1. Structural Framework for the Mechanism of Archaeal Exosomes in RNA Processing 
Katharina Büttner, Katja Wenig, Karl-Peter Hopfner 
Molecular Cell 
Volume 20, Issue 3, Pages (November 2005)
DOI: /j.molcel
Copyright © 2005 Elsevier Inc. Terms and Conditions

2. Figure 1 Architecture of the Archaeal Exosome
Ribbon models of the nine subunit archaeal Rrp4- ([A] top view; [B] side view) and Csl4-exosomes ([C] top view). Both exosomes contain a ring of alternating Rrp41 (blue) and Rrp42 (green) subunits. Three tungstate moieties (magenta and red cpk model) indicate three active sites in the Rrp41:Rrp42 ring. Three Rrp4 subunits (orange, with annotated NT, S1, and KH domains) or Csl4 subunits (red, with annotated N-terminal (NT), S1, and zinc-ribbon (ZN) domains plus yellow zinc ions) bind to the top face of the Rrp41-Rrp42 ring in both exosome complexes, respectively.
(D) A slice view through the molecular surface of the Cls4-exosome reveals the channel that proceeds from S1 domains (S1 pore) via the central processing chamber to a pore at the RNase-PH domains (PH pore). A 10 Å narrow neck restricts the channel diameter between the S1 pore and the processing chamber.
Molecular Cell  , DOI: ( /j.molcel )
Copyright © 2005 Elsevier Inc. Terms and Conditions

3. Figure 2 Phosphorolytic Active Sites
(A) Ribbon model of the Rrp4-exosome (bottom view) superimposed with anomalous difference density (magenta) derived from tungstate binding experiments. The density patches reveal likely phosphate binding moieties and define the location of the three phosphorolytic active sites in active site pockets of the central chamber.
(B) Stereo plot of a close-up view of the active site pocket in the interface of Rrp41 (blue) and Rrp42 (green). The Rrp4 exosome is shown as a ribbon model and uses the color code of Figure 1. The side chains of notable conserved active side residues are shown as annotated color-coded sticks. Tungsten anomalous difference density is shown with 4σ (yellow) and 8σ (magenta) contour and indicates a possible RNA path from the central channel into the active site pocket. A tungstate moiety (WO4) is modeled into the 8σ contour and may denote the position of the attacking phosphate.
(C) RNase activity and polyadenylation assays. Protein, phosphate, or ADP was added as indicated, and substrate oligo(rA) 30-mer (S) and polyadenylated (PA) or phosphorolytically degraded (P) nucleotide products were resolved by denaturing polyacrylamide gel electrophoresis. The Rrp4-exosome (REx) shows that processive degradation (lanes 2 and 3) and processive polyadenylation (lane 7) of a 30-mer oligo(rA) depended on phosphate (lane 4) or ADP. Both activities are abolished or severely inhibited by the D180A mutation in the active site (lanes 6 and 9) or the R65E mutation in the neck loop (lanes 5 and 8). Controls show that Csl4-exosome (CEx, lane 12) and the Rrp41:Rrp42 complex (41:42, lanes 10 and 11) possess a similar activity, whereas Csl4 (lane 13) and Rrp4 (lane 14) subunits alone do not possess RNase activity in vitro.
Molecular Cell  , DOI: ( /j.molcel )
Copyright © 2005 Elsevier Inc. Terms and Conditions

4. Figure 3 The Rrp4 and Csl4 Structures
(A) Domain organization of afuRrp4, with highlighted and annotated secondary structure. Rrp4 possesses an N-terminal, all-β domain (red), a middle S1 domain (orange), and a C-terminal KH domain (yellow). The KH domain contains the conserved GXXG motif that is implicated in nucleic acid recognition. The N-terminal and S1/KH domain bind to two adjacent Rrp41 (blue) and Rrp42 (green) molecules, thereby stabilizing the exosome assembly.
(B) Detailed view of the interface of the N-terminal domain of Rrp4 (red) and the C-terminal helix of Rrp41 (blue). The backbone is shown as a ribbon, and interface side chains are displayed as color-coded sticks. Notable residues that are discussed in the text are annotated.
(C) Local mobility of the S1 domain revealed by superposition of the three Rrp41:Rrp42:Rrp4 trimers in the structurally asymmetric Rrp4-exosome. The three Rrp4 subunits are shown in red, orange, and yellow, whereas Rrp41 and Rrp42 are shown in blue and green, respectively. The S1 domains can at least move by 5–6 Å (double-headed arrow), a distance that could allow threading of RNA one base at a time.
(D) Structure-based sequence alignment of Archaeoglobus fulgidus (af), Sulfolobus solfataricus (so), Saccharomyces cerevisiae (y), and human (h) Rrp4 or Rrp40. Conservation is indicated by shading and the darker color represents higher conservation. The secondary structure is shown on top of the alignment with the annotation and color code of (A). The GXXG motif is indicated below the alignment.
(E) Domain organization of afuCsl4 with highlighted and annotated secondary structure is shown. Cls4 possesses an N-terminal, all-β domain (red), a middle S1 domain (orange), and a C-terminal zinc-ribbon domain (yellow, with magenta zinc ion and zinc anomalous difference density). The location of the Ski4-1 mutation in the interface of S1 and zinc-ribbon domains is indicated by a black triangle. The location of N-terminal and S1 domains is similar to those of Rrp4, but the zinc-ribbon domain adopts a unique position.
(F) Detailed view of the interface of the N-terminal domain of Cls4 (red) and the C-terminal helix of Rrp41 (blue), similar to (B), is shown.
(G) Structure-based sequence alignment of af, so, y, and h Cls4 with highlighted shading and secondary structures (see [D]). The location of the Ski4-1 Gly→Glu mutation is indicated below the alignment.
Molecular Cell  , DOI: ( /j.molcel )
Copyright © 2005 Elsevier Inc. Terms and Conditions

