Membrane Protein Insertion in Bacteria from a Structural Perspective

Membrane proteins are inserted into the lipid bilayer in bacteria by two pathways. The Sec machinery is responsible for the insertion of the majority of the membrane proteins after targeting by the SRP/FtsY components. However, there is also a class of membrane proteins that insert independent of the Sec machinery. These proteins require a novel protein called YidC. Recently, the structural details of the Sec machinery have come to light via X-crystallographic analysis. There are now structures of the membrane-embedded Sec protein-conducting channel, the SecA ATPase motor, and the targeting components. The structures give clues to how a polypeptide is translocated across the membrane and how the transmembrane segments of a membrane protein are released from the Sec complex. Additionally, the structure of the targeting components sheds light on how the membrane substrates are selected for transport and delivered to the membrane. Introduction Membrane proteins are ubiquitous in nature and comprise around 30% of the total proteins within the cell. Membrane proteins play vital functions for the cell. They act as receptors where they are involved in transmitting information from the extracellular environment into the interior of the cell. Membrane proteins also function as transporters to move sugars, amino acids and other energy rich molecules and ions into the cell. Other functions of membrane proteins include energy harvesting and energy transduction roles in photosynthesis and oxidative phosphorylation, as well as functions in lipid synthesis and catabolism. Given the wide variety of functions, there is a diversity of membrane protein structures. However, generally almost all bacterial inner-membrane integral membrane proteins have helical transmembrane segments that range from 20 to 30 residues in length, with tryptophan and tyrosine residues being enriched near phospholipid headgroups and the connecting loops between helical transmembrane segments tend to be short.1 In this review, we will bring the reader up on the latest developments in bacterial membrane protein biogenesis with a focus on structural aspects of the targeting and translocation components that facilitate insertion. In the field of membrane protein biogenesis, there are at least four main problems. (1) How do membrane proteins with hydrophobic surfaces avoid aggregating in the cytoplasm? *Ross E. Dalbey—Department of Chemistry, The Ohio State University, Columbus, Ohio 43210, U.S.A. E-mail: [email protected] Protein Movement Across Membranes, edited by Jerry Eichler. ©2005 Eurekah.com. 05Eichler(Dalbey) 1/6/05, 2:18 PM 55 © 20 05 C op yr ig ht Eu re ka h / L an de s Bi os ci en ce D o N ot D is tri bu te Protein Movement Across Membranes 56 (2) How are hydrophilic domains translocated across the membrane? (3) How are hydrophobic domains integrated into the membrane? (4) What are the energetics of membrane protein insertion? Not surprising, there are proteins that catalyze the targeting of proteins to the membrane and the insertion into the lipid bilayer. In bacteria, there are two pathways used for membrane protein insertion; the Sec-pathway and YidC pathway. The majority of proteins use the Sec pathway for insertion (Fig. 1A). A subset of proteins insert by a Sec-independent pathway involving YidC (Fig. 1B). The goal of understanding the molecular events involved in membrane protein assembly is not only of significant scientific interest in the membrane biogenesis area but is essential for the understanding of the disease states that result when these events go wrong. Figure 1. Schematic depiction of the two known membrane protein integration (assembly) pathways. A) The Sec-dependent pathway (the heterotrimer SecDFyajC and the ATPase SecA are not shown). B) The YidC pathway. The PDB coordinates used for the large ribosomal subunit from Deinococcus radiodurans were 1NKW, the PDB coordinates used for SecYEβ from Methanococcus janaschii, were 1RHZ, The PDB coordinated used for Ffh from Sulfolobus solfataricus and FtsY from Thermus aquaticus, were 1QZW and 1RJ9, respectively. The program PyMol was used to make this figure. 