Thermoacidophilic crenarchaea of the genus Sulfolobus contain six AAA (ATPase associated with various cellular activities) proteins, including a proteasome-associated ATPase, a Vps4 (vacuolar protein sorting 4) homologue, and two Cdc48 (cell-division cycle 48)-like proteins. The last two AAA proteins are deeply branching divergent members of this family without close relatives outside the Sulfolobales. Both proteins have two nucleotide-binding domains and, unlike other members of the family, they seem to lack folded N-terminal domains. Instead, they contain N-terminal extensions of approx. 50 residues, which are predicted to be unstructured, except for a single transmembrane helix. We have analysed the two proteins, MBA (membrane-bound AAA) 1 and MBA2, by computational and experimental means. They appear to be monophyletic and to share a common ancestor with the Cdc48 clade. Both are membrane-bound and active as nucleotidases upon heterologous expression in Escherichia coli. They form ring complexes, which are stable after solubilization in a mild detergent and whose formation is dependent on the presence of the N-terminal extensions.

Background

AAA (ATPase associated with various cellular activities) proteins represent a large and heterogeneous family, whose common activity appears to be the unfolding of proteins and disassembly of protein complexes. Structurally, AAA proteins are characterized by an N-terminal domain, which is thought to mediate substrate recognition, followed by one or two extended ATPase domains. Although some AAA proteins have been purified in dimeric form, it is thought that, in all cases, the active form involves assembly to a hexameric ring. Proteins of the Cdc48 (cell-division cycle 48) clade represent the main AAA lineage with two tandem nucleotidase domains (named D1 and D2), which form separate rings and give the particle the appearance of a stacked doughnut. Whereas the ATPase domains exhibit a high degree of sequence conservation in all AAA proteins, the N-terminal domains are divergent, clade-specific and belong to several unrelated folds [18].

In a large-scale phylogenetic analysis we found that, in addition to the five large well-described clades of AAA proteins, the family contained one further large clade and multiple minor ones, most of which were branching close to the root. Several proteins proved impossible to assign to any branch; among these ‘orphans’ were two proteins from the thermoacidophilic crenarchaeon Sulfolobus solfataricus, KEGG accession numbers Sso2420 and Sso2831, and their homologues [8]. In an effort to understand their origin and cellular function, we have initiated the biochemical and structural characterization of these two atypical ATPases.

Bioinformatic analyses

Sso2420 and Sso2831 are proteins of approx. 600 residues, with two canonical AAA domains (Figure 1). N-terminally, they contain extensions of approx. 50 residues which are predicted to be unstructured, except for a single transmembrane segment per protein. We have therefore named them MBA1 and MBA2 respectively, for membrane-bound ATPase 1 and 2. The absence of a folded N-domain appears to be unique among AAA proteins.

Multiple sequence alignment of the S. solfataricus MBA1 and MBA2 proteins and their homologues

Figure 1
Multiple sequence alignment of the S. solfataricus MBA1 and MBA2 proteins and their homologues

The predicted secondary structures of MBA1 and MBA2 were computed on the Quick2D server [20]; α-helices and β-strands are labelled as coils and arrows respectively. AAA D1 and D2 domain boundaries are indicated above and below the sequences. The asterisk marks mutations generated in both D1 and D2 Walker A motifs of MBA1ΔN. Shaded areas indicate amino acid residues shared between all sequences: black shading indicates highly conserved identical residues, whereas similar residues are depicted with grey shading. KEGG accession numbers: S. solfataricus MBA1 Sso2420, and MBA2 Sso2831; S. acidocaldarius MBA1 Saci_0292, and MBA2 Saci_0877; S. tokodaii MBA1 ST0548, and MBA2 ST1711; Metallosphaera sedula MBA1 Msed_0634.

Figure 1
Multiple sequence alignment of the S. solfataricus MBA1 and MBA2 proteins and their homologues

The predicted secondary structures of MBA1 and MBA2 were computed on the Quick2D server [20]; α-helices and β-strands are labelled as coils and arrows respectively. AAA D1 and D2 domain boundaries are indicated above and below the sequences. The asterisk marks mutations generated in both D1 and D2 Walker A motifs of MBA1ΔN. Shaded areas indicate amino acid residues shared between all sequences: black shading indicates highly conserved identical residues, whereas similar residues are depicted with grey shading. KEGG accession numbers: S. solfataricus MBA1 Sso2420, and MBA2 Sso2831; S. acidocaldarius MBA1 Saci_0292, and MBA2 Saci_0877; S. tokodaii MBA1 ST0548, and MBA2 ST1711; Metallosphaera sedula MBA1 Msed_0634.

