crystal Structure of a
funding from the Austrian Science Fund (SFB F4411 and P-30486-BBL). The Transcribing RNA Polymerase II Complex Reveals a Complete Transcription
synchrotron data were collected at beamline P13 operated by EMBL Hamburg Bubble. Mol. Cell 59, 258–269.
at the PETRA III storage ring (DESY, Hamburg, Germany). We would like to Bernecky, C., Herzog, F., Baumeister, W., Plitzko, J.M., and Cramer, P. (2016).
thank Thomas Schneider and the team of beamline P13 for assistance in using Structure of transcribing mammalian RNA polymerase II. Nature 529, 551–554.
the beamline. Bernecky, C., Plitzko, J.M., and Cramer, P. (2017). Structure of a transcribing
RNA polymerase II-DSIF complex reveals a multidentate DNA-RNA clamp.
AUTHOR CONTRIBUTIONS Nat. Struct. Mol. Biol. 24, 809–815.
Biasini, M., Bienert, S., Waterhouse, A., Arnold, K., Studer, G., Schmidt, T.,
C.G. processed the complete vRNAP cryo-EM data, built and refined the Kiefer, F., Gallo Cassarino, T., Bertoni, M., Bordoli, L., and Schwede, T.
model, analyzed the model, and prepared figures. H.S.H. processed the (2014). SWISS-MODEL: modelling protein tertiary and quaternary structure us-
core vRNAP cryo-EM data, built and refined the model, analyzed the model, ing evolutionary information. Nucleic Acids Res. 42, W252–W258.
and prepared figures. K.B. established the purification procedure of the Bond, C.S., and Schüttelkopf, A.W. (2009). ALINE: a WYSIWYG protein-
vRNAP complexes. J.B. performed functional vRNAP assays, purified the sequence alignment editor for publication-quality alignments. Acta Crystal-
vRNAP complexes, and prepared figures. S.N. prepared cryo-EM samples logr. D Biol. Crystallogr. 65, 510–512.
of core vRNAP and acquired cryo-EM data. A.A.S. and Q.Z. generated the re- Borukhov, S., Sagitov, V., and Goldfarb, A. (1993). Transcript cleavage factors
combinant Vaccinia virus strain and provided virus for vRNAP purification. C.D. from E. coli. Cell 72, 459–466.
assisted with cryo-EM data collection and processing. A.S. and H.U. per-
Broyles, S.S. (2003). Vaccinia virus transcription. J. Gen. Virol. 84, 2293–2303.
formed MS analysis and XLMS, respectively. A.H. and M.E. performed
RNA-seq analysis of complete vRNAP. B.B. performed initial cryo-EM analysis Broyles, S.S., and Moss, B. (1986). Homology between RNA polymerases of
of the complete vRNAP complex and assisted with cryo-EM data collection of poxviruses, prokaryotes, and eukaryotes: nucleotide sequence and transcrip-
the complete vRNAP complex. P.C., A.A.S., and U.F. designed and supervised tional analysis of vaccinia virus genes encoding 147-kDa and 22-kDa subunits.
the project and acquired funding. C.G., H.S.H., P.C., and U.F. wrote the Proc. Natl. Acad. Sci. USA 83, 3141–3145.
manuscript. Broyles, S.S., and Moss, B. (1988). DNA-dependent ATPase activity associ-
ated with vaccinia virus early transcription factor. J. Biol. Chem. 263,
10761–10765.
DECLARATION OF INTERESTS
Broyles, S.S., Li, J., and Moss, B. (1991). Promoter DNA contacts made by the
Q.Z. is a full-time employee and A.S. and U.F. were former salaried employees vaccinia virus early transcription factor. J. Biol. Chem. 266, 15539–15544.
of the Genelux Corporation. Q.Z., A.S., and U.F. are company shareholders of Buratowski, S. (2009). Progression through the RNA polymerase II CTD cycle.
the Genelux Corporation. Mol. Cell 36, 541–546.

