The crystal structure of the conserved GTPase of SRP54 from the archaeon Acidianus ambivalens and its comparison with related structures suggests a model for the SRP–SRP receptor complex

Research Article

515

The crystal structure of the conserved GTPase of SRP54 from the archaeon Acidianus ambivalens and its comparison with related structures suggests a model for the SRP–SRP receptor complex Guillermo Montoya1, Kai te Kaat1, Ralph Moll2, Günter Schäfer2 and Irmgard Sinning1,3* Background: Protein targeting to the endoplasmic reticulum in eukaryotes and to the cell membrane in prokaryotes is mediated by the signal recognition particle (SRP) and its receptor (SR). Both contain conserved GTPase domains in the signal-peptide-binding proteins (SRP54 and Ffh) and the SR proteins (SRα and FtsY). These GTPases are involved in the regulation of protein targeting. Most studies so far have focussed on the SRP machinery of mammals and bacteria, leaving the SRP system of archaea less well understood.

Addresses: 1European Molecular Biology Laboratory, Structural Biology Programme, Postfach 10.2209, D-69017 Heidelberg, Germany, 2Institute of Biochemistry, Medical University of Lübeck, Ratzeburger Allee 160, D23538 Lübeck, Germany and 3BZH, University Heidelberg, INF 328, D-69120 Heidelberg, Germany.

Results: We report the crystal structure of the conserved GTPase (NG-Ffh) from the thermophilic archaeon Acidianus ambivalens at 2.0 Å resolution and of the Thr112→Ala mutant, which is inactive in GTP hydrolysis. This is the first structure of an SRP component from an archaeon and allows for a detailed comparison with related structures from Escherichia coli and thermophilic bacteria. In particular, differences in the conserved consensus regions for nucleotide binding and the subdomain interfaces are observed, which provide information about the regulation of the GTPase. These interactions allow us to propose a common signalling mechanism for the SRP–SR system.

Key words: archaea, GTPase, protein targeting, SRP, X-ray diffraction

*Corresponding author. E-mail: [email protected]

Received: 28 January 2000 Accepted: 14 March 2000 Published: 2 May 2000 Structure 2000, 8:515–525 0969-2126/00/$ – see front matter © 2000 Elsevier Science Ltd. All rights reserved.

Conclusions: The overall structure of SRP-GTPases is well conserved between bacteria and archaea, which indicates strong similarities in the regulation of the SRP-targeting pathway. Surprisingly, structure comparisons identified a homodimeric ATP-binding protein as the closest relative. A heterodimer model for the SRP–SR interaction is presented.

Introduction Protein targeting is essential for all living cells. The signal recognition particle (SRP) and its receptor (SR) are part of a well conserved machinery that is employed in protein transport to the endoplasmic reticulum (ER) in eukaryotes and to the plasma membrane in prokaryotes. SRP is a stable cytosolic ribonucleoprotein complex that binds to the signal peptide of a polypeptide chain emerging at the ribosome. The subsequent interaction with the SR is necessary to address the ribosome nascent chain complex (RNC) to the translocation machinery in the lipid bilayer (reviewed in [1,2]). The key components of protein targeting to the ER in eukaryotes and to the plasma membrane in prokaryotes are highly conserved. The SRP pathway of bacteria consists of a single protein named Ffh (Fifty four homologue), a 4.5S RNA and the docking protein FtsY [3,4]. The individual components share sequence homology with their eukaryotic counterparts as well as functional similarities [3,5]. Recently it has been shown that the SRP pathway in Escherichia coli is

essential for the insertion of a number of polytopic membrane proteins, whereas a number of pre-proteins can be targeted in an SRP-independent pathway via SecB [6–9]. During targeting Ffh/4.5S RNA and FtsY form a complex in a GTP-dependent manner, followed by the reciprocal stimulation of the GTPases in both Ffh and FtsY [10–12]. Ffh, FtsY and their respective homologues in eukaryotes are multidomain proteins and contain a conserved GTPase (see Figure 1a). The so-called N and G domains form the common core, flanked by either an N-terminal charged domain in the receptor proteins, which is likely to mediate membrane attachment [13] or a C-terminal M domain in Ffh, which binds the SRP RNA and the signal peptide [14,15]. Structural information is available for the conserved GTPase domains of FtsY (the SRα homologue) from E. coli [16] and the apo form as well as the GDP and GDP–Mg2+ complexes of Ffh from the thermophilic bacterium Thermus aquaticus [17,18]. The structure of the M domain from mammalian SRP54 [19] and of T. aquaticus has also been determined [20].

516

Structure 2000, Vol 8 No 5

Figure 1 (a)

Charged domain

N

G 12

NH 2

NH 2

3 4 5 COOH

FtsY

COOH

SRα

NH 2

COOH Ff h

NH 2

COOH SRP54

NH 2

Structure of A. ambivalens Ffh. a) Domain structure of the SRP-GTPase family. Within the G domain the five consensus elements are depicted in yellow and the I-box in light green. (b) Cα ribbon representation of the NG domain of Ffh from A. ambivalens drawn with MOLSCRIPT [49] and Raster3D [50]. The three segments are shown in red (N domain), blue (G domain) and green (I-box). Helices are labelled alphabetically in capital letters (A–D) in the N domain. Helices and strands in the G domain are labelled with numbers according to their appearance in the sequence. (c) Stereo diagram of a Cα representation of the core domain of Ffh (blue). The N and C termini (3 and 297, respectively) are close in space. (d) Stereoview of a (2|Fo|–|Fc|) sigmaaweighted electron-density map at 2.0 Å contoured at 1.0σ showing the interaction of the conserved loop between the B and C-helices of the N domain with residues in the N cap of α5.

M

COOH

p21Ras

(b)

(c)

(d)

Structure

Research Article Crystal structure of A. ambivalens Ffh Montoya et al.

517

Table 1 Data collection and phasing statistics. Data set

Native 1

Native 2

T112A mutant

KAu(CN)2

SeMet protein

Number of crystals Source (Å) Temperature (K) Maximum resolution (Å) Measured reflections Unique reflections I/σ Completeness (%) Rsym* (%) No. of sites Rmerge† Rcullis‡ Rkraut§ Combined figure of merit (SHARP) Phasing power# Density modification (Solomon) FOM

1 1.54 100 2.85 111,572 7679 12.1 100 6.6

1 BM14 0.99 100 2.00 451,808 20,842 18.2 100 5.1

1 BM14 0.99 100 1.95 450,721 21,100 15.3 99.6 6.9

1 1.54 100 2.95 117,741 6594 9.1 99.6 8.6 2 8.7 0.56 0.06

1 BM14 0.99 100 2.9 100,599 6867 12.1 98.3 5.1 4 6.3 0.8 0.13

2.0/1.0

0.5/1.4

0.43 ∆PHI 2.8° 0.96

*Rsym = ΣhΣi | Ii(h) – | / ΣhΣi Ii(h), where h is the index for unique reflections, Ii(h) is the ith measurement of reflection h and the mean intensity of all measurements of I(h). †R merge = Σh | Fnat – Fderiv| / Σh |Fnat|, where Fnat and Fderiv are the structurefactor amplitudes of the native and the derivative, respectively. ‡R cullis = Σ ||Fderiv| ± |Fnat||–|FHcalc|/ Σ ||Fderiv| ± |Fnat||, where the

summations are over the centric reflections, FHcalc is the calculated structure factor for heavy atoms. §R kraut = Σ ||F + deriv| – |F + Hcalc|| + ||F – deriv| – |F – Hcalc||/ Σ | F + deriv| + |F – deriv|, where the summations are over the acentric reflections. #Phasing power = FH /E, where FH and E are the rms of the heavy-atom structure factor and lack of closure error, respectively. F > 2σ(F) (all data).