Searches for homologues uncovered close relatives of MBA1 and MBA2 in Sulfolobus acidocaldarius and Sulfolobus tokodaii, and an MBA1 homologue in a closely related genus, Metallosphaera. This phylogenetic spectrum is by far the most restrictive that we have yet encountered for any AAA protein. All MBAs show extensive sequence conservation in the ATPase domains and substantial divergence in the N-terminal extensions, to the point where the extensions of MBA1 and MBA2 can only be matched via their transmembrane segments (Figure 1).

In S. solfataricus, MBA1 and MBA2 are encoded in putative operons, MBA1 together with an amine oxidase (KEGG accession number Sso2422) and MBA2 with a highly conserved mechanosensitive channel (KEGG accession number Sso2829) and a transmembrane protein unique to S. solfataricus (KEGG accession number Sso2830). This genomic context is not conserved in the other Sulfolobus species, each of which shows a different gene arrangement. This is in contrast with the other AAA proteins of these organisms whose genomic context is preserved throughout.

We re-analysed the phylogenetic position of MBAs in the context of the AAA family by cluster analysis [9] and Bayesian phylogeny [10]. By focusing on these proteins, we hoped to improve on our earlier analysis, which was broadly aimed at evaluating phylogenetic relationships throughout the family [8]. Cluster analysis consistently recovered MBA D1 and D2 domains separately, at the periphery of the main Cdc48 cluster. Their positions did not, however, allow us to match them specifically with the D1 or D2 clade of Cdc48 (Figure 2). Whereas the D1 domains remained separate, the D2 domains reproducibly converged on to the same part of the map. In contrast, phylogenetic analysis with the program MrBayes [10] recovered D1 domains of MBAs and Cdc48 as a monophyletic clade, but the branching order of the D2 domains was more complicated and did not give a clear assignment. Taken together, these results are best compatible with the assumption that MBAs are monophyletic and originated from the Cdc48 lineage after duplication of the AAA domain.

Cluster map of the core AAA family

Figure 2
Cluster map of the core AAA family

In this analysis, we have omitted the deep-branching clades of the tree presented in [8]. The map was obtained in CLANS [9] with attraction and repulsion values set to 10, and a dampening of 1. The clusters were reproducible under a wide range of other settings we tested, establishing their significance, and reproduced well the map reported previously. For the present Figure, the map was computed in three-dimensions, collapsed to two-dimensions and re-equilibrated with the same settings. In addition to the groups that we described in [8], a new group has emerged containing ORFs from recently sequenced environmental bacteria (bacTRAP), as well as a small group of four Wolbachia sequences which appeared to be devoid of any close relatives. AMA, Archaeoglobus fulgidus and methanogenic archaea; ARC, AAA forming ring-shaped complexes; NSF, N-ethylmaleimide-sensitive factor.

Figure 2
Cluster map of the core AAA family

In this analysis, we have omitted the deep-branching clades of the tree presented in [8]. The map was obtained in CLANS [9] with attraction and repulsion values set to 10, and a dampening of 1. The clusters were reproducible under a wide range of other settings we tested, establishing their significance, and reproduced well the map reported previously. For the present Figure, the map was computed in three-dimensions, collapsed to two-dimensions and re-equilibrated with the same settings. In addition to the groups that we described in [8], a new group has emerged containing ORFs from recently sequenced environmental bacteria (bacTRAP), as well as a small group of four Wolbachia sequences which appeared to be devoid of any close relatives. AMA, Archaeoglobus fulgidus and methanogenic archaea; ARC, AAA forming ring-shaped complexes; NSF, N-ethylmaleimide-sensitive factor.

Biochemical analysis

We undertook the biochemical characterization of MBA1 following heterologous expression in Escherichia coli with a C-terminal His6-tag. The protein was expressed in high quantities and localized to the E. coli cell membrane. Upon solubilization with the detergent DDM (n-dodecyl-β-D-maltoside), we purified the recombinant protein of 69.2 kDa, which, using CD spectroscopy, appeared to be folded and formed a mixture of dimers and higher oligomers in solution. The apparent molecular mass of the oligomers, determined using size-exclusion chromatography, was in the range 440–480 kDa, which corresponds to either hexameric or heptameric complexes. Cryo-electron microscopy corroborated the presence of ring-shaped particles and single-particle reconstruction suggested heptameric complexes, without achieving the level of significance needed to establish this point unambiguously (Figure 3). A mutant lacking the N-terminal transmembrane helix (residues 1–27), MBA1ΔN, was expressed under the same conditions and was recovered from the soluble fraction upon cell lysis. MBA1ΔN was folded according to CD spectroscopy and migrated as a dimer in analytical gel-sizing columns.