1548 Cell 179, 1537–1550, December 12, 2019

Chen, V.B., Arendall, W.B., 3rd, Headd, J.J., Keedy, D.A., Immormino, R.M., Henikoff, S. (1993). Transcriptional activator components and poxvirus DNA-
Kapral, G.J., Murray, L.W., Richardson, J.S., and Richardson, D.C. (2010). dependent ATPases comprise a single family. Trends Biochem. Sci. 18,
MolProbity: all-atom structure validation for macromolecular crystallography. 291–292.
Acta Crystallogr. D Biol. Crystallogr. 66, 12–21. Hillen, H.S., Bartuli, J., Grimm, C., Dienemann, C., Bedenk, K., Urlaub, H., Sza-
Cheung, A.C.M., and Cramer, P. (2011). Structural basis of RNA polymerase II lay, A., Fischer, U., and Cramer, P. (2019). Structural basis of Poxvirus tran-
backtracking, arrest and reactivation. Nature 471, 249–253. scription: transcribing and capping complexes. Cell 179. Published online
Chi, P.-I., and Liu, H.-J. (2012). Molecular Signaling and Cellular Pathways for December 12, 2019. https://doi.org/10.1016/j.cell.2019.11.023.
Virus Entry. International Scholarly Research Notices 2013, 1–8. Hindman, R., and Gollnick, P. (2016). Nucleoside Triphosphate Phosphohy-
Cho, E.J., Takagi, T., Moore, C.R., and Buratowski, S. (1997). mRNA capping drolase I (NPH I) Functions as a 50 to 30 Translocase in Transcription Termina-
enzyme is recruited to the transcription complex by phosphorylation of the tion of Vaccinia Early Genes. J. Biol. Chem. 291, 14826–14838.
RNA polymerase II carboxy-terminal domain. Genes Dev. 11, 3319–3326. Hirata, A., Klein, B.J., and Murakami, K.S. (2008). The X-ray crystal structure of
Christen, L.A., Piacente, S., Mohamed, M.R., and Niles, E.G. (2008). Vaccinia RNA polymerase from Archaea. Nature 451, 851–854.
virus early gene transcription termination factors VTF and Rap94 interact with Hoffmann, N.A., Jakobi, A.J., Moreno-Morcillo, M., Glatt, S., Kosinski, J., Ha-
the U9 termination motif in the nascent RNA in a transcription ternary complex. gen, W.J.H., Sachse, C., and Müller, C.W. (2015). Molecular structures of un-
Virology 376, 225–235. bound and transcribing RNA polymerase III. Nature 528, 231–236.
Cotter, C.A., Earl, P.L., Wyatt, L.S., and Moss, B. (2017). Preparation of Cell Hsin, J.-P., and Manley, J.L. (2012). The RNA polymerase II CTD coordinates
Cultures and Vaccinia Virus Stocks. Curr. Protoc. Protein Sci. 89, transcription and RNA processing. Genes Dev. 26, 2119–2137.
5.12.1–5.12.18.
Jasnovidova, O., and Stefl, R. (2013). The CTD code of RNA polymerase II: a
Cramer, P., Bushnell, D.A., Fu, J., Gnatt, A.L., Maier-Davis, B., Thompson, structural view. Wiley Interdiscip. Rev. RNA 4, 1–16.
N.E., Burgess, R.R., Edwards, A.M., David, P.R., and Kornberg, R.D. (2000).
Jones, D.T. (1999). Protein secondary structure prediction based on position-
Architecture of RNA polymerase II and implications for the transcription mech-
specific scoring matrices. J. Mol. Biol. 292, 195–202.
anism. Science 288, 640–649.
Kane, E.M., and Shuman, S. (1992). Temperature-sensitive mutations in the
Cramer, P., Bushnell, D.A., and Kornberg, R.D. (2001). Structural basis of tran-
vaccinia virus H4 gene encoding a component of the virion RNA polymerase.
scription: RNA polymerase II at 2.8 angstrom resolution. Science 292,
J. Virol. 66, 5752–5762.
1863–1876.
Kates, J.R., and McAuslan, B.R. (1967). Poxvirus DNA-dependent RNA poly-
Deng, L., and Shuman, S. (1998). Vaccinia NPH-I, a DExH-box ATPase, is the
merase. Proc. Natl. Acad. Sci. USA 58, 134–141.
energy coupling factor for mRNA transcription termination. Genes Dev. 12,
538–546. Kato, S.E.M., Moussatche, N., D’Costa, S.M., Bainbridge, T.W., Prins, C.,
Strahl, A.L., Shatzer, A.N., Brinker, A.J., Kay, N.E., and Condit, R.C. (2008).
Emsley, P., and Cowtan, K. (2004). Coot: model-building tools for molecular
Marker rescue mapping of the combined Condit/Dales collection of tempera-
graphics. Acta Crystallogr. D Biol. Crystallogr. 60, 2126–2132.
ture-sensitive vaccinia virus mutants. Virology 375, 213–222.
Emsley, P., Lohkamp, B., Scott, W.G., and Cowtan, K. (2010). Features and
Kettenberger, H., Armache, K.-J., and Cramer, P. (2003). Architecture of the
development of Coot. Acta Crystallogr. D Biol. Crystallogr. 66, 486–501.
RNA polymerase II-TFIIS complex and implications for mRNA cleavage. Cell
Engel, C., Sainsbury, S., Cheung, A.C., Kostrewa, D., and Cramer, P. (2013). 114, 347–357.
RNA polymerase I structure and transcription regulation. Nature 502, 650–655.
Kettenberger, H., Armache, K.-J., and Cramer, P. (2004). Complete RNA poly-
Eustermann, S., Schall, K., Kostrewa, D., Lakomek, K., Strauss, M., Moldt, M., merase II elongation complex structure and its interactions with NTP and
and Hopfner, K.-P. (2018). Structural basis for ATP-dependent chromatin re- TFIIS. Mol. Cell 16, 955–965.
modelling by the INO80 complex. Nature 556, 386–390.
Knutson, B.A., and Broyles, S.S. (2008). Expansion of poxvirus RNA polymer-
Fabrega, C., Shen, V., Shuman, S., and Lima, C.D. (2003). Structure of an ase subunits sharing homology with corresponding subunits of RNA polymer-
mRNA capping enzyme bound to the phosphorylated carboxy-terminal ase II. Virus Genes 36, 307–311.
domain of RNA polymerase II. Mol. Cell 11, 1549–1561.
Koonin, E.V., and Yutin, N. (2001). Nucleo-cytoplasmic Large DNA Viruses
Falkner, F.G., and Moss, B. (1990). Transient dominant selection of recombi- (NCLDV) of Eukaryotes (American Cancer Society).
nant vaccinia viruses. J. Virol. 64, 3108–3111.
Kyrieleis, O.J., Chang, J., de la Peña, M., Shuman, S., and Cusack, S. (2014).
Fernández-Tornero, C., Moreno-Morcillo, M., Rashid, U.J., Taylor, N.M.I., Crystal structure of vaccinia virus mRNA capping enzyme provides insights
Ruiz, F.M., Gruene, T., Legrand, P., Steuerwald, U., and Müller, C.W. (2013). into the mechanism and evolution of the capping apparatus. Structure 22,
Crystal structure of the 14-subunit RNA polymerase I. Nature 502, 644–649. 452–465.
Fish, R.N., and Kane, C.M. (2002). Promoting elongation with transcript cleav- Li, J., Pennington, M.J., and Broyles, S.S. (1994). Temperature-sensitive
age stimulatory factors. Biochim. Biophys. Acta 1577, 287–307. mutations in the gene encoding the small subunit of the vaccinia virus early
Fontaine, K.A., Camarda, R., and Lagunoff, M. (2014). Vaccinia virus requires transcription factor impair promoter binding, transcription activation, and
glutamine but not glucose for efficient replication. J. Virol. 88, 4366–4374. packaging of multiple virion components. J. Virol. 68, 2605–2614.
Gnatt, A.L., Cramer, P., Fu, J., Bushnell, D.A., and Kornberg, R.D. (2001). Liu, Y., Schmidt, B., and Maskell, D.L. (2010). MSAProbs: multiple sequence
Structural basis of transcription: an RNA polymerase II elongation complex alignment based on pair hidden Markov models and partition function poste-
at 3.3 A resolution. Science 292, 1876–1882. rior probabilities. Bioinformatics 26, 1958–1964.
Hagler, J., and Shuman, S. (1993). Nascent RNA cleavage by purified ternary Liu, X., Li, M., Xia, X., Li, X., and Chen, Z. (2017). Mechanism of chromatin re-
complexes of vaccinia RNA polymerase. J. Biol. Chem. 268, 2166–2173. modelling revealed by the Snf2-nucleosome structure. Nature 544, 440–445.
Harlen, K.M., and Churchman, L.S. (2017). The code and beyond: transcription Luo, Y., Hagler, J., and Shuman, S. (1991). Discrete functional stages of
regulation by the RNA polymerase II carboxy-terminal domain. Nat. Rev. Mol. vaccinia virus early transcription during a single round of RNA synthesis
Cell Biol. 18, 263–273. in vitro. J. Biol. Chem. 266, 13303–13310.
He, Y., Yan, C., Fang, J., Inouye, C., Tjian, R., Ivanov, I., and Nogales, E. (2016). Luo, Y., Mao, X., Deng, L., Cong, P., and Shuman, S. (1995). The D1 and D12
Near-atomic resolution visualization of human transcription promoter opening. subunits are both essential for the transcription termination factor activity of
Nature 533, 359–365. vaccinia virus capping enzyme. J. Virol. 69, 3852–3856.