The SRP pathway of archaea is much less well characterised. SRP54 and SRα (FtsY) homologues have been assigned in a number of the recently completed genomic sequences based on sequence similarities. For SRP54 from the thermophilic archaeon Acidianus ambivalens the similarity to the bacterial and mammalian homologues was recently described by the ability of SRP54 to bind to the 7S RNA as well as its GTPase activity [21]. The precise composition of the SRP from archaea, however, is still not known [22].

solved by multiple isomorphous replacement with anomalous scattering (MIRAS) using gold and selenomethionine derivatives (for data collection and phasing statistics see Table 1). A complete model of the engineered fragment comprising amino acids 3–293 could be built. Four Table 2 Refinement statistics.

Structure determination and overall structure

Resolution range (Å) Data: Native 2: No. observed reflections I/σ >0 No. unique reflections F/σ >0 No. unique reflections F/σ >2 Overall completeness (25 Å–2.0 Å) (%) No. reflections in test set (%) No. non-hydrogen atoms of protein No. water sites Rsym (%) Rfree (%) Rcryst (%) rmsd from ideal stereochemistry Bond lengths (Å) Bond angles (°) Dihedral angles (°) Impropers (°) Average B factor (Å2) For all atoms For mainchain

Recombinant A. ambivalens NG-Ffh (residues 1–293 with a His6-tag at the C terminus) was purified from E. coli and crystallised using the hanging drop method in the orthorhombic space group, C2221 [23]. The structure was

Rcryst and Rfree are for data F/σ > 0. Rfree = Σ|Fobs–Fcalc|/ΣFobs for 5% of data randomly selected and excluded from the refinement. Rcryst = Σ|Fobs – Fcalc|/ΣFobs for the remaining 95% of the data included in the refinement.

In this paper we describe the crystal structure of the conserved GTPase of the SRP54 homologue from A. ambivalens (NG-Ffh). This is the first structure of an SRP component from an archaeon and reveals local differences to the previously solved structures of Ffh and FtsY [16–18]. This allows for a detailed comparison of the common core of SRP-GTPases. How SRP and SR form a complex during the targeting process and which regions are involved in this interaction are still open questions. Based on the structural homology of SRP-GTPases with other nucleotide-binding proteins a model for the SRP–SR complex is proposed.

Results and discussion

25–2.0 451,808 20,842 20,460 99.7 5 2333 102 5.1 27.5 21.3 0.011 1.1 22.4 1.0 39.9 35.2

518

Structure 2000, Vol 8 No 5

Data Bank [PDB] entries 1FFH and 3NG1) and NG-FtsY from E. coli (PDB entry 1FTS) [16–18] (Figure 3). All three structures can be superimposed with an overall root mean squared deviation (rmsd) below 2 Å. The G domains of A. ambivalens and T. aquaticus superimpose with an rmsd of 1.5 Å, slightly better than each of them does with E. coli FtsY (rmsd 1.7 Å and 1.6 Å, respectively). Significant differences in the subdomain organisation can be noticed. These are the organisation of the α helices of the N domain and their orientation with respect to the GTPase domain and the conformation and orientation of the I-box with respect to the GTPase domain. The I-box of A. ambivalens Ffh described here superimposes better with the I-box from E. coli FtsY than with the one from T. aquaticus. It should be noticed that the overall rmsd between the two available apo structures of T. aquaticus (3NG1 and 1FFH) is significantly higher than the rmsd between the apo form (3NG1) and the GDP and the GDP–Mg2+-bound structures of the same organism. Taken together, these observations show the high degree of flexibility within the subdomains of the conserved core of SRP-GTPases. The differences observed between the

histidine residues of the His6-tag at the C terminus were well ordered (residues 294–297; refinement statistics are given in Table 2). The structure can be divided into three segments based on sequence alignments and structural data (Figure 1) [16–18]: the N domain (residues 1–84) and a GTPase domain (99–297) with an α–β–α insertion (I-box) [16] comprising residues 138–180. The N domain is formed by an antiparallel four-helix bundle. These helices pack together by hydrophobic interactions that are conserved throughout the SRP-GTPase family. A peptide of 15 residues links the N to the G domain and is tightly packed against the protein surface. The GTP-binding domain shows the canonical α–β fold (known from small GTPases such as p21ras and Ran) with seven out of eight β strands forming a parallel β sheet surrounded by six α helices and loops connecting them. The consensus elements for nucleotide binding are well conserved (Figure 2) [24]. Overall comparison of the conserved core domain structures

The structure described here is overall rather similar to the known structures of NG-Ffh from T. aquaticus (Protein Figure 2

A

* FFH_AA FFH_AQUA FFH_ECOLI SRP54_SACH SRP54_HUMAN SRa_HUMAN SRa_SACH FTSY_AA FTSY_ECOLI

. . . . . D . . .

. . . . . E . . .

. . . . . G . . .

. . . . . A . . .

. . . . . A . . .

. . . . . Q . . A

. . . . . T . . R

. . . . . L . . L

. . . . . T . . K

S . . . . K . . R

K . . M M P . . S

L M M V V S . P L

*L D N L F F L L A . T L

Q D A A T . K K

Q N D D K . E T

L L L L G . K K

R S T G G T . G E

B

D A D K R L . S N

10 A V R L R L R I K I GG . . I G L G

E R Q K K . K I .

90 P K V L P V T L N A F V A WT MD A K D E E E I . . .

R Q S N T M . S S

K E R S S F . K .

F A T A A G . L .

L I L V L M . N G

T G R N . L . I F

G R N N R K . F I

S . . A S G . N S

S L I I L L V I L

20 S . R G S G S N S N V G L G F R F R

Y N A N N Y Y Y F

100 V I M L WF V V L I I M V I M V V T V F S V I M V I L

. R R T A S N Y G

. G G Q T K K R K

. R R D I S T E K

. I L D I L I I I

Y T T F N S N K D

D E E T E R E E .

K E D T E E S D .

A D N S V D D D .

V L V V L M L I .

E K K D N E K D D

T S A S S V V V V

110 GK GK GK GK GK GK GK GK GK

D A D V A S S D D

F T T M M V V I L

I L L L L L L I F

30 K E R E R E K G K E D K E K E E E E

L I V I V M L L L

C

Q R R V C R T R E

β1 FFH_AA FFH_AQUA FFH_ECOLI SRP54_SACH SRP54_HUMAN SRa_HUMAN SRa_SACH FTSY_AA FTSY_ECOLI

S K V C V Q T E E

N E A K K P P K P

. . . . . Q G . .

. . . . . R V . .

. . . . . R S . .

. . . . . V V . .

. . . . . D D . .