Cryo-electron micrograph of wild-type MBA1 in solution

Figure 3
Cryo-electron micrograph of wild-type MBA1 in solution

Arrows indicated ring-shaped particles. Single particle reconstruction suggests the presence of putative heptameric oligomers (lower panel).

Figure 3
Cryo-electron micrograph of wild-type MBA1 in solution

Arrows indicated ring-shaped particles. Single particle reconstruction suggests the presence of putative heptameric oligomers (lower panel).

MBA1 showed ATPase activity at temperatures above 65°C as judged by Malachite Green assays [11]. We were unable to detect any activity at room temperature or 37°C. MBA1ΔN was active under the same conditions, albeit to only about a third of wild-type rates. Because of its easier malleability, we used this form to characterize further the enzymatic activity. MBA1ΔN had a temperature optimum of 75°C (Figure 4A), which corresponds to the optimal growth temperature of S. solfataricus; no activity was observed beyond 85°C. The pH optimum was 7.5 (Figure 4B), one unit higher than the cytoplasmic pH of about 6.5 [12]. MBA1ΔN was most active in the presence of Mg2+ cations; Co2+ and Mn2+ ions gave 70 and 40% of the activity measured with Mg2+ ions respectively.

Biochemical characteristics of the ATPase activity of MBA1ΔN

Figure 4
Biochemical characteristics of the ATPase activity of MBA1ΔN

(A) Temperature-dependence of the activity. (B) pH-dependence of the activity. Results are means of three independent measurements.

Figure 4
Biochemical characteristics of the ATPase activity of MBA1ΔN

(A) Temperature-dependence of the activity. (B) pH-dependence of the activity. Results are means of three independent measurements.

In order to estimate the relative contribution of the two ATPase domains to the overall activity, we generated MBA1ΔN mutants lacking the lysine of the Walker A motif in D1 (MBA1ΔN-K108A) or D2 (MBA1ΔN-K391A). This residue is essential in co-ordinating the bound nucleotide, and domains lacking this residue are therefore catalytically inactive. Mutant K108A had only approx. 65% of the activity shown by MBA1ΔN, whereas K391A had over 80%, indicating that the bulk of the activity is due to the first AAA domain (Figure 5).

Comparison of the relative ATPase activity of wild-type MBA1, wild-type MBA2, MBA1ΔN and two Walker A mutants thereof

Figure 5
Comparison of the relative ATPase activity of wild-type MBA1, wild-type MBA2, MBA1ΔN and two Walker A mutants thereof

The experiments were conducted in triplicate at optimum temperature (75°C) and pH (7.5). The specific activity of MBA1 was 2.8 μmol ·mg−1 · h−1 in the presence of 1 mM ATP and 5 mM Mg2+, which is in the range of activities observed previously for VAT [VCP (valosin-containing protein)-like ATPase from Thermoplasma acidophilum] [21] or p97 [22].

Figure 5
Comparison of the relative ATPase activity of wild-type MBA1, wild-type MBA2, MBA1ΔN and two Walker A mutants thereof

The experiments were conducted in triplicate at optimum temperature (75°C) and pH (7.5). The specific activity of MBA1 was 2.8 μmol ·mg−1 · h−1 in the presence of 1 mM ATP and 5 mM Mg2+, which is in the range of activities observed previously for VAT [VCP (valosin-containing protein)-like ATPase from Thermoplasma acidophilum] [21] or p97 [22].

We also expressed the recombinant MBA2 protein with a C-terminal His6-tag in E. coli. The full-length form was recovered from the membrane fraction and the N-terminally deleted form lacking the transmembrane sequence, MBA2ΔN, from the cytosolic fraction, but part of the latter protein appeared to be membrane-associated; both proteins were folded according to CD spectroscopy. MBA2 was enzymatically active, but showed significantly less activity than MBA1 under the same conditions (Figure 5). Size-exclusion chromatography analysis also suggested the formation of larger oligomers by the full-length protein under certain conditions, but we are lacking an electron microscopy analysis at this time.