Cell 179, 1537–1550, December 12, 2019 1549

Martinez-Rucobo, F.W., Sainsbury, S., Cheung, A.C.M., and Cramer, P. Rould, M.A., Perona, J.J., Söll, D., and Steitz, T.A. (1989). Structure of E. coli
(2011). Architecture of the RNA polymerase-Spt4/5 complex and basis of uni- glutaminyl-tRNA synthetase complexed with tRNA(Gln) and ATP at 2.8 A res-
versal transcription processivity. EMBO J. 30, 1302–1310. olution. Science 246, 1135–1142.
McCracken, S., Fong, N., Rosonina, E., Yankulov, K., Brothers, G., Siderovski, Sainsbury, S., Niesser, J., and Cramer, P. (2013). Structure and function of the
D., Hessel, A., Foster, S., Shuman, S., and Bentley, D.L. (1997). 50 -Capping en- initially transcribing RNA polymerase II-TFIIB complex. Nature 493, 437–440.
zymes are targeted to pre-mRNA by binding to the phosphorylated carboxy- Scheres, S.H.W. (2012). RELION: implementation of a Bayesian approach to
terminal domain of RNA polymerase II. Genes Dev. 11, 3306–3318. cryo-EM structure determination. J. Struct. Biol. 180, 519–530.
Mirzakhanyan, Y., and Gershon, P.D. (2017). Multisubunit DNA-Dependent Schrodinger, L.L.C. (2015). The PyMOL Molecular Graphics System,
RNA Polymerases from Vaccinia Virus and Other Nucleocytoplasmic Large- Version 1.8.
DNA Viruses: Impressions from the Age of Structure. Microbiol. Mol. Biol. Stebbins, C.E., Borukhov, S., Orlova, M., Polyakov, A., Goldfarb, A., and Darst,
Rev. 81, e00010–e00017. S.A. (1995). Crystal structure of the GreA transcript cleavage factor from Es-
Mohamed, M.R., and Niles, E.G. (2000). Interaction between nucleoside cherichia coli. Nature 373, 636–640.
triphosphate phosphohydrolase I and the H4L subunit of the viral RNA poly- Stothard, P. (2000). The sequence manipulation suite: JavaScript programs for
merase is required for vaccinia virus early gene transcript release. J. Biol. analyzing and formatting protein and DNA sequences. Biotechniques 28,
Chem. 275, 25798–25804. 1102–1104, 1104.
Mohamed, M.R., Christen, L.A., and Niles, E.G. (2002). Antibodies directed Tate, J., and Gollnick, P. (2011). Role of forward translocation in nucleoside
against an epitope in the N-terminal region of the H4L subunit of the vaccinia triphosphate phosphohydrolase I (NPH I)-mediated transcription termination
virus RNA polymerase inhibit both transcription initiation and transcription of vaccinia virus early genes. J. Biol. Chem. 286, 44764–44775.
termination, in vitro. Virology 299, 142–153.
Tate, J., and Gollnick, P. (2015). The role of vaccinia termination factor and cis-
Moss, B. (2012). Poxvirus cell entry: how many proteins does it take? Viruses 4, acting elements in vaccinia virus early gene transcription termination. Virology
688–707. 485, 179–188.
Moss, B. (2013). Poxvirus DNA replication. Cold Spring Harb. Perspect. Biol. 5, Tegunov, D., and Cramer, P. (2019). Real-time cryo-electron microscopy data
a010199. preprocessing with Warp. Nat. Methods 16, 1146–1152.
Munyon, W., Paoletti, E., and Grace, J.T., Jr. (1967). RNA polymerase activity Vos, S.M., Pöllmann, D., Caizzi, L., Hofmann, K.B., Rombaut, P., Zimniak, T.,
in purified infectious vaccinia virus. Proc. Natl. Acad. Sci. USA 58, 2280–2287. Herzog, F., and Cramer, P. (2016). Architecture and RNA binding of the human
Neyer, S., Kunz, M., Geiss, C., Hantsche, M., Hodirnau, V.-V., Seybert, A., En- negative elongation factor. eLife 5, e14981.
gel, C., Scheffer, M.P., Cramer, P., and Frangakis, A.S. (2016). Structure of Vos, S.M., Farnung, L., Boehning, M., Wigge, C., Linden, A., Urlaub, H., and
RNA polymerase I transcribing ribosomal DNA genes. Nature 540, 607–610. Cramer, P. (2018). Structure of activated transcription complex Pol II-DSIF-
Ngo, T., Mirzakhanyan, Y., Moussatche, N., and Gershon, P.D. (2016). Protein PAF-SPT6. Nature 560, 607–612.
Primary Structure of the Vaccinia Virion at Increased Resolution. J. Virol. 90, Wang, S.P., and Shuman, S. (1996). A temperature-sensitive mutation of the
9905–9919. vaccinia virus E11 gene encoding a 15-kDa virion component. Virology 216,
Noe Gonzalez, M., Sato, S., Tomomori-Sato, C., Conaway, J.W., and Con- 252–257.
away, R.C. (2018). CTD-dependent and -independent mechanisms govern Yang, B., Wu, Y.-J., Zhu, M., Fan, S.-B., Lin, J., Zhang, K., Li, S., Chi, H., Li,
co-transcriptional capping of Pol II transcripts. Nat. Commun. 9, 3392. Y.-X., Chen, H.-F., et al. (2012). Identification of cross-linked peptides from
Opalka, N., Chlenov, M., Chacon, P., Rice, W.J., Wriggers, W., and Darst, S.A. complex samples. Nat. Methods 9, 904–906.
(2003). Structure and function of the transcription elongation factor GreB Zhang, K. (2016). Gctf: Real-time CTF determination and correction. J. Struct.
bound to bacterial RNA polymerase. Cell 114, 335–345. Biol. 193, 1–12.
Pettersen, E.F., Goddard, T.D., Huang, C.C., Couch, G.S., Greenblatt, D.M., Zhang, Y., Ahn, B.Y., and Moss, B. (1994). Targeting of a multicomponent tran-
Meng, E.C., and Ferrin, T.E. (2004). UCSF Chimera–a visualization system scription apparatus into assembling vaccinia virus particles requires RAP94,
for exploratory research and analysis. J. Comput. Chem. 25, 1605–1612. an RNA polymerase-associated protein. J. Virol. 68, 1360–1370.
Piacente, S.C., Christen, L.A., Mohamed, M.R., and Niles, E.G. (2003). Effect of Zhang, G., Campbell, E.A., Minakhin, L., Richter, C., Severinov, K., and Darst,
selected mutations in the C-terminal region of the vaccinia virus nucleoside S.A. (1999). Crystal structure of Thermus aquaticus core RNA polymerase at
triphosphate phosphohydrolase I on binding to the H4L subunit of the viral 3.3 A resolution. Cell 98, 811–824.
RNA polymerase and early gene transcription termination in vitro. Virology Zheng, S.Q., Palovcak, E., Armache, J.-P., Verba, K.A., Cheng, Y., and Agard,
310, 109–117. D.A. (2017). MotionCor2: anisotropic correction of beam-induced motion for
Plaschka, C., Hantsche, M., Dienemann, C., Burzinski, C., Plitzko, J., and improved cryo-electron microscopy. Nat. Methods 14, 331–332.
Cramer, P. (2016). Transcription initiation complex structures elucidate DNA Zimmermann, L., Stephens, A., Nam, S.-Z., Rau, D., Kübler, J., Lozajic, M., Ga-
opening. Nature 533, 353–358. bler, F., Söding, J., Lupas, A.N., and Alva, V. (2018). A Completely Reimple-
Polyakov, A., Richter, C., Malhotra, A., Koulich, D., Borukhov, S., and Darst, mented MPI Bioinformatics Toolkit with a New HHpred Server at its Core.
S.A. (1998). Visualization of the binding site for the transcript cleavage factor J. Mol. Biol. 430, 2237–2243.
GreB on Escherichia coli RNA polymerase. J. Mol. Biol. 281, 465–473. Zivanov, J., Nakane, T., Forsberg, B.O., Kimanius, D., Hagen, W.J., Lindahl, E.,
Rohou, A., and Grigorieff, N. (2015). CTFFIND4: Fast and accurate defocus and Scheres, S.H. (2018). New tools for automated high-resolution cryo-EM
estimation from electron micrographs. J. Struct. Biol. 192, 216–221. structure determination in RELION-3. eLife 7, 163.