. . . . . M L . .

L A A L L L L N L

*.

F L M V V R R . .

. . T . D E . .

. . . C . I I R N

. . . E . . Q F V

G G G G D . S E E

G G E S P . K R .

D E E E G . R N .

K A N E V . S L .

I L L P P Q E K .

P . A K T R G R .

D . A K K R K S G

K K Q R G Q C P K

I D P K K R D K A

P R P T Q P P P P

K R M T T D Q T E

S A A A A H Q E Q

L L L L L L L L L

I M L L L I I L L

S D E E E A T E I

A A A S A K K S A

40 D V D V D V D V D V N V N V D V D V

N N A N N A A S G

V L L I I A P L V

K E P A K D E E E

L V V L L I A V T

V A V V V A A T T

F R R S K V D D R

V P V V T V V V V

130 GL L L L V GL C L L I L I I V ML

α1 L L M F F F I F M

V V A V V C V F V

G G G G G G G G G

V L L L L V V V V

Q Q Q Q Q N N N N

G G G G G G G G G

T T T T T S S T T

T T T T T T T T T

T T S S T N N T T

A A V C C L L I I

G A G T S A S A G

K K K K K K K K K

L L L L L I L V L

A A G A A S A A A

S D E K Q Q Y K K

L F F L L L L I I

50 T N V E I N R N R E C E T Q L E I T

K R R N N S Q D N

I V V I V V V L L

K R K R K A S K T

E E E S S N H S E

R E K Q A K D S G

P P P A A A S A A

140 A A A A A A GA GA GA GA A A A A

FFH_AA FFH_AQUA FFH_ECOLI SRP54_SACH SRP54_HUMAN SRa_HUMAN SRa_SACH FTSY_AA FTSY_ECOLI

. . . . . . R . .

. . . . . . G . .

. . . . . . K . .

. . . . . G T . .

. . . . . G G . .

. . . . . R N . .

. . . . . T D . .

. . . . . M Y . .

V V V I I V V I I

P P D P P Q E P P

V V F F F L L L V

Y L F Y Y F F V I

G E P G G E E K A

E V S S S K A G Q

E G G E E G G G G

K E Q T M K G G A

D S K D D D S D D

A P P P P A D P S

. . . . . . L . .

. . . . . . V . .

V E V A V A T A A

G S D K I G K S S

I I I V I I I V V

A R V A A A A A I

170 K R R R N A E E S E ME K Q F D F D

G V A G G A A A A

V E L I V I I I I

E E K N E A K Q Q

K K R S Q . . K E

. G E . . L L . .

K K K K R E Q K Q

K . H R K N N A Q

G G K G G G N G G

F K R R K K F K WK F S F K L K K S

V V V V I A V S A

β4 K K E K K F Y A A

F A A F F A A A A

L R K K K R R K K

S L L K N N D S A

E E K E E Q Q R R

K L V L D E V I S

N G G S L G G G R

. V V R M M T V L

G A S C C A A A A

. L N S . G A N R

. E K E . T N R D

P T T T A T S D A

T P P T S V V L L

Y A G N G T E D Y

I . . A L S N E G

E . . Q N T T L L

R . . T K V A V L

R . . K R K R K K

70 E WF E V I QE F K L I K MI QA L E S L K S L E E M

150 QQ E K E Q I R T K R R E N QK QR

I V V A A L L L N

G G G R R S A E N

. . . . . A Q . .

. . . . . L L . .

. . . . . H M . .

. . . . . P D . .

. . . . . P D . .

K P S L S L K S . . F S WT S E A E

I L V Q Q Q T K G

K A K K H E K N E

I T I T A S A S I

V V V V V L L I L

Y Y R F F V T M A

D E N D K Q Q E K

E A E E E I I I V

80 L L L L L L L L D

. . . . . H H . .

. . . . . . V . .

. . . . . . R . .

. . . . . . G . .

. . . . . . S . .

. . . . . . K . .

. . . . . . N . .

. . . . . . K . .

Y L A E E V V L V

230 D L S V N T QQ A Q D Q QQ E Q S Q

α2 A A A A A C C S G

D D D D D D D D D

V T V T T T T T T

Y Q Y F F F F F F

R R R R R R R R R

L R I F F V V Q V

E E K D D E E E E

Q Q Q Q Q Q Q Q Q

L L L L L L L L L

Q R E K K R R A Q

Q L T Q Q T V I V

L G L G L A N A N A H T H V H C WG

. . . . . E S . .

. . . . . K T . .

G2

α3 160 P G MD D V Y T Y T GY GY K Y H T

L A A L I L L L A

β2

120 Y F Y L Y Y K F L V Y Y Y Y Y F WL F WL Y ML R QF

G1

β3

D 60 E . K Q H E E N E E K V S K K K K Q

180 K M A R F Y K F N F GF N F GI N I

E D D D E D D D D

I L V I I V I V V

I I L I I V V V L

I L L I I L L L I

V V V V V V M I A

D D D D D D D D D

α4 T T T T T T T T T

A A A S S A A A A

190 GR GR GR GR GR GR GR GR GR

H L L H H M R M L

G . . . . . . . .

Y . . . . . . . .

G Q H H K Q H H Q

E I V Q Q D N V N

E D D E E N D D K

200 L L L M MM L F L F L M L M L T L M

A E E E D A P K S

A P A E S P T D H

D E E H D C T D E

280 E L GL A L D L D F D L D L D L D L

E G D Q E T S E E

E E E E E A P E E

M L I M M L L L L

K A K I L A K K K

N R Q E Q K S R K

I L V I V L F I I

β5 Y K H S A I A V V

E E A N N T D R R

210 A I V L S I V I A I V N QA I S V M

K G N K Q T K K K

. . . . . . . . K

. . . . . . . . L

. . . . . . . . D

. . . . . . . . V

. . . . . . . . E

. . . . . . . . A

P P P P P P P P P

D D V N D D D N H

E E E Q N L K L E

I I I . L L L I L

L L L . L M M F F

G G G . G K S S A

M M M . M A . . R

G G G . G . . . E

D D D . D . . . D

I V V . I . . . .

E A . . . . . . .

. S . . . . . . .

. L . . . . . . .

V V T T I V I K V

T L L I V L I L M

L L F M Y F M L L

V V V V V V V V T

220 I D L D V D L D MD GE GE I D I D

α5 A A A A A A A A A

S M M S S L L L S

I T T I I V V A T

G G G G G G G G G

Q Q Q Q Q N T N Q

K E D A A E D D N

A A A A C A S A A

A A A S A L A A A

S R K K K V K K K

K A A A A K N Y L

F F F F F F F F F

G3

β6 FFH_AA FFH_AQUA FFH_ECOLI SRP54_SACH SRP54_HUMAN SRa_HUMAN SRa_SACH FTSY_AA FTSY_ECOLI

N D N K K N N E H

Q E E E D R D N E

A K A S K A A V A

S V L S V L F V V

240 K . G. P . D . D . A D GK D . G.

. . . . . H . . .

. . . . . S . . .

. . . . . M . . .

. . . . . A . . .

. . . . . Q . . .

. . . . . T . . .

. . . . . P G . .

. . . . . R R . .

. . . . . L N . .