Conclusions

MBA1 (Sso2420) and MBA2 (Sso2831) of S. solfataricus are novel members of the AAA family, with some unusual properties. They are difficult to assign to any AAA clade with any degree of confidence based on their sequences. They occur only in Sulfolobus and Metallosphaera, showing the most restricted phylogenetic spectrum yet observed for an AAA protein. They lack a folded N-domain; instead, they have short N-terminal extensions containing a transmembrane sequence. Their genomic context is highly diverse even between different species of the same genus. These properties can be either interpreted as signs of an extreme specialization towards a specific lineage, or as ancestral features of ‘living fossils’ which have been displaced by more modern AAA forms almost everywhere else.

In considering their evolutionary origin, we find two observations particularly relevant: (i) the fact that their D1 and D2 domains cluster into separate groups suggests that they had a common ancestor which already contained two differentiated AAA domains; and (ii) the fact that they form satellite clusters of Cdc48 and, furthermore, that at least their D1 domains consistently group with the D1 domains of Cdc48 in phylogenetic analyses suggests that this common ancestor was originally a member of the Cdc48 lineage.

Unlike Cdc48, however, at least MBA1 might form heptameric complexes, whose formation is dependent on the N-terminal extensions. These properties would again be unusual, since oligomerization in the Cdc48 clade is driven by the ATPase rings and since heptamerization has only been reported in a few members of the larger AAA+ superfamily, but not in AAA proteins [7,13] {we note that the recently reported heptameric oligomerization of a truncated Vps4 (vacuolar protein sorting 4) protein [14] is contradicted by studies of the full-length protein [5,15]}. Establishing the oligomeric state of MBA1 unambiguously is a priority for further studies. Towards this end, we are pursuing crystallographic studies of both MBA1 and MBA2. In addition, we are attempting to characterize the biological role of these proteins via homologous expression [16,17] and the generation of genetic knockouts [18,19]. The expression of His10-strep-tagged proteins in S. solfataricus cells [16] will hopefully allow us to isolate and identify binding partners and will help to clarify the cellular localization of the proteins at various stages of the cell cycle.

Funding

This work was supported by institutional funds from the Max Planck Society.

Molecular Biology of Archaea: Biochemical Society Focused Meeting held at University of St Andrews, U.K., 19–21 August 2008. Organized and Edited by Stephen Bell (Oxford, U.K.) and Malcolm White (St Andrews, U.K.).