1550 Cell 179, 1537–1550, December 12, 2019

STAR+METHODS

KEY RESOURCES TABLE

REAGENT or RESOURCE SOURCE IDENTIFIER

Bacterial and Virus Strains
GLV-1h68 Genelux Corporation N/A
GLV-1h439 This paper N/A
BL21(DE3)pLysS Promega Corporation L1195
Deposited Data
Core vRNAP structure This paper PDB: 6RIC
Core vRNAP cryo-EM map This paper EMD-4888
Complete vRNAP structure This paper PDB: 6RFL
Complete vRNAP cryo-EM map This paper EMD-4868
E11 crystal structure This paper PDB: 6RFG
INO80 Chromatin Remodeler Structure Eustermann et al., 2018 PDB: 6FHS
S.cerevisiae Pol II crystal structure Armache et al., 2005 PDB: 1WCM
B.taurus Pol II elongation complex structure Bernecky et al., 2016 EMD-3218
Vaccinia capping enzyme crystal structure Kyrieleis et al., 2014 PDB: 4CKB
SNF2 Structure Liu et al., 2017 PDB: 5X0X
Bacterial tRNAGln Structure Rould et al., 1989 PDB: 1GSG
S.cerevisiae Pol II – TFIIS crystal structure Sainsbury et al., 2013 PDB: 4BBR
S.cerevisiae Pol II backtracked reactivation structure Cheung and Cramer, 2011 PDB: 3PO3
Experimental Models: Cell Lines
HeLaS3 Sigma-Aldrich 87110901
African green monkey kidney fibroblasts (CV-1) ATCC CCL-70
Oligonucleotides
A24R-5: CCTAGTAGTAACGAAACATCCATCGATAC This paper N/A
A24Rtag-3: CCTTGTAATCAGCGTAATCTGGAACATCGT This paper N/A
ATGGGTACATGGTCACCAGAAAAGACGGCTTGAG
A25Ltag-5: GTTCCAGATTACGCTGATTACAAGGATGA This paper N/A
CGACGATAAGGCAGCATAATATTCTAGTTTGGTAGTA
GATACATATCAATATCATCA
A25L-3: TCGATCGCCTTCAGAAAGACTG This paper N/A
A24R-5: CCTAGTAGTAACGAAACATCCATCGATAC This paper N/A
A25L-3: TCGATCGCCTTCAGAAAGACTG This paper N/A
In-vitro transcription template strand: IDT DNA N/A
GACTTATGATCGGATAAGAGTCCAGCCAATGACA
GA TGCCTCATAGCC
In-vitro transcription non-template strand: IDT DNA N/A
GGCTATGAGGCATCCCATGCGTTGAGGACTCTTA
TC CGATCATAAGTC
In-vitro transcription RNA primer: This paper N/A
GAGUUGUAAUAACAAGGGAAAUGUCAUUGG
Recombinant DNA
Plasmid pSB24 Luo et al., 1991 pSB24
Software and Algorithms
MotionCor2 Zheng et al., 2017 https://msg.ucsf.edu/em/software/
motioncor2.html
Gctf Zhang, 2016 https://www.mrc-lmb.cam.ac.uk/kzhang/
(Continued on next page)

Cell 179, 1537–1550.e1–e6, December 12, 2019 e1

Continued
REAGENT or RESOURCE SOURCE IDENTIFIER
Gautomatch Kai Zhang https://www.mrc-lmb.cam.ac.uk/kzhang/
Relion 3.0 Scheres, 2012 https://github.com/3dem/relion
Warp Tegunov and Cramer, 2019 http://www.warpem.com
Phenix Adams et al., 2010 http://www.phenix-online.org
PyMol 1.8 Schrodinger, 2015 http://www.pymol.org
Coot Emsley and Cowtan, 2004 https://www2.mrc-lmb.cam.ac.uk/personal/
pemsley/coot/
UCSF Chimera Pettersen et al., 2004 https://www.cgl.ucsf.edu/chimera/
Aline Bond and Schüttelkopf, 2009 https://bondxray.org/software/aline.html

LEAD CONTACT AND MATERIALS AVAILABILITY

Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact Utz
Fischer (utz.fischer@biozentrum.uni-wuerzburg.de). All unique/stable reagents generated in this study are available from the Lead
Contact with a completed Materials Transfer Agreement.

EXPERIMENTAL MODEL AND SUBJECT DETAILS

Male African green monkey kidney fibroblasts (CV-1) were purchased from the American Type Culture Collection (ATCC No.
CCL-70). Human HeLa S3 cells (female) were cultured in a 37 C incubator equilibrated with 5% CO2 and 95% humidified atmo-
sphere. Both cell lines were cultured in DMEM (GIBCO) supplemented with 10% FCS and 1% Penicillin/Streptomycin.

METHOD DETAILS

Generation of recombinant Vaccinia virus GLV-1h439

GLV-1h439 was derived from GLV-1h68 with a HA-tag and FLAG-tag inserted at the end of A24R gene (encoding vRNAP subunit
Rpo132). For insertion of the HA/FLAG-doubletag, an A24R transfer vector was constructed. DNA fragments (termed A and B), flank-
ing about 500 bps of each side of the insertion site of the A24R gene were first amplified via PCR with primers A24R-5/A23R-tag3
(product A) and A25Ltag-5/A25L-3 (product B) as listed in the Key Resources Table. A second round of PCR linked A and B fragments
into product C with primers A24R-5 and A25L-3. The PCR product C was cloned into the pCR-Blunt II-TOPO vector using Zero Blunt
TOPO PCR cloning Kit (Invitrogen). The resulting construct pCRII-A24Rtag4 was sequence confirmed. A p7.5E-gpt cDNA fragment
(E.coli xanthine-guanine phosphoribosyltransferase gene under the control of vaccinia 7.5 early promoter), released by Xba I and PstI
restriction digest from the TK transfer vector, was then subcloned into pCRII-A24Rtag4. The gpt selection-expression cassette was
located outside the Vaccinia virus DNA that directs homologous recombination into the virus genome, allowing for transient dominant
selection of vaccinia recombinants (Falkner and Moss, 1990). The final construct A24Rtag-gpt2 was sequence confirmed, and used
to make recombinant virus GLV-1h439, with GLV-1h68 as the parental virus.

Viral replication analysis

Replication of recombinant GLV-1h439 and GLV-1h68 was performed using a standard plaque assay(Cotter et al., 2017). HeLa S3
cells were grown in 24-well plates and infected with virus at a multiplicity of infection (MOI) of 1. After incubation for 1 h at 37 C, me-
dium was replaced by fresh growth medium and samples were collected 2, 24, 48 and 72 h post viral infection (hpi). After three freeze-
thaw cycles, lysates were titrated by plaque assay on CV-1 cells. The assay was performed in triplicate and all samples were
measured in duplicates.

vRNAP purification
For purification of vRNAP from infected cells, HeLa S3 cells were grown in 15-cm plates up to 80%–90% of confluence. The cells
were infected with purified GLV-1h439 with a MOI of 1.2. After 24 h the cells were pelleted and resuspended in lysis buffer
(50 mM HEPES, pH 7.5, 150 mM NaCl, 1,5 mM MgCl2, 0.5% [v/v] NP-40, 1 mM DTT, and complete EDTA-free protease inhibitor
cocktail [Sigma-Aldrich]). For vRNAP purification, the extract was incubated for 3 h at 4 C with 200 mL anti-FLAG Agarose (Sigma).
Beads were washed four times with buffer containing 50 mM HEPES, pH 7.5, 150 mM NaCl, 1.5 mM MgCl2, 0.1% [v/v] NP-40 and
1 mM DTT and equilibrated with elution buffer (50 mM HEPES, pH 7.5, 150 mM NaCl, 1.5 mM MgCl2 and 1 mM DTT). The bead-
bound proteins were eluted with 3 x FLAG peptide, resolved in 12% Bis-Tris gels and visualized by silver staining. For purification
of native vRNAP, the eluate of the anti-FLAG column was concentrated to 1 mg/mL and layered on top of a 10% - 30% sucrose

e2 Cell 179, 1537–1550.e1–e6, December 12, 2019

gradient and centrifuged for 16 h and 35.000 rpm at 4 C in a Beckman 60Ti swing-out rotor. Gradient fractions were fractionated
manually, separated by SDS-PAGE and proteins visualized by silver staining.