I V L F V I L Y L

G T T G A D D D T

T G G A S G F G G

I L V I V I F V I

I V V I I V I I T

α6

*I T K M* L L L V L I L L

T T T T T S T T

K K K K K K K K

L V M L F C V L

250 D G D G D G D G D G D T D T D A D G

. . . . . I V . .

T D D H H D G D T

A A A A A D E A A

K R R R K K M K K

G G G G G V L G G

G G G G G G G G G

G A A G G A T V V

A A A A A A M V I

*L L L I L I V L F

260 S A S A S I S A S A S M N M S L S V

β7 V R R V V T V A A

A H H A A Y Y Y D

A V I A A I A E Q

T T T T T T T L F

G G G N K S G K G

G4

Sequence alignment of the conserved core domain of SRP54 (human and Sacharomyces cerevisiae), Ffh (E. coli, T. aquaticus and A. ambivalens) and FtsY (E. coli and A. ambivalens) based on the three-dimensional structure of NG-Ffh from A. ambivalens. The start and end of the structure are marked by black squares. Residues

A K K T S K I K I

T P P P P P P P P

I I I I I I I V I

270 K F Y F K F I F I F V F L F I F R Y

β8

*I G T G E K I A L I I V V L I

G G G G G G G G

V V T T T V T V

S G G G G G G G

E E E E Q Q Q E

K K H H T T N R

G5

P T I I Y Y Y I

E E E E E R R I R

V P P K P S T P P

α7

*F N P R R *F F F F F L L F F

Y H S K N S S K

P P P T A V V A

E D K Q K K D D

R L R I S F P F A V WA WF D F

290 V A A G A S I S I S V A V N I E I E

R R R . K A T R A

Structure

involved in the hydrophobic packing close to G4 are labelled with an asterisk. Secondary structure was assigned with DSSP based on the A. ambivalens structure. The GTP-binding consensus elements are included in yellow boxes and the conserved loop between the B and C helices in a green box.

Research Article Crystal structure of A. ambivalens Ffh Montoya et al.

519

Figure 3 Comparison of the conserved core domains of Ffh from A. ambivalens and T. aquaticus with FtsY from E. coli. (a) The NG domains are shown in a ribbon diagram, side by side. Differences in subdomain organization can be seen. Colour code and representation are as in Figure 1b. (b) Superposition of the three structures (based on the G domains) with the lsq routine in the program O [47]. A. ambivalens Ffh (red), E. coli FtsY (green), T. aquaticus Ffh (blue, 3NG1). I-box and switch II region (sII) are marked. G elements are numbered 1, 2, 4 and 5.

NG-Ffh and NG-FtsY structures are, therefore, unlikely to be the consequence of thermal stability as proposed earlier [18], but rather reflect the variability in subdomain orientation and interaction. Structure of the nucleotide-binding site

The structure of the consensus elements (for their sequences see Figure 2) involved in nucleotide binding in A. ambivalens Ffh is similar to the previously published SRP-GTPases with the P-loop, G3, G4 and G5 regions superimposing best. The G1 (or P-loop) shows a network of hydrogen bonds with the G3 region, as in small GTPases. The sidechain of Lys111 interacts with the mainchain oxygen of Thr188. The OH group of Thr112 is forming a weak interaction with the sidechain of Asp187. The structure of the P-loop is further supported by a hydrogen bond between the sidechain of Gln107 and Arg138 in the I-box. The importance of these local interactions became evident from the analysis of the Thr112→Ala mutant, which is no longer able to hydrolyse GTP (KtK, unpublished observations). The crystals of the Thr112Ala mutant are isomorphous with the native crystals and diffract to 2.0 Å [23]. The refined structure is very similar to the wild-type NG-Ffh with an rmsd of 0.24 Å for 291 Cα atoms and an Rcryst of 23% (Rfree 27%). The major difference to the wild-type structure is the absence of interpretable electron density for residues 108–112, which indicates a severe destabilisation

of the P-loop. The loss of the polar hydroxyl group of Thr112 abolishes the interaction with the sidechain of Asp187 and the hydrogen bond between the sidechains of Gln107 and Arg138. In the structure from T. aquaticus Ffh in complex with GDP–Mg2+ the sidechain of Thr112 is involved in coordinating the Mg2+ ion, whereas in the apo form it interacts with Asp187 [18]. A similar interaction has been reported in p21ras between the corresponding Ser17 and Asp57 in both, the GDP and the GMP–PNP bound forms [25]. The complete loss of GTPase activity observed for the Thr112Ala mutant of A. ambivalens can be explained by the inability to coordinate the Mg2+ ion. It is interesting to note that recent kinetic data from E. coli Ffh suggest that nucleotide binding to Ffh is independent from Mg2+ [26] in contrary to what has been observed for other GTPases, such as Rab proteins [27]. In small GTPases the G2 region (adjacent to switch I) is involved in Mg2+-binding and in the modulation of GTPase activity by its interaction with GTPase regulatory proteins [24]. This region is well conserved in SRPGTPases although some variation can be noticed in SRP54 homologues from bacteria, archaea (see Figure 2) and cyanobacteria (not shown). In A. ambivalens this element is very similar to the one in E. coli Ffh as the strictly conserved aspartic acid residue is followed by a short hydrophobic residue and a tyrosine (‘135DVY’; in

520

Structure 2000, Vol 8 No 5

Figure 4

Cα representation of a superposition of switch I and switch II regions from NG-Ffh of A. ambivalens (red) with T. aquaticus apo (dark blue, 3NG1) and the GDP-bound form (light blue, 1NG1). The conserved residues Asp135 and Arg191 (salt link in 3NG1) and Tyr137 are shown with the sidechain. In A. ambivalens the G2 motif is modified to ‘135DVYR’ (compared with ‘DTQR’ in T. aquaticus). The superposition also shows that there is no significant change in the mainchain structure between the apo and GDP-bound form of T. aquaticus. This figure and figures 5–7 have been created with MOLSCRIPT [49] and Raster3D [50].

single-letter amino acid code). The invariant Asp135 is oriented towards the active site similar to the corresponding Asp330 in E. coli FtsY and in both structures no salt link with the invariant Arg191 (Arg386 in FtsY) in the G3 region is seen in contrast to T. aquaticus Ffh. Instead, stacking of the aromatic ring of Tyr137 with the guanidinium moiety of Arg191 is observed (Figure 4). As a consequence of the local interactions the G2 region moves towards the active site, which also affects the I-box. In A. ambivalens the I-box is packed more closely to the G domain, similar to its position in E. coli FtsY (see also Figure 3). There is no direct protein–protein interaction between the I-box and the G2 region in A. ambivalens. The interaction is mediated through four water molecules. In small GTPases the G2 region contains a conserved threonine residue. In p21ras this residue (Thr35) was shown to play a role in GTP hydrolysis by interaction with the Mg2+ ion, the γ-phosphate and Ser17 and to rotate away after hydrolysis [25]. In E. coli FtsY, substitution of the corresponding Thr331 by alanine results in an approximately tenfold decrease of GTP hydrolysis (KtK, unpublished observations) similar to what has been reported for the