5. Figure 4 Electrostatic Surface and Proposed RNA Binding Sites
Molecular surfaces of Rrp4- and Csl4 exosomes in different views, colored according to electrostatic potential (red, 7 kt/e− to blue +7 kT/e−). Rrp4- ([A], top view) and Csl4-exosome ([B], top view) contain strong positive patches at the S1 and KH domains. In contrast, the exterior of the RNase-PH domain ring is highly negatively charged ([C], bottom view of the Csl4 exosome). The strongest positive potential reaches from the S1 pore via the neck loops into the processing chamber ([D]: slice view of the Rrp4 exosome), indicating a likely path for RNA. The location of the functionally important neck residue R65rrp41 is indicated by (∗). The positive patches on the S1 and KH domains match the RNA binding sites of related OB fold and KH domains. RNA derived from a superposition of exosome S1 and KH domains with nucleic acid complexes of OB fold or KH domains are shown as a yellow tube (Figure S3)
Molecular Cell  , DOI: ( /j.molcel )
Copyright © 2005 Elsevier Inc. Terms and Conditions

6. Figure 5 3′→5′ Processing Complexes in All Kingdoms of Life
(A) A model for a heterotrimeric Csl4-Rrp4-Rrp4 cap on the molecular surface of the RNase-PH domain (gray). Placement of a Csl4 subunit (red ribbon model) into an Rrp4 cap (orange ribbon models, the third subunit is substituted by Csl4). The placement was guided by superimposing the RNase-PH domains of Rrp4- and Csl4-exosomes in silico. The model indicates that Csl4 and Rrp4 type subunits can form heterotrimeric caps without evident steric clashes, consistent with our biochemical observations. Such a model is a good approximation of the eukaryotic exosome Csl4-Rrp4-Rrp40 cap.
(B) Evolutionary conservation of RNA degradation in all kingdoms of life. Ribbon representations of the Rrp4 exosome complex (only three out of nine subunits are shown) and bacterial polynucleotide phosphorylase (PNPase, only one out of three subunits is shown). The comparison reveals similar overall arrangement of the catalytic RNase-PH domain dimer (blue and green) and the phosphorolytic active sites (with magenta tungstate moieties). S1 (orange) and KH (yellow) domains are located at related but not identical positions with respect to the active sites. However, S1 and KH domains of PNPase are incompletely resolved and may be mobile. Whereas the precise mode of RNA recognition between PNPase and exosome probably differs to some extent, the overall mechanistic principle of RNA degradation, inferred from the domain arrangements, appears remarkably conserved.
Molecular Cell  , DOI: ( /j.molcel )
Copyright © 2005 Elsevier Inc. Terms and Conditions

7. Figure 6 Proposed Mechanism for Core Exosomes
(A) Functional analysis of the neck. Shown is a slice through the Rrp4-exosome (in color-coded ribbon representation) with highlighted neck side chains and tungstate difference density (yellow mesh). Tungstate density and the R65rrp41E mutation (annotated) indicate that the neck region is important for RNA processing and presumably relays RNA from the S1 domains to the active sites (AS).
(B) Gel shift assays of the wild-type and R65E containing Rrp4-Exosome (REx) and Rrp41:Rrp42 ring (41:42) is shown. R65E changes the shift pattern of the Rrp4-exosome (lanes 2–4 and 5–7) but abolishes RNA binding by the Rrp41:Rrp42 ring (lanes 8–10 and 11–13). The data are consistent with a model in which RNA binds to S1 and/or KH domains and enters the processing chamber in the Rrp41:Rrp42 ring via the neck containing R65.
(C) Proposed model for RNA degradation and trimming by exosomes. Our structural and functional results support a model in which single-stranded RNA 3′ tails reach the active sites (AS) via the S1 pore and neck. For simplicity, only two out of three active sites and S1 domains are shown. The neck and S1 pores restrict processive degradation to fully unfolded RNA substrates. Secondary structures can stall at the S1 pore and abort processive degradation. Accessory factors with helicase activity could stimulate processive degradation in part by providing unfolded, protein-deprived RNA. Sufficiently stable secondary structures of protein:RNA complexes are only trimmed, leaving a tail of about eight nucleotides that span the distance between the active sites and the S1 domains.
Molecular Cell  , DOI: ( /j.molcel )
Copyright © 2005 Elsevier Inc. Terms and Conditions