Initiation assay
Plasmid pSB24, containing a G-less cassette downstream of a synthetic vaccinia virus early promoter was generously provided by
Dr. Steven Broyles (Purdue University). Construction of the pSB24 vector with Vaccinia virus early termination signal was described in
(Luo et al., 1991). Briefly, by standard genetic manipulation, the sequence from BamHI site to HindIII site of the pSB24 was replaced
with the duplex oligonucleotides. The insert sequences include three tandem copies of Vaccinia early termination signal. A typical
in vitro transcription had a volume of a 50ml and contained 40 mM Tris-HCl, pH 7.9, 1 mM DTT, 2mM spermidine, 6 mM MgCl2,
1 mM ATP, 1 mM CTP, 1 mM GTP, 0.1mM UTP, 20 mCi ⍺[32P]-UTP [6000 Ci/mmol], 80 mM SAM, 400 ng of NdeI-linearized
pSB24 template as well as purified core or complete vRNAP (Luo et al., 1991). The reaction was incubated at 30 C for the indicated
time points before RNA was extracted and precipitated with isopropanol. Transcripts were analyzed by denaturing gel electropho-
resis and visualized by autoradiography.

Mass spectrometry analysis

For protein identification by in-gel digestion each gel lane was cut into 15 slices. The gel bands were destained with 30% acetonitrile
in 0.1 M NH4HCO3 (pH 8.0), shrunk with 100% acetonitrile, and dried in a vacuum concentrator (Concentrator 5301, Eppendorf, Ger-
many). Digests were performed with 0.1 mg trypsin per gel band overnight at 37 C in 0.1 M NH4HCO3 (pH 8.0). After removing the
supernatant, peptides were extracted from the gel slices with 5% formic acid, and extracted peptides were pooled with the super-
natant. Nano LC-MS/MS analyses were performed on an Orbitrap Fusion (Thermo Scientific) equipped with a PicoView Ion Source
(New Objective) and coupled to an EASY-nLC 1000 (Thermo Scientific). Peptides were loaded on capillary columns (PicoFrit, 30 cm x
150 mm ID, New Objective) self-packed with ReproSil-Pur 120 C18-AQ, 1.9 mm (Dr. Maisch) and separated with a 30-min linear
gradient from 3% to 30% acetonitrile and 0.1% formic acid and a flow rate of 500 nL/min. Both MS and MS/MS scans were acquired
in the Orbitrap analyzer with a resolution of 60,000 for MS scans and 15,000 for MS/MS scans. HCD fragmentation with 35% normal-
ized collision energy was applied. A Top Speed data-dependent MS/MS method with a fixed cycle time of 3 s was used. Dynamic
exclusion was applied with a repeat count of 1 and an exclusion duration of 30 s; singly charged precursors were excluded from se-
lection. Minimum signal threshold for precursor selection was set to 50,000. Predictive AGC was used with AGC a target value of 2e5
for MS scans and 5e4 for MS/MS scans. EASY-IC was used for internal calibration. Data analysis was performed against UniProt
Vaccinia Virus database with PEAKS 8.5 software (Bioinformatics Solution Inc.) with the following parameters: parent mass tolerance:
8 ppm, fragment mass tolerance: 0.02 Da, enzyme: trypsin, variable modifications: oxidation (M), pyro-glutamate (N-term. Q), phos-
phorylation (STY), carbamidomethylation (C). Results were filtered to 1% PSM-FDR by target-decoy approach.

Cross-linking mass spectrometry (XLMS)

Protein cross-linking of purified complexes and subsequent mass spectrometry was performed as described previously (Vos et al.,
2018). Briefly, samples were crosslinked with BS3 (Thermo Fisher Scientific) and incubated for 30 min at 30 C. The reaction was
quenched by adding 100 mM Tris-HCl pH 7.5 and 20 mM ammonium bicarbonate (final concentrations) and incubation for 15 min
at 30 C. Proteins were precipitated with 300 mM sodium acetate pH 5.2 and four volumes of acetone overnight at 20 C. The protein
was pelleted by centrifugation, briefly dried, and resuspended in 4 M urea and 50 mM ammonium bicarbonate. Crosslinked proteins
were reduced with DTT and alkylated (Vos et al., 2016). After dilution to 1 M urea with 50 mM ammonium bicarbonate (pH 8.0), the
crosslinked protein complex was digested with trypsin in a 1:50 enzyme-to-protein ratio at 37 C overnight. Peptides were acidified
with trifluoroacetic acid (TFA) to a final concentration of 0.5% (v/v), desalted on MicroSpin columns (Harvard Apparatus) following
manufacturer’s instructions and vacuum-dried. Dried peptides were dissolved in 50 mL 30% acetonitrile/0.1% TFA and peptide
size exclusion (pSEC, Superdex Peptide 3.2/300 column on an ÄKTAmicro system, GE Healthcare) was performed to enrich for
crosslinked peptides at a flow rate of 50 mL min1. Fractions of 50 mL were collected. Fractions containing the crosslinked peptides
(1–1.7 ml) were vacuum-dried and dissolved in 2% acetonitrile/0.05% TFA (v/v) for analysis by LC–MS/MS.
Crosslinked peptides were analyzed as technical duplicates on an Orbitrap Fusion or Orbitrap Fusion Lumos Tribrid Mass Spec-
trometer (Thermo Fisher Scientific), coupled to a Dionex UltiMate 3000 UHPLC system (Thermo Fisher Scientific) equipped with an in-
house-packed C18 column (ReproSil-Pur 120 C18-AQ, 1.9 mm pore size, 75 mm inner diameter, 30 cm length, Dr. Maisch GmbH).
Samples were separated applying the following 58 min gradient: mobile phase A consisted of 0.1% formic acid (v/v), mobile phase
B of 80% acetonitrile/0.08% formic acid (v/v). The gradient started with 5% B, increasing to 8% B on Fusion and 15% on Fusion
Lumos, within 3 min, followed by 8%–42% B and 15%–46% B within 43 min accordingly, then keeping B constant at 90% for
6 min. After each gradient the column was again equilibrated to 5% B for 6 min. The flow rate was set to 300 nL min1. MS1 spectra
were acquired with a resolution of 120,000 in the Orbitrap covering a mass range of 380–1580 m/z. Injection time was set to 60 ms and
automatic gain control target to 5 3 105. Dynamic exclusion covered 10 s. Only precursors with a charge state of 3–8 were included.
MS2 spectra were recorded with a resolution of 30,000 in the Orbitrap, injection time was set to 128 ms, automatic gain control target
to 5 3 104 and the isolation window to 1.6 m/z. Fragmentation was enforced by higher-energy collisional dissociation at 30%.
Raw files were converted to mgf format using ProteomeDiscoverer 1.4 (Thermo Scientific, signal-to-noise ratio 1.5, 1,000–
10,000 Da precursor mass). For identification of crosslinked peptides, files were analyzed by pLink (v. 1.23), pFind group

Cell 179, 1537–1550.e1–e6, December 12, 2019 e3

(Yang et al., 2012) using BS3 as crosslinker and trypsin as digestion enzyme with maximal two missed cleavage sites. Carbamido-
methylation of cysteines was set as a fixed modification, oxidation of methionines as a variable modification. Searches were conduct-
ed in combinatorial mode with a precursor mass tolerance of 5 Da and a fragment ion mass tolerance of 20 ppm The used database
contained all proteins within the complex. The false discovery rate was set to 0.01. Results were filtered by applying a precursor mass
accuracy of ± 10 ppm Spectra of both technical duplicates were combined and evaluated manually.