05Eichler(Dalbey) 1/6/05, 2:19 PM 56 © 20 05 C op yr ig ht Eu re ka h / L an de s Bi os ci en ce D o N ot D is tri bu te 57 Membrane Protein Insertion in Bacteria from a Structural Perspective Insertion by the Sec-Translocase Mediated Pathway Many membrane proteins inserted by the Sec pathway are targeted to the membrane by the evolutionarily-conserved Signal Recognition Particle (SRP) route. In this pathway, the cytosolic component SRP, comprised of Ffh and the 4.5S RNA binds to the membrane protein and targets the protein to the SRP receptor FtsY. SRP binds to the hydrophobic region of the membrane protein as it emerges from the ribosomal tunnel (Fig. 1A). Then, the ribosome/ mRNA/nascent membrane protein/Ffh complex is targeted to FtsY that is associated with the membrane. Insertion of a protein into the membrane is initiated by a cleavable signal peptide or a noncleaved transmembrane segment. The transmembrane segments are integrated into the membrane and the hydrophilic domains are either translocated across the membrane or remain within the cytoplasm. The membrane protein uses the Sec translocase for insertion into the membrane and translocation of hydrophilic domains across the membrane (Fig. 1A). In E. coli, the Sec translocase is comprised of the SecYEG protein-conducting channel and the trimeric SecDFYajC complex (for review see ref. 6). The protein YidC interacts with the hydrophobic regions of membrane proteins during the insertion of the protein into the membrane. In some cases, the membrane-associated ATPase SecA is required for the translocation of large hydrophilic domains of membrane proteins. Targeting The targeting components Ffh and FtsY are important for the insertion of membrane proteins as depletion of Ffh and FtsY within the cell has been shown to inhibit the insertion of a variety of membrane proteins. The SRP component Ffh in E. coli is homologous to the 54 KDa subunit of the eukaryotic SRP11 which is comprised of 6 polypeptides and a 7S RNA component. Ffh exists in complex with a 4.5S RNA instead of the 7S RNA seen in the eukaryotic complex. SRP Ffh has been shown to bind to signal peptides of exported proteins and hydrophobic segments of membrane proteins.10,12 For membrane proteins containing multiple hydrophobic regions, it may be sufficient for Ffh to bind to the first hydrophobic domain and target the protein to the membrane. Efficient membrane targeting of proteins which have hydrophobic surfaces is important as it prevents aggregation in the aqueous cytoplasm. The SRP receptor in bacteria (FtsY) is simpler than the SRP receptor (SR) in eukaryotes which contain two subunits, SRα and SRβ. The membrane-associated protein FtsY is homologous to the SRα subunit. Both FtsY and Ffh are essential bacterial proteins.13,14 Ffh has been shown to form a complex with FtsY, in a GTP-dependent manner. Following GTP hydrolysis, the Ffh and FtsY complex disassembles from the targeted nascent protein and the nascent chain can insert into the Sec machinery. Interestingly, it has been found that the GTPase activity of Ffh is stimulated by FtsY and the GTPase activity of FtsY is stimulated by Ffh. In order to provide insight into the protein targeting mechanism, it is very useful to obtain structural knowledge of the targeting components. Ffh contains three domains, i.e, the amino-terminal N domain, the GTPase G domain and the methionine rich M domain (Fig. 2A).16,17 The M-domain is connected to the N and G-domains by a flexible linker. The crystal structure of the M domain from Thermus aquaticus reveals a hydrophobic groove lined with methionine residues that has been proposed to bind to the signal peptide or the membrane anchor domain of the nascent polypeptide.17 Interestingly, a crystal structure of the E. coli Ffh domain with the domain IV of the 4.5S RNA suggests that the signal sequence recognition domain is comprised of both protein and RNA (SRP)(Table 1A). A structure of the complete SRP54 (Ffh) in complex with helix 8 of the SRP RNA component revealed the overall juxtaposition of the M, G and N domains relative to each other. 05Eichler(Dalbey) 1/6/05, 2:19 PM 57 © 20 05 C op yr ig ht Eu re ka h / L an de s Bi os ci en ce D o N ot D is tri bu te Protein Movement Across Membranes 58 Numerous structures are available for the NG domains of the Ffh from archaeal homologs. These structures have been solved both in the presence and absence of GDP or nonhydrolyzable GTP analogs (see Table 1B). The N domain is comprised of a four-helix bundle, which is closely associated with the G domain (Ras-like GTPase) that has a core made up of a five-stranded β-sheet surrounded by α-helices. The G domain also contains an Insertion Box Domain (IBD) which is unique to the SRP GTPases. A similar structural Figure 2. A) A ribbon diagram of the overall structure of the SRP core from the archaeon Sulfolobus solfataricus. The structure reveals the interdomain communication between the N domain, the G domain, the M domain and helix 8 of SRP RNA. The RNA is shown in a stick diagram. The PDB coordinates 1QZW and the program PyMol were used to make this figure. B) A ribbon diagram with transparent surface showing the heterodimeric complex of the signal recognition particle protein Ffh and its receptor FtsY from the species Thermus aquaticus. Ffh is rendered in a darker shade and FtsY is shown in a lighter shade. The bound GTP analogue molecules are shown in van der Waal’s spheres. The N-terminal domains (N domain) and the GTP binding domains (G domain) for each protein are labeled. The PDB coordinates 1RJ9 and the program PyMol were used to make this figure. 