Thr35→Ala mutant of p21ras [28]. Because in A. ambivalens and E. coli Ffh the corresponding threonine residue has been substituted by a valine, one of the adjacent residues has to perform its role in GTP hydrolysis. It is tempting to propose that Tyr137 might be this residue. The invariant Asp135, however, is another possible candidate for Mg2+ binding. The superposition of A. ambivalens Ffh with the p21ras/GAP in complex with a transition state analog [29] places Asp135 within hydrogen-bonding distance of the Mg2+ ion, which favours the latter idea. The same can be observed for Asp330 in E. coli FtsY. Here, the local structure appears to be stabilised by the interactions of Thr331 with the I-box [16]. In T. aquaticus Thr136 is interacting with the I-box in a similar way. Here a salt link between Asp135 and Arg191 was proposed as a major structural component (Figure 4). This salt link is also present when GDP is bound, but lost when GDP and Mg2+ are present. It is not surprising that the presence of GDP and GDP Mg2+ has only a minor effect on the local structure [18]. One would expect this region to change significantly upon binding of GTP or transition-state analogs, whereas the GDP-bound protein is already prone for nucleotide release and is therefore more similar to the apo structure. The G3 region (‘187DTAG’) is strictly conserved and superimposes well for the three available structures of SRP-GTPases. The adjacent switch II region is well ordered in all these structures, whereas high temperature factors and multiple conformations are often found in the corresponding region of p21ras and related GTPases. In A. ambivalens the conserved Arg191 interacts with Tyr137 in the G2 region (Figure 4) and a two-residue insertion (Gly193 and Tyr194) is present (Figure 5). This insertion is unique for Ffh from A. ambivalens. Interestingly, the OH group of Tyr194 is engaged in a network of hydrogen bonds with the mainchain of Val106 and Gly108 and with the sidechain of Thr110 in the P-loop, mediated by water molecules. Moreover, Tyr194 is connected to Thr247 and Lys248 in the G4 region in an indirect way through several water molecules. By these interactions, Tyr194 might sense the occupancy of the active site and communicate the information between the G1 and G4 regions and the I-box. Given that A. ambivalens is a thermophile, this insertion seems to contradict one of the proposed principles of thermostability [30]. Helix α4 is kinked, which brings the switch II region, below the P-loop when compared with the structures of other SRP GTPases (Figures 3b, 5). Glu197 interacts with Lys227 and a number of interactions link the switch II region via helix α5 to the N domain (see above). This might allow for conformational changes of the N domain to be transmitted to the switch II region, which suggests an elegant way to sense the occupancy of the active site via interdomain interactions. In the GTPase superfamily an additional consensus element for nucleotide binding can be assigned, the G5

Research Article Crystal structure of A. ambivalens Ffh Montoya et al.

521

Figure 5 Interdomain and intradomain interactions in NG-Ffh might sense the nucleotide-binding state. The conserved B–C loop (residues 36–41) of the N domain interacts closely with residues in α5 (mainly with Gln226) by polar interactions. Tyr194 of the glycine–tyrosine insertion in the switch II region is involved in hydrophobic interactions with Val106 and Ile224 as well as in a network of interactions that connect the switch II region with the G4 region and even the N domain. Polar interactions can be observed between Glu197 and the mainchains of Leu231 and Lys227.

region or ‘closing loop’ [18,31]. This region interacts with the ribose ring of the nucleotide and is well conserved in the SRP-GTPase family. In A. ambivalens this region (residues 272–282) forms a β-hairpin loop that is stabilised mainly by polar interactions. These involve the sidechain of Lys117 and the mainchains of Gly275, Glu280 and Leu278. Gly273 is hydrogen bonded to the mainchains of Met249 and Lys248 in the G4 region, thereby contributing to the β-sheet core of the G domain. The G5 region is well ordered in Ffh of A. ambivalens and its overall conformation is very similar to the corresponding regions in FtsY from E. coli [16] and the recent structures from T. aquaticus [18]. A common signalling mechanism

Although the interface between the N and G domain contains highly conserved sequence motifs (‘36LISADVN’ and ‘252TAKGGG’ in A. ambivalens, see Figure 2) that have been implicated in signal transduction [16,18], the details of the interactions are quite different in A. ambivalens. In particular, the loop connecting G4 to helix α6 adopts a different conformation possibly because Thr252 (Asp250 in T. aquaticus) does not form hydrogen bonds with the glycine residues of the so-called ‘DARGG’ motif [18]. In addition, in A. ambivalens this motif contains three glycines instead of two in T. aquaticus, which as a consequence shifts the conserved Ser260 so that it cannot interact with the N domain directly. The sidechain and the mainchain oxygens of Ser260 interact through two water molecules with the mainchain of Asn42 and Leu83, respectively (Figures 1d and 5). In E. coli FtsY similar interactions are observed for Ser459. In T. aquaticus the corresponding Ser258 interacts with the mainchain nitrogen of Val44. Given that Ser260 in α6 is highly conserved throughout the SRP-GTPase family (Figure 2), it might play an important role in these interactions. Helix α6 in A. ambivalens is less strongly involved in the interaction with the N domain than in T. aquaticus and the

number of hydrophobic contacts is significantly lower. Hydrogen bonds important for the N–G interaction include the mainchain nitrogen of Gly257 and the mainchain oxygen of Ala39. Mainchain and sidechain atoms of Gln226 in helix α5 are involved in hydrogen bonds with Asp40 and Asn42 in the loop between the B and C helices in the N domain. All structures of the conserved core domains of the prokaryotic SRP and SR show a close interaction of the N and G domain, whereas the angle between the domains is variable (see Figure 3). The hypothesis that nucleotide binding is monitored or sensed by the N domain is supported by recent structural and biochemical data. A mutational study in the ‘ALLEADV’ motif in the N domain showed defects in signal sequence binding by SRP54 [32]. The magnitude of the defect was independent of the preprotein substrate, which suggested that the mutations do not alter the specificity of signal sequence recognition. In contrast, the same authors reported that mutations in the guanosine triphosphatase consensus sequences had no effect on signal sequence binding, but severely impaired protein translocation activity. These results strongly suggest that the event of signal peptide binding is transmitted to the G-binding domain via a conformational change in the N domain. This change might affect the conserved hydrophobic packing of residues in the NG interface (Ile4, Phe289, Phe284, Phe84, Met249, Leu259, Ile246 and Ile272) with residues in the G4 region (Figure 6). On the other hand, an important structural change upon signal peptide binding might occur in the switch II region, which is indirectly connected via α5 with the conserved loop (‘SLISADV’ in A. ambivalens, see above, Figures 1d and 5). These conformational changes would then allow for the binding of the nucleotide after the M domain is loaded with the signal peptide. In this state SRP is competent for complex formation with the receptor. A similar role might be assigned to the homologous amino acids in the conserved core in the SRP receptor with the

522

Structure 2000, Vol 8 No 5

Figure 6

Hydrophobic interactions close to the G4 region. A conserved hydrophobic core is formed by Ile4, Phe289, Phe284, Phe84, Met249, Leu259, Ile246 and Ile272.