RNaseq analysis
Libraries were generated from the isolated RNA fraction following the Ion Torrent Ion Total RNA-seq kit v2 (Thermo Fisher; Art. No.
4475936) protocol with the following modifications. Before the libraries were generated 40 ng of the gel-purified RNA was digested
with 10U of RNase T1 (Thermos Fisher; Art. No. EN0541) for 1 min at room temperature. After PCI extraction and ethanol precipita-
tion, the RNA was pre-treated with 5 U Antarctic phosphatase (New England Biolabs; Art. No. M0289) for 30 min at 37 C. After heat
inactivation at 65 C, the RNA was phosphorylated by 20 U T4 polynucleotide kinase (New England Biolabs; Art. No. M0201) for
60 min at 37 C. Adaptor ligation was carried out for 16 hours at 16 C followed by an incubation of 10 min at 50 C. Reverse transcrip-
tion (RT) was performed employing SuperScript III with incubations at 42 C, 50 C and 55 C for 45, 15 and 10 min, respectively. The
RT reactions were purified and the cDNA was amplified by Platinum PCR SuperMix High Fidelity. Resulting libraries were sequenced
using a Ion Proton (Ion Torrent TM) with high-Q.

Structure determination of core vRNAP

Following sucrose gradient purification, fraction 11 (Figure 1B) was diluted 1:50 and concentrated in a Vivaspin concentrator to a
concentration of roughly 50 mg/mL to remove the sucrose. For cryo-EM analysis the sample was centrifuged for 2h at 21,000 g
and diluted 1:1 in a buffer containing 20mM HEPES, pH 7.5, 200mM (NH4)2SO4, 1mM MgCl2 and 5mM 2-mercaptoethanol. 4ml of
sample were applied to glow discharged UltrAu 2/2 (Quantifoil) grids at 4 C and 95% humidity in a Vitrobot (FEI Company), blotted
for 8.5 s at blot force 14 and plunge-frozen in liquid ethane. Cryo-EM data was collected on a Titan Krios G2 electron microscope (FEI
Company) operated at 300 kV with a K2 direct electron detection device operated in counting mode (Gatan) and an energy filter
(Gatan) set to a slit width of 15 eV. Movie stacks of 39 frames were acquired with a total dose of 55 e-/Å2 in counting mode at
a nominal magnification of 165,000x, corresponding to a calibrated pixel size of 0.81 Å/pixel. Dose weighting and motion
correction was performed using MotionCor2 (Zheng et al., 2017). Per-micrograph contrast-transfer function (CTF) estimation was
done using Gctf (Zhang, 2016), as implemented in Relion (Scheres, 2012). A subset of 4,065 particles was manually picked from
the micrographs and used for reference-free 2D classification in Relion and the resulting class averages were used to generate refer-
ence projections. These were then used as templates for automated particle picking using Gautomatch (https://www.mrc-lmb.cam.
ac.uk/kzhang/).
A total of 479,618 particles were extracted with a box size of 300 pixels in Relion and subjected to reference-free 2D classification
followed by initial global 3D refinement using the B.taurus Pol II elongation complex structure as reference (EMD 3218) (Bernecky
et al., 2016), which yielded a reconstruction at 3.1 Å overall resolution (Figure S2). Further 3D classification revealed two distinct
states of vRNAP corresponding to ‘open’ and ‘closed’ state clefts, similar to the motion observed previously for Pol II (Cramer
et al., 2000; 2001). As the two reconstructions did not show any further differences and the closed state class contained more par-
ticles, this class was used for further refinement. Per-particle CTF and motion correction was performed on this particle subset using
Warp (Tegunov and Cramer, 2019) and CTF and beam tilt refinement was additionally performed using Relion. The resulting final
reconstruction from 3D refinement in Relion achieved an overall resolution of 2.8 Å after post-processing with a sharpening B-factor
of 79 Å2. This cryo-EM density was of excellent quality, with clear sidechain densities for the majority of the complex and occasional
density for bound ions. However, we refrained from modeling ions or waters, with the exception of the catalytic metal ion A as its
location and identity can be inferred from previous crystallographic studies as well as the structural zinc ions which are each com-
plexed by four cysteine or histidine residues. In addition to the well-resolved core, the cryo-EM map showed fragmented densities on
either side of the vRNAP cleft which were not of sufficient quality for model building. To improve these regions, soft masks encom-
passing them were cut out from the global reconstruction that was previously low-pass filtered to 10 Å. Focused 3D classification
using these masks and the particle subset used in the global refinement was then used to identify particle subpopulations with strong
occupancy in the desired region. These particle subpopulations were then subjected to focused 3D refinement, which was initially run
without a reference mask until the refinement converged to local searches, from where on the respective mask was provided for
alignment of particles within the masked region. Post-processing of these maps was performed in Relion using the same soft masks
also used in focused classification and refinement. This approach yielded improved densities for the previously poorly resolved
regions.
The initial model of core vRNAP was constructed by docking homology models of Rpo147 and Rpo132 generated by Swissmodel
(Biasini et al., 2014) into the cryoEM density, followed by manual rebuilding of all residues in Coot (Emsley et al., 2010). Subunits
Rpo35, Rpo22, Rpo19, Rpo18 and Rpo7 were built de novo in Coot. The density for the most distal strands of Rpo18 was weak
and improved only moderately upon focused classification and refinement, thus indicating potential mobility. Subunit Rpo30 was
built de novo in the improved map obtained by focused refinement for its binding region. Cross-linking coupled to mass-spectrom-
etry indicated that the initially fragmented densities remaining on either side of the cleft represent Rap94 (Figure S2H), and these re-
gions could be built de novo after focused classification and refinement in the respective maps. The Rap94 linker regions L2 and L4

e4 Cell 179, 1537–1550.e1–e6, December 12, 2019

could be partially built de novo in the global reconstruction. After fitting of all models, very weak density remained at the back of
vRNAP, which corresponds to the B-homology domain of Rap94. Extensive focused classification and refinements efforts on this
region yielded improved maps around the B-ribbon and B-cyclin domains, but these were not of sufficient quality for reliable model
building and we thus omitted these parts from the core vRNAP model. In total, the structure contains models for Rpo147 (UniProt
B9U1I2; res. 2-207; 217-1268), Rpo132 (UniProt: B9U1Q1; res. 8-122; 126-418; 422-448; 458-789; 797-825; 841-1162), Rpo35
(UniProt: B9U1R2; res. 3-305), Rpo22 (UniProt: B9U1I0; res. 1-184); Rpo19 (UniProt: B9U1M4; res. 61-164), Rpo18 (UniProt:
B9U1K4; res. 2-108; 136-159), Rpo7 (UniProt: B9U1G3; res. 2-62), Rpo30 (UniProt: B9U1D1; res. 23-62; 67-151) and Rap94
(UniProt: B9U1I7; res. 106-134; 160-316; 588-619; 627-650; 655-795). The structure was refined using phenix.real_space_refine
(Adams et al., 2010) against a composite map generated from the global refinement map and the focused refinement map using phe-
nix.combine_focused_maps by weighting the individual parts according to their cross-correlation with the model. To validate this
approach, the model was similarly refined against the locally sharpened density obtained during the Relion local resolution estima-
tion, which yielded comparable final results. The final structure displays excellent stereochemistry, as verified by Molprobity (Chen
et al., 2010) (Table S4).
Figures were created with PyMol (Schrodinger, 2015) and UCSF Chimera (Pettersen et al., 2004). Angular distribution plots were
created using a tool distributed with Warp (Tegunov and Cramer, 2019). Sequence identity scores in Table S1 were calculated using
Ident and Sim (https://www.bioinformatics.org/sms2/ident_sim.html) (Stothard, 2000) with the structure-based alignments as input.