Abbreviations

     
  • AAA

    ATPase associated with various cellular activities

  •  
  • Cdc48

    cell-division cycle 48

  •  
  • MBA

    membrane-bound AAA

  •  
  • ORF

    open reading frame

References

References
1
Lupas
A.N.
Martin
J.
AAA proteins
Curr. Opin. Struct. Biol.
2002
, vol. 
12
 (pg. 
746
-
753
)
2
Hanson
P.I.
Whiteheart
S.W.
AAA+ proteins: have engine, will work
Nat. Rev. Mol. Cell Biol.
2005
, vol. 
6
 (pg. 
519
-
529
)
3
Hartman
J.J.
Vale
R.D.
Microtubule disassembly by ATP-dependent oligomerization of the AAA enzyme katanin
Science
1999
, vol. 
286
 (pg. 
782
-
785
)
4
Scott
A.
Chung
H.Y.
Gonciarz-Swiatek
M.
Hill
G.C.
Whitby
F.G.
Gaspar
J.
Holton
J.M.
Viswanathan
R.
Ghaffarian
S.
Hill
C.P.
, et al. 
Structural and mechanistic studies of VPS4 proteins
EMBO J.
2005
, vol. 
24
 (pg. 
3658
-
3669
)
5
Yu
Z.
Gonciarz
M.D.
Sundquist
W.I.
Hill
C.P.
Jensen
G.J.
Cryo-EM structure of dodecameric Vps4p and its 2:1 complex with Vta1p
J. Mol. Biol.
2008
, vol. 
377
 (pg. 
364
-
377
)
6
Ogura
T.
Wilkinson
A.J.
AAA+ superfamily ATPases: common structure–diverse function
Genes Cells
2001
, vol. 
6
 (pg. 
575
-
597
)
7
Diemand
A.V.
Lupas
A.N.
Modeling AAA+ ring complexes from monomeric structures
J. Struct. Biol.
2006
, vol. 
156
 (pg. 
230
-
243
)
8
Frickey
T.
Lupas
A.N.
Phylogenetic analysis of AAA proteins
J. Struct. Biol.
2004
, vol. 
146
 (pg. 
2
-
10
)
9
Frickey
T.
Lupas
A.
CLANS: a Java application for visualizing protein families based on pairwise similarity
Bioinformatics
2004
, vol. 
20
 (pg. 
3702
-
3704
)
10
Ronquist
F.
Huelsenbeck
J.P.
MrBayes 3: Bayesian phylogenetic inference under mixed models
Bioinformatics
2003
, vol. 
19
 (pg. 
1572
-
1574
)
11
Geladopoulos
T.P.
Sotiroudis
T.G.
Evangelopoulos
A.E.
A malachite green colorimetric assay for protein phosphatase activity
Anal. Biochem.
1991
, vol. 
192
 (pg. 
112
-
116
)
12
She
Q.
Singh
R.K.
Confalonieri
F.
Zivanovic
Y.
Allard
G.
Awayez
M.J.
Chan-Weiher
C.C.
Clausen
I.G.
Curtis
B.A.
De Moors
A.
, et al. 
The complete genome of the crenarchaeon Sulfolobus solfataricus P2
Proc. Natl. Acad. Sci. U.S.A.
2001
, vol. 
98
 (pg. 
7835
-
7840
)
13
Costa
A.
Pape
T.
van Heel
M.
Brick
P.
Patwardhan
A.
Onesti
S.
Structural studies of the archaeal MCM complex in different functional states
J. Struct. Biol.
2006
, vol. 
156
 (pg. 
210
-
219
)
14
Hartmann
C.
Chami
M.
Zachariae
U.
de Groot
B.L.
Engel
A.
Grütter
M.G.
Vacuolar protein sorting: two different functional states of the AAA-ATPase Vps4p
J. Mol. Biol.
2008
, vol. 
377
 (pg. 
352
-
363
)
15
Xiao
J.
Xia
H.
Yoshino-Koh
K.
Zhou
J.
Xu
Z.
Structural characterization of the ATPase reaction cycle of endosomal AAA protein Vps4
J. Mol. Biol.
2007
, vol. 
374
 (pg. 
655
-
670
)
16
Albers
S.V.
Jonuscheit
M.
Dinkelaker
S.
Urich
T.
Kletzin
A.
Tampe
R.
Driessen
A.J.
Schleper
C.
Production of recombinant and tagged proteins in the hyperthermophilic archaeon Sulfolobus solfataricus
Appl. Environ. Microbiol.
2006
, vol. 
72
 (pg. 
102
-
111
)
17
Hobel
C.F.
Albers
S.V.
Driessen
A.J.
Lupas
A.N.
The Sulfolobus solfataricus AAA protein Sso0909, a homologue of the eukaryotic ESCRT Vps4 ATPase
Biochem. Soc. Trans.
2008
, vol. 
36
 (pg. 
94
-
98
)
18
Albers
S.V.
Driessen
A.J.
Conditions for gene disruption by homologous recombination of exogenous DNA into the Sulfolobus solfataricus genome
Archaea
2007
, vol. 
2
 (pg. 
i
-
v
)
19
Worthington
P.
Hoang
V.
Perez-Pomares
F.
Blum
P.
Targeted disruption of the α-amylase gene in the hyperthermophilic archaeon Sulfolobus solfataricus
J. Bacteriol.
2003
, vol. 
185
 (pg. 
482
-
488
)
20
Biegert
A.
Mayer
C.
Remmert
M.
Söding
J.
Lupas
A.
The MPI Toolkit for protein sequence analysis
Nucleic Acids Res.
2006
, vol. 
34
 (pg. 
335
-
339
)
21
Pamnani
V.
Tamura
T.
Lupas
A.
Peters
J.
Cejka
Z.
Ashraf
W.
Baumeister
W.
Cloning, sequencing and expression of VAT, a CDC48/p97 ATPase homologue from the archaeon Thermoplasma acidophilum
FEBS Lett.
1997
, vol. 
404
 (pg. 
263
-
268
)
22
Peters
J.M.
Walsh
M.J.
Franke
W.W.
An abundant and ubiquitous homo-oligomeric ring-shaped ATPase particle related to the putative vesicle fusion proteins Sec18p and NSF
EMBO J.
1990
, vol. 
9
 (pg. 
1757
-
1767
)