peculiarity that here signalling should be coupled to the direct interaction of the acidic domain with the membrane [13] or an unknown membrane protein. Therefore the conserved core domain might provide a common scaffold that undergoes similar conformational changes in Ffh and FtsY because of signalling from the third domain depending on signal peptide binding or membrane association. As a consequence, SRP and SR adopt a state competent for complex formation and protein translocation [32]. The high conservation of the ‘LISADVN’ and ‘TAKGGG’ motifs and of residues involved in the hydrophobic packing in the NG interface of SRP and SR (see Figure 2) supports the hypothesis of a common signalling mechanism. SRP–SR complex formation

Ffh and FtsY form a complex during protein targeting that dissociates after GTP hydrolysis. The structure of the SRP–SRP receptor complex is crucial to understand the mechanism by which signal peptide release is coupled to the stimulation of the GTPase in both Ffh and FtsY. The recent structures of Ffh and FtsY, the structures of related proteins and the homology between SRP-GTPases and other GTPases allow us to propose a model for their interaction. The domain alignment program DALI [33] identified the nitrogenase iron protein (NIP) (PDB entries 1NIPA and 1N2C) [34], dethiobiotin synthetase (PDB entry 1DAI) [35], and other ATP-utilising proteins as the closest related structures to the conserved core domains of FtsY and Ffh. This similarity has been noticed previously

[36]. These proteins are homodimers. Surprisingly, these structures were found to be more similar to SRP-GTPases than other GTPases such as elongation factor Tu (EF-Tu). The topologies of NIP and the SRP-GTPases are remarkably similar in the central β sheet (as seen in Figure 7a,b), varying in the number and the organization of α helices on both sides. The rmsd between Ffh, FtsY and NIP is 1.9 Å and 2.0 Å for 141 and 139 Cα atoms, respectively. These numbers are similar to the ones obtained for the superpositions of the different SRP-GTPases (see above). NIP was shown to undergo nucleotide-dependent conformational changes and the active site shares similarities with Gtα and p21ras [34]. Interestingly, in NIP the Mg2+ ion interacts with Ser16 (Thr112 in A. ambivalens, Ser17 in p21ras) and Asp39 (Asp330 in FtsY, Asp135 in A. ambivalens, Thr35 in p21ras). This supports our previous suggestion that Asp135 in A. ambivalens might be involved in Mg2+ binding (see above). The P-loops of the monomers come very close in the NIP dimer and catalytically important residues cross the dimer interface reminiscent of the arginine fingers in the GTPase–GAP complexes [29,37]. The NIP dimer closes when the transition state analogue ADP-AlF4– is bound and this state displays the lowest rmsd to the structures of Ffh and FtsY. We have superimposed Ffh and FtsY with this form of the dimer and thereby created a model of the SRP–SR complex, which is consistent with a number of criteria that are known for the interaction. Because Ffh and FtsY form a complex during protein targeting and the NG domains share significant homology in their sequences as well as their three-dimensional structures, they can be regarded as a heterodimer. In the model the switch I and II regions are involved in the interaction, similar to the known structures of GTPases in complex with regulatory proteins [29,37–41]. The GTPases of Ffh and FtsY are both activated upon complex formation and have been suggested to act as GTPase-activating proteins (GAPs) for each other [12]. GAPs contain conserved arginine residues that are involved in catalysis by stabilizing the transition state [29]. SRP-GTPases contain two strictly conserved arginine residues in the G2 and G3 regions (Arg138 and Arg191 in Ffh) that might play the role of the arginine fingers. The conformational change that must occur upon binding of GTP has not been considered (Ffh and FtsY are both in their apo forms), which explains one clash in the interface between the I-box and switch II. Our model fullfills another important requirement for complex formation of Ffh and FtsY as the headto-head interaction between the NG domains leaves enough room for the two additional domains, which are absent in the model: the M domain of Ffh in complex with the RNA and signal peptide as part of the RNC and the acidic domain of FtsY for its interaction with the membrane [13]. Taking these points together the heterodimer model presented here might be of relevance for the SRP–SR complex and allows now for a detailed analysis of the proposed interactions.

Research Article Crystal structure of A. ambivalens Ffh Montoya et al.

Figure 7 A model for the SRP–SR complex (see text for details). Stereoview Cα trace of a superposition of (a) A. ambivalens Ffh (green) with one NIP monomer (orange, from 1N2C) and (b) E. coli FtsY (blue) with the second monomer of NIP (orange, from 1N2C). (c) Cα trace of the model of the SRP–SR complex: the interface involves the switch I and II regions and the I-box of both molecules. The NG domains of A. ambivalens Ffh (red) and E. coli FtsY (blue) are at a relative angle of about a 90°.

523

524

Structure 2000, Vol 8 No 5

Biological implications The SRP pathway from archaea is much less well characterized than from other organisms and a number of components might still not be identified. The NG-Ffh structure from Acidianus ambivalens is the first structure of an SRP component from an archaeon. The structure described here allows for a detailed comparison of SRP-GTPases, which puts the conclusions drawn on a more solid basis. The comparison of the known structures of SRPGTPases shows a rather high degree of conformational flexibility in subdomain organisation and points towards a role of this flexibility in the regulation of SRP-GTPases. The sequence similarity between Ffh and FtsY and the high number of known sequences allow us to perform phylogenetic studies of the SRP-GTPase family. These suggest that the SRP-GTPases might have evolved earlier than the small GTPases, probably with other nucleotidebinding proteins [42]. This is reflected by structure comparisons that reveal a higher similarity of SRP-GTPases to nucleotide-binding proteins than to other GTPases. The model proposed here for the complex of Ffh and FtsY is in agreement with the current knowledge of this interaction and also takes the interactions of small GTPases with regulatory proteins into account. Given that SRP and its receptor have been proposed to act as GTPase-activating proteins (GAPs) for each other, the question remains whether there are arginine fingers in SRP-GTPases. The heterodimer model presented here sets the ground for a detailed study. Site-directed mutagenesis and spectroscopic techniques can now be used to validate the model. The mutual stimulation of SRP and SR might be yet another variation of the GAP theme.

Materials and methods Cloning, expression and crystallisation Methods used to express and purify recombinant full-length Ffh and NG-Ffh have been described [21,23]. The engineered fragment of Ffh was concentrated to 8 mg/ml. Crystals of NG-Ffh were grown using the hanging-drop vapour diffusion method at 20°C. Orthorhombic crystals were obtained for the NG-Ffh fragment at 20°C. The protein drops were prepared by mixing 1.5 µl of the protein solution with 1.5 µl of well buffer that contained 100 mM acetate buffer pH 5.0, 100 mM calcium acetate and 18% PEG 8000. The maximum size was around 0.3 × 0.4 × 0.3 mm within one to two weeks. The crystals appear as tetragonal prisms often containing cavities. Others were shorter and more cube-like. One of the latter ones was used for the synchrotron data collection. The crystals diffracted to 2.8 Å using a rotating anode as X-ray source and to 2.0 Å with synchrotron radiation.