Structure determination of complete vRNAP

Sample were prepared as for the core vRNAP. For cryo-EM data collection, R 1.2/1.3 holey carbon grids (Quantifoil) were glow dis-
charged for 90 s (Plasma Cleaner model PDC-002. Harrick Plasma Ithaca, NY/USA) at medium power and 3.5 mL of C2 sample was
applied inside a Vitrobot Mark IV (FEI) at 4 C and 100% relative humidity. The grids were blotted for 3 s and with blot force 5 and
plunged into liquid ethane. The Cryo-EM datasets were collected with a Thermo Fisher Titan Krios G3 and a Falcon III camera
(Thermo-Fischer). Data was acquired with EPU at 300 keV and a primary magnification of 75,000 (calibrated pixel size 1.0635 Å)
in movie-mode with 25 fractions per movie and integrating the electron-signal. The total exposure was 50 e/Å2 over an exposure
time of 4.5 s with 2 exposures per hole.
Dose-weighted, motion-corrected sums of the micrograph movies were calculated with Motioncorr2 (Zheng et al., 2017). The
contrast-transfer function of each micrograph was fitted with CTFFind4 (Rohou and Grigorieff, 2015). An initial set of 1,500 particles
was selected manually and subjected to a 2D-Classification in Relion3-beta (Zivanov et al., 2018). 12 reasonable class averages were
selected as templates for subsequent automated particle picking within Relion and 256,452 particles were picked from 2,224 micro-
graphs. The dataset was then cleaned up by four cycles of 2D classification and particle sorting followed by manual selection of clas-
ses based on the appearance of their class averages resulting in a final dataset of 190,000 good particles. A subset of 20,000 particles
was used to generate an initial model. An initial 3D classification with Relion yielded two major classes which differed obviously in the
density for VTF/CE, and were subjected to 3D refinement. The class of the large particle yielded a 3.3 Å reconstruction. A second
round of automated particle picking was performed with projections from the reconstruction of the large particle as picking templates
and yielded a dataset of 858,702 particles. This dataset was then cleaned up by four cycles of 2D classification and particle sorting
followed by manual particle selection resulting in a final dataset of 618,338 good particles. A 3D classification of this dataset yielded
only highly similar classes and the reconstruction using the full, unclassified dataset yielded the highest resolution of 2.98 Å. With
further per-particle CTF refinement including a per-dataset beam-tilt refinement and per-particle motion-correction (‘polishing’)
within Relion3 we obtained a reconstruction with 2.75 Å resolution.
For model building and refinement, the complete vRNAP density was unambiguously docked with the previously built core vRNAP
model, the crystallographic models of VTF/CE (PDB: 4CKB) (Kyrieleis et al., 2014), the E11 homodimer and bacterial tRNAGln ex-
tracted from PDB entry 1GSG (Rould et al., 1989). The residual density for VETF-l365-436, NPH-I, and Rap94 was assigned and traced
manually within Coot (Emsley et al., 2010) with the guidance of secondary structure predictions from PsiPred (Jones, 1999) and XLMS
data. The final model was refined with Phenix.real_space_refine including an ADP refinement step. During refinement secondary
structure, mild Ramachandran and reference model restraints from the VTF/CE and E11 crystallographic models were imposed. After
a further cycle of manual inspection and automated refinement, water molecules were placed with Coot and a final round of refine-
ment with Phenix.real_space_refine was applied.

X-Ray structure determination of E11

Bacterially overexpressed, hexa-histidine tagged E11 protein was bound to Ni-NTA-Agarose, eluted with 200 mM Imidazole and
dialysed against TBS. The tag was cleaved with tobacco etch virus protease and a final gel filtration chromatography was performed.
Crystals were obtained with the hanging drop vapor diffusion method with reservoir solution containing 20% PEG 4000. For crystal-
lographic phase determination the crystals were derivatized with sodium ethylmercurithiosalicylate and a SAD experiment was per-
formed at beamline MX1/P13 of the PETRA III storage ring of the Deutsches Elektronen-Synchrotron (DESY). Phasing and initial
model building were performed with Phenix.autosol. The model was then refined against a native dataset collected at the same
beamline with Phenix.refine and completed manually within Coot. After three more cycles of manual corrections and automated
refinement including water placement and TLS refinement, the R-factors converged.

Cell 179, 1537–1550.e1–e6, December 12, 2019 e5

QUANTIFICATION AND STATISTICAL ANALYSIS

No statistical methods were used to predetermine sample size. The experiments were not randomized, and the investigators were
not blinded to allocation during experiments and outcome assessment.

DATA AND CODE AVAILABILITY

Coordinate files for the core and complete vRNAP complex structures were deposited with the Protein Data Bank and are available
under accession codes 6RIC and 6RFL, respectively. The cryo-EM density maps for the core and complete vRNAP complexes were
deposited with the Electron Microscopy Data Base (EMDB) and are available under accession codes EMD-4888 and EMD-4868,
respectively. Coordinate and structure files for the E11 crystal structure were deposited with the Protein Data Bank and are available
under accession code 6RFG.