Data collection and processing The crystals belong to the orthorhombic space group C2221 with cell dimensions a = 64.86 Å, b = 128.02 Å, c = 72.00 Å and contain one molecule per asymmetric unit. Data were processed using the programs DENZO and Scalepack [43] or MOSFLM and Scala (CCP4 suite) [44] depending on the data set. A summary of the statistics is given in Table 1. Data were collected in-house with a MAR-image plate detector mounted on a rotating anode X-ray generator. All data were collected at 100K using an Oxford cryo-system. A cryo buffer with 15% MPD was added in small amounts to replace the protein solution around the crystals. The crystals were then directly flash-frozen in the

cryo-stream using a nylon loop. A range of heavy-metal soaks (up to 10 days) were prepared using different concentrations and times. A first derivative was obtained by soaking the native crystals in mother liquor containing 5 mM of KAu(CN)2. Selenomethionine protein was prepared [23]. The selenium incorporation was checked by mass spectrometry that showed full incorporation for all five sites available in the protein. The native and KAu(CN)2 data sets were collected in-house using a rotating anode. The selenomethionine data were collected on the BM14 beam-line at the ESRF at a wavelength of 0.99 Å. Anomalous data were collected for both derivatives.

Patterson solution, phase calculation and structure determination After data processing the crystallographic calculations were performed using the CCP4 suite of programs [44]. Isomorphous difference Patterson maps at 4 Å resolution for the gold derivative provided a clear solution for two sites. Anomalous difference Patterson for this derivative yielded a solution consistent with the isomorphous difference Patterson function. Cross-difference Fourier maps using the phases of the gold derivative revealed the positions of five selenium atoms in the selenium derivative. The sidechain of the fifth methionine was not wellenough defined to reveal the position of the last selenium atom. Heavy-atom parameters were refined initially using the program PHASES [45] and the refined values were used in the program SHARP [46]. Initial MIRAS phases at 2.8 Å resolution with a figure of merit (FOM) of 0.43 using both isomorphous and anomalous data were obtained. A molecular mask was built using a solvent content of 37%. Solvent flattening using SOLOMON (CCP4 suite) [44] improved the map. Fourier maps were then calculated and inspected to locate and build interpretable pieces of secondary structure using the graphics program O [47]. Initially a polyalanine model was used to represent the secondary structure. The resulting electron-density map allowed to place 290 Cα carbons out of 301 residues. There was one loop region (region 108–111) with initially weak electron density. The density in this region improved during refinement permitting to include those residues in the model. The sequence could be unambiguously assigned and the sidechains were built in.

Refinement The current model contains 295 out of 297 residues of the protein. The model has been refined using a 2.0 Å native data set collected at ESRF at 100K and rebuilt against (2|Fo|–|Fc|) maps using the Rfree as a monitor of model quality. The refinement was performed using the maximum likelihood and bulk-solvent correction options implemented in the program REFMAC (CCP4 suite) [44]. The initial Rfactor and Rfree were 41.0 and 44.9, respectively. Solvent water molecules were identified using the program ARP and checked in the map using the wat_pekpik routine in O. The average B factor of 35.2 Å2 for the mainchain of the current model is in good agreement with the estimated B factor of 26.7 Å2 from the Wilson plot. The Ramachandran plot (not shown) showed that over 92.5% of the mainchain torsion angles are in the most favoured regions with no amino acids in disallowed regions. The geometric parameters were checked with the program WHAT_CHECK in the WHATIF package [48].

Accession numbers Accession codes will be deposited at the PDB.

Acknowledgements We thank Andy Thompson from the EMBL-outstation in Grenoble, for help with data collection at BM-14 and the ESRF for access to BM-14. We are also grateful to A Murzin (Cambridge) for helpful suggestions on the NIP based model. We acknowledge the support of this work through the EUnetwork grant EUCT96/0035 to IS, an FCI grant to KtK, a grant from the Spanish Ministery of Education and Science to GM and a grant Scha 125/17-3 from the Deutsche Forschungsgemeinschaft to GSch.

References 1. Lütcke, H. (1995). Signal recognition particle (SRP), a ubiquitous initiator of protein translocation. Eur. J. Biochem. 228, 531-550. 2. Walter, P. & Johnson, A.P. (1994). Signal sequence recognition and protein targeting to the endoplasmic reticulum membrane. Annu. Rev. Cell Biol. 10, 89-117.

Research Article Crystal structure of A. ambivalens Ffh Montoya et al.

3. Luirink, J., High, S., Wood, H., Giner, A., Tollervey, D. & Dobberstein, B. (1992). Signal sequence recognition by an Escherichia coli ribonucleoprotein complex. Nature 359, 741-743. 4. Luirink, J., et al., & Kusters, R. (1994). An alternative protein targeting pathway in Escherichia coli: studies on the role of FtsY. EMBO J. 13, 2289-2296. 5. Bernstein, H.D., Poritz, M.A., Freymann, D.M. & Walter, P. (1993). Functional substitution of the signal recognition particle 54-kDa subunit by its Escherichia coli homolog. Proc. Natl Acad. Sci. USA 340, 482-486. 6. Seluanov, A. & Bibi, E. (1997). FtsY, the prokaryotic signal recognition particle receptor homologue, is essential for biogenesis of membrane proteins. J. Biol. Chem. 272, 2053-2055. 7. Ulbrandt, N.D., Newitt, J.A. & Bernstein, H.D. (1997). The E. coli signal recognition particle is required for the insertion of a subset of inner membrane proteins. Cell 88, 187-196. 8. de Gier, J.W., et al., & von Heijne, G. (1998). Differential use of the signal recognition particle translocase targeting pathway for inner membrane protein assembly in Escherichia coli. Proc. Natl Acad. Sci. USA 95, 14646-14651. 9. Beck, K., Wu, L.F., Brunner, J. & Muller, M. (2000). Discrimination between SRP- and SecA/SecB-dependent substrates involves selective recognition of nascent chains by SRP and trigger factor. EMBO J. 19, 134-143. 10. Kusters, R., et al., & Luirink, J. (1995). The functioning of the SRP receptor FtsY in protein-targeting in E. coli is correlated with its ability to bind and hydrolyse GTP. FEBS Lett. 372, 253-258. 11. Miller, J.D., Bernstein, H.D. & Walter, P. (1994). Interaction of E. coli Ffh/4.5S ribonucleoprotein and FtsY mimics that of mammalian signal recognition particle and its receptor. Nature 367, 657-659. 12. Powers, T. & Walter, P. (1995). Reciprocal stimulation of GTP hydrolysis by two directly interacting GTPases. Science 269, 1422-1424. 13. de Leeuw, E., et al., & Sinning, I. (2000). Anionic phospholipids are involved in membrane association of FtsY and stimulate its GTPase activity. EMBO J. 19, 531-541. 14. Zopf, D., Bernstein, H.D. & Walter, P. (1993). GTPase domain of the 54-kD subunit of the mammalian signal recognition particle is required for protein translocation but not for signal sequence binding. J. Cell Biol. 120, 1113-1121. 15. Lütcke, H., High, S., Römisch, K., Ashford, A.J. & Dobberstein, B. (1992). The methionine-rich domain of the 54 kDa subunit of signal recognition particle is sufficient for the interaction with signal sequences. EMBO J. 11, 1543-1551. 16. Montoya, G., Svensson, C., Luirink, J. & Sinning, I. (1997). Crystal structure of the NG domain from the signal-recognition particle receptor FtsY. Nature 385, 365-368. 17. Freymann, D.M., Keenan, R.J., Stroud, R.M. & Walter, P. (1997). Structure of the conserved GTPase domain of the signal recognition particle. Nature 385, 361-364. 18. Freymann, D.M., Keenan, R.J., Stroud, R.M. & Walter, P. (1999). Functional changes in the structure of the SRP GTPase on binding GDP and Mg2+GDP. Nat. Struct. Biol. 6, 793-801. 19. Clemons, W.M. Jr., Gowda, K., Black, S.D., Zwieb, C. & Ramakrishnan, V. (1999). Crystal structure of the conserved subdomain of human protein SRP54M at 2.1 Å resolution: Evidence for the mechanism of signal peptide binding. J. Mol. Biol. 292, 697-705. 20. Keenan, R.J., Freymann, D.M., Walter, P. & Stroud, R.M. (1998). Crystal structure of the signal sequence binding subunit of the signal recognition particle. Cell 94, 181-191. 21. Moll, R., Schmidtke, S. & Schäfer, G. (1999). Domain structure, GTPhydrolyzing activity and 7S RNA binding of Acidianus ambivalens ffhhomologous protein suggest an SRP-like complex in archaea. Eur. J. Biochem. 259, 441-448. 22. Pohlschroeder, M., Prinz, W.A., Hartmann, E. & Beckwith, J. (1997). Protein translocation in the three domains of life, Variations of a theme. Cell 91, 563-566. 23. Montoya, G., te Kaat, K., Moll, R., Schäfer, G. & Sinning, I. (1999). Crystallization and preliminary X-ray diffraction studies on the conserved GTPase domain of the signal recognition particle from Acidianus ambivalens. Acta Crystallogr. D 55, 1949-1951. 24. Bourne, H.R., Sanders, D.A. & McCormick, F. (1990). The GTPase superfamily, a conserved switch for diverse cell functions. Nature 348, 125-132. 25. Schlichting, I., et al., & Petsko, G.A. (1990). Time-resolved X-ray crystallographic study of the conformational change in Ha-Ras p21 protein on GTP hydrolysis. Nature 345, 309-315.

525

26. Jagath, J.R., Rodnina, M.V., Lentzen, G. & Wintermeyer, W. (1998). Interaction of guanine nucleotides with the signal recognition particle from Escherichia coli. Biochemistry 37,15408-15413. 27. Simon, I., Zerial, M. & Goody, R.S. (1996). Kinetics of interaction of Rab5 and Rab7 with nucleotides and magnesium ions. J. Biol. Chem. 271, 20470-20478. 28. John, J., et al., & Wittinghofer, A. (1993). Kinetic and structural analysis of the Mg(2+)-binding site of the guanine nucleotide-binding protein p21H-ras. J. Biol. Chem. 268, 923-929. 29. Scheffzek, K., et al., & Wittinghofer, A. (1997). The Ras–RasGAP complex: structural basis for GTPase activation and its loss in oncogenic Ras mutants. Science 277, 333-338. 30. Vieille, C., Burdette, D.S.& Zeikus, J.G. (1996). Thermozymes. Biotechnol. Ann. Rev. 2, 1-83. 31. Kjeldgaard, M., Nyborg, J. & Clark, B.F. (1996). The GTP binding motif: variations on a theme. FASEB J. 10,1347-1368. 32. Newitt, J.A. & Bernstein, H.D. (1997). The N-domain of the signal recognition particle 54-kDa subunit promotes efficient signal sequence binding. Eur. J. Biochem. 245, 720-729. 33. Holm, L. & Sander, C. (1997). Dali/FSSP classification of threedimensional protein folds. Nucl. Acids Res. 25, 231-234. 34. Schindelin, H., Kisker, C., Schlessman, J.L., Howard, J.B. & Rees, D.C. (1997). Structure of ADP × AIF4(–)-stabilized nitrogenase complex and its implications for signal transduction. Nature 387, 370-376. 35. Huang, W., Lindqvist, Y., Schneider, G., Gibson, K.J., Flint, D. & Lorimer, G. (1994). Crystal structure of an ATP-dependent carboxylase, dethiobiotin synthetase, at 1.65 Å. Structure 2, 407-414. 36. Schlessman, J.L., Woo, D., Joshua-Tor, L., Howard, J.B & Rees, D.C (1998). Conformational variability in structures of the nitrogenase iron proteins from Azotobacter vinelandii and Clostridium pasteurianum. J. Mol. Biol. 280, 669-685. 37. Rittinger, K., Walker, P.A., Eccleston, J.F., Smerdon, S.J. & Gamblin, S.J. (1997). Structure at 1.65 Å of RhoA and its GTPase-activating protein in complex with a transition-state analogue. Nature 389, 758-762. 38. Nassar, N., Horn, G., Herrmann, C., Scherer, A., McCormick, F. & Wittinghofer, A. (1995). The 2.2 Å crystal structure of the Ras-binding domain of the serine/threonine kinase c-Raf1 in complex with Rap1A and a GTP analogue. Nature 375, 554-560. 39. Goldberg, J. (1999). Structural and functional analysis of the ARF1–ARFGAP complex reveals a role for coatomer in GTP hydrolysis. Cell 96, 893-902. 40. Kawashima, T., Berthet-Colominas, C., Wulff, M., Cusack, S. & Leberman R. (1996). The structure of the Escherichia coli EF-Tu–EFTs complex at 2.5 Å resolution. Nature 379, 511-518. 41. Wang, Y., Jiang, Y., Meyering-Voss, M., Sprinzl, M. & Sigler, P. (1997). Crystal structure of the EF-Tu–EF-Ts complex from Thermus thermophilus. Nat. Struct. Biol. 4, 650-656. 42. Gribaldo, S. & Cammarano, P. (1998). The root of the universal tree of life infered from anciently duplicated genes encoding components of the protein-targeting machinery. J. Mol. Evol. 47, 508-516. 43. Otwinowski, Z. & Minor, W. (1997). Processing of X-ray diffraction data collected in oscillation model. Methods Enzymol. 276, 307-326. 44. The CCP4 suite programs for X-ray crystallography (1994). Acta Crystallogr. D 50, 760-776. 45. Furey, W. & Swaminathan, S. (1997). PHASES-95: a program package for processing and analyzing diffraction data from macromolecules. Methods Enzymol. 277, 590-600. 46. De la Fortelle, E. & Bricogne, G. (1997). Maximum-likelihood heavy atom parameter refinement for multiple isomorphous replacement and multiwavelength anomalous diffraction methods. Methods Enzymol 276, 472-494. 47. Jones, T.A., Zou, J.Y., Cowan, S.W.& Kjeldgaard, M. (1991). Improved methods for building protein models in electron density maps and the location of errors in these models. Acta Crystallogr. A 47,110-119. 48. Vriend, G. (1990) WHAT IF: A molecular modeling and drug design program. J. Mol Graph. 8,52-56. 49. Kraulis, P. (1991). MOLSCRIPT: a program to produce both detailed and schematic plots of protein structures. J. Appl. Crystallogr. 24, 946-950. 50. Merritt, E.A. & Murphy, M.E. (1994). Raster3D version 2.0. A program for photorealistic molecular graphics. Acta Crystallogr. D 50, 896-873.