e6 Cell 179, 1537–1550.e1–e6, December 12, 2019

Supplemental Figures

Figure S1. Purification and Activity of vRNAP Complexes, Related to Figure 1

(A) Schematic representation of modified Vaccinia virus genes. A DNA fragment en-coding a HA-FLAG-tag was fused in GLV-1h439 to the 30 end of A24R,
allowing the expression of C-terminally tagged Rpo132.
(B) Replication of GLV-1h439 in comparison to its parental virus GLV-1h68. Virus titer was determined for the indicated time points from infected cells and cell
culture supernatant, respectively.
(C) Schematic representation of the purification strategy.
(D) Scheme of the pSB24 template (top) and nucleic-acid scaffold with RNA in red, template DNA in blue, and non-template chain in light pink (bottom) as used for
the transcription assays in Figures 1C and 1D.
(legend on next page)
Figure S2. Structure Determination of Core vRNAP, Related to Figure 2
(A) Exemplary cryo-EM micrograph of the core vRNAP dataset.
(B) The 32 best aligning class averages from unsupervised 2D classification.
(C) Cryo-EM processing workflow for structure determination.
(D) Focused classification and refinement workflow for improved local maps.
(E) Fourier Shell Correlation (FSC)-plots for cryo-EM reconstructions used.
(F) Angular distribution plot for the global reconstruction of core vRNAP.
(G) Local resolution estimation for the global reconstruction of core vRNAP as implemented in Relion.
(H) Bis(sulfosuccinimidyl)suberate (BS3) crosslinks identified by mass spectrometry used for positioning of Rap94 domains. (Left) Overview of the core vRNAP
structure with regions where strong crosslinks occurred indicated. (Indent 1-3) Proteins are shown in cartoon representation with coloring as in Figure 2.
Crosslinked lysine residues are shown as sticks. Selected strong crosslinks are shown as lines.
(legend on next page)
Figure S3. Structure-Based Sequence Alignment of Rpo147 and S. cerevisae Rpb1, Related to Figure 3
(A) Schematic depiction of Vaccinia Rpo147 and the homologous S.cerevisiae Pol II subunit Rpb1 with domains indicated. Insertions and deletions are indicated
by connecting lines, with differing regions shown with dashed lines. Regions with differing fold are indicated by crossed connecting lines.
(B) Structure-based sequence alignment with secondary structure elements depicted and colored according to domains as in Figures 2A and 2C. Sheet regions
are shown as arrows, helical region as cylinders. Invariant residues are colored in dark blue and conserved residues in light blue. Regions differing in fold are
colored in green (vRNAP-specific) and red (Pol II-specific). The alignment was generated with MSAProbs (Liu et al., 2010) within the MPI Bioinformatics Toolkit
(Zimmermann et al., 2018), visualized using Aline (Bond and Schüttelkopf, 2009) and manually edited by comparison to the S.cerevisiae Pol II structure
(PDB 1WCM) (Armache et al., 2005). In Rpo147, helices a8 and a9 in the polymerase clamp core domain are shortened. Helices a27, a28, a32 and a34, which are
located in the foot domain of Rpb1, are absent. The jaw domain is substantially reduced, lacking Rpb1 regions 1158-1188 and 1245-1253.
(legend on next page)
Figure S4. Structure-Based Sequence Alignment of Rpo132 and S. cerevisae Rpb2, Related to Figure 3.
(A) Schematic depiction of Vaccinia Rpo132 and the homologous S.cerevisiae Pol II subunit Rpb2 with domains indicated. Insertions and deletions are indicated
by connecting lines, with differing regions shown with dashed lines. Regions with differing fold are indicated by crossed connecting lines.
(B) Structure-based sequence alignment with secondary structure elements depicted and colored according to domains as in Figures 2A and 2C. Sheet regions
are shown as arrows, helical region as cylinders. Invariant residues are colored in dark blue and conserved residues in light blue. Regions differing in fold are
colored in green (vRNAP-specific) and red (Pol II-specific). The alignment was generated with MSAProbs (Liu et al., 2010) within the MPI Bioinformatics Toolkit
(Zimmermann et al., 2018), visualized using Aline (Bond and Schüttelkopf, 2009) and manually edited by comparison to the S.cerevisiae Pol II structure (PDB
1WCM) (Armache et al., 2005). Helices a7 and a8 in the lobe domain are extended in the Rpo132. In the protrusion domain, the region between a11 and a12 differs
between the yeast and viral proteins. The most prominent differences are located in the external domains, in particular in the regions between b16 and b17, a16
and a17, and between a19 and b24. The region following b28 (res. 784-797), which contacts upstream DNA in the yeast Pol II (Barnes et al., 2015), is reduced and
adopts a different conformation in the viral enzyme.
(legend on next page)
Figure S5. Structure-Based Sequence Alignment of Rpo35, Rpo22, Rpo19, Rpo18, and Rpo7 with Corresponding S. cerevisae Pol II Subunits,
Related to Figure 3
Structure-based sequence alignments with secondary structure elements depicted and colored according to domains as in Figure 3. Sheet regions are shown as
arrows, helical region as cylinders. Invariant residues are colored in dark blue and conserved residues in light blue. Regions differing in fold are colored in green
(vRNAP-specific) and red (Pol II-specific). The alignment was generated with MSAProbs (Liu et al., 2010) within the MPI Bioinformatics Toolkit (Zimmermann et al.,
2018), visualized using Aline (Bond and Schüttelkopf, 2009) and manually edited by comparison to the S.cerevisiae Pol II structure (PDB 1WCM) (Armache
et al., 2005).
(A) Schematic depiction of Vaccinia Rpo35 and Rpo7 and the homologous S.cerevisiae Pol II subunits Rpb3, Rpb11 and Rpb10 with domains indicated and
structure-based sequence alignment between the proteins. Insertions and deletions are indicated by connecting lines, with differing regions shown with dashed
lines. Regions with differing fold are indicated by crossed connecting lines. The region resembling the non-conserved domain of Rpb3 responsible for interactions
with Rpb10 and Rpb12 is reduced in Rpo35, with the Zn-binding motif lacking altogether.
(B) Schematic depiction of Vaccinia Rpo22, Rpo19 and Rpo18 and the homologous S.cerevisiae Pol II subunits Rpb5, Rpb6 and Rpb7 with domains indicated and
structure-based sequence alignments. Depiction as in (A). Like Rpb7, Rpo18 binds to the polymerase core via its K1 helical turn and its tip loop in the amino
terminal tip domain. These elements form a wedge between the N-terminal region of Rpo147, the switch 5 region, the Rpo132 anchor, and helix a1 of Rpo19, all of
which are conserved between Vaccinia and Pol II. The Rpo18 tip domain may therefore restrict movement of the clamp, as proposed for Rpb7 in Pol II (Armache
et al., 2003). The C-terminal domain of Rpo19 forms a b-barrel-like structure but appears tilted toward the polymerase body compared to Rpb4/7.
(legend on next page)
Figure S6. Structure Determination of Complete vRNAP, Related to Figure 4
(A) Exemplary cryo-EM micrograph of the complete vRNAP complex dataset.
(B) Selected Class averages from unsupervised 2D classification in Relion.
(C) Cryo-EM processing workflow for structure determination.
(D) Local resolution estimates mapped to the cryo EM density isosurface representation.
(E) Angular particle orientation map.
(F) Fourier Shell Correlation (FSC)-plot.
Figure S7. Sequence Alignment of Rpo30 and S. cerevisiae TFIIS and Structural Details of NPH-I and E11, Related to Figures 4, 5, and 6
(A) Structure-based sequence alignment of Rpo30 and S. cerevisiae TFIIS with secondary structure elements depicted and colored according to domains as in
Figure 6. Sheet regions are shown as arrows, helical region as cylinders. Invariant residues are colored in dark blue and conserved residues in light blue. Regions
differing in fold are colored in green (vRNAP-specific) and red (Pol II-specific). The alignment was generated with MSAProbs (Liu et al., 2010) within the MPI
Bioinformatics Toolkit (Zimmermann et al., 2018), visualized using Aline (Bond and Schüttelkopf, 2009) and manually edited by comparison to the S. cerevisiae

(legend continued on next page)

Pol II structure (PDB 1WCM) (Armache et al., 2005). The Zink-binding regions are highlighted in pink and the conserved acidic residues of TFIIS that enter the Pol II
active site (DEP motif) are highlighted in green.
(B) Fold and topology of the E11 crystal structure. Topology (left). Fold and secondary structure elements in cartoon style (right). The two protomers of the
homodimer are in orange and yellow, respectively.
(C) Comparison of the ATPase domains of NPH-I to those of the chromatin remodelers INO80 (PDB 6FHS) (Eustermann et al., 2018) and SNF2 (from PDB ID 5X0X)
(Liu et al., 2017). The characteristic structural elements are color-coded and labeled.