05Eichler(Dalbey) 1/6/05, 2:19 PM 58 © 20 05 C op yr ig ht Eu re ka h / L an de s Bi os ci en ce D o N ot D is tri bu te 59 Membrane Protein Insertion in Bacteria from a Structural Perspective Table 1A. SRP Protein/RNA complexe structures PDB R Resolution ID Source Method Description Value [Å] Reference 1DUL E. coli X-ray Domain IV of 4.5 S RNA, 0.199 1.8 Batey et al. domain M of Ffh 200018 1HQ1 E. coli X-ray 4.5S RNA, M-domain 0.151 1.5 Batey et al. of Ffh 2001 67 1QZW S. solfataricus X-ray The complete SRP 54 (Ffh) 0.340 4.1 Rosendal with helix 8 et al. 200316 arrangement is found in the N and G domains of E. coli FtsY (SRα), which has been solved to 2.2Å resolution.19 The structure of the catalytic core (N and G domains) formed by the Ffh/FtsY complex from T. aquaticus has been solved to 1.9 Å resolution in complex with the nonhydrolyzable GTP analog GMP-PCP.20,21 The structures show that Ffh and FtsY form a quasi-two-fold symmetrical hetero-dimer having interaction surfaces both in the N-domain and the G-domain but with the majority of the protein-protein interactions occur between the G-domains (Fig. 2B). Comparison with structures of the uncomplexed proteins shows there are major conformational changes that occur upon formation of the heterodimer. Binding of GTP verses GDP results in small structural adjustments in the free proteins. The structures reveal that the 3' OH of the GTPs are essential for Ffh/FtsY association, activation and catalysis. The structures show that there is a shared composite active site containing the two GTPs at the interface, explaining why the reason why binding of Ffh to FtsY is GTP-dependent and why the complex disassembles after GTP hydrolysis. The structural rearrangement upon complex formation results in bringing catalytic residues in the IBD loop into the active site. The only interactions at the active site between the GTPases occur between the nucleotides. The GTP molecules are aligned head to tail such that the γ-phosphate of each GTP is hydrogen-bonded to the other GTP’s ribose 3' OH group. Hydrolysis of the GTP releases the γ-phosphate. This essentially breaks the contact between the active sites and the GTP substrate and initiates the Ffh/FtsY dissociation. All the three-dimensional structural information of the bacterial and archaeal SRP targeting components currently available are listed in (Table 1A, B, and C). The Signal Recognition Particle Database (SRPDB) (http://psyche.uthct.edu/dbs/SRPDB/SRPDB.html) provides up to date access to alignments of the SRP and SR sequences and phylogenic analysis of these proteins and RNAs. The function of the SRP/FtsY domains become more clear upon structural analysis. Not only do the structures shed light on how the SRP Ffh M domain binds to the signal peptide, but they also deepen our understanding into why Ffh and FtsY respectively acts as each other’s GTPase activating protein. The structures of the Ffh/FtsY (NG domain) complex reveal that Ffh and FtsY interact via the NG domains with the two GTPs forming a composite active site and explains why the targeting of ribosome nascent chain-bound Ffh to FtsY requires GTP (Fig. 5A). The transfer of the nascent membrane protein to the SecY complex cannot take place until Ffh bound to FtsY dissociates from the nascent chain. This only occurs after GTP has been hydrolyzed from Ffh and FtsY. 05Eichler(Dalbey) 1/6/05, 2:19 PM 59 © 20 05 C op yr ig ht Eu re ka h / L an de s Bi os ci en ce D o N ot D is tri bu te Protein Movement Across Membranes 60 Ta bl e 1B . SR P Ff h an f F ts Y P ro te in s tr uc tu re s PD B I D So ur ce M et ho d D es cr ip ti on R V al ue R es ol ut io n [Å ] R ef er en ce 1F FH T. a qu at ic us X -r ay N a nd G d om ai ns o f F fh 0. 18 6 2. 0 Fr ey m an n et a l. 19 97 68 2F FH T. a qu at ic us X -r ay M d om ai n of F fh 0. 25 7 3. 2 K ee na n et a l. 19 98 17 1N G 1 T. a qu at ic us X -r ay N a nd G d om ai ns o f F fh w ith G D P bo un d 0. 18 9 2. 0 Fr ey m an n et a l. 19 99 69 2N G 1 T. a qu at ic us X -r ay N a nd G d om ai ns o f F fh w ith G D P bo un d 0. 20 0 2. 0 Fr ey m an n et a l. 19 99 69 3N G 1 T. a qu at ic us X -r ay N a nd G d om ai ns o f F fh w ith n o G D P bo un d 0. 19 9 2. 3 Fr ey m an n et a l. 19 99 69 1J 8M A . a m bi va le ns X -r ay G d om ai n of F fh 0. 21 9 2. 0 M on to ya e t a l. 20 00 70 1J 8Y A . a m bi va le ns X -r ay G d om ai n of F fh , T 11 2A m ut an t 0. 22 7 2. 0 M on to ya e t a l. 20 00 70 1J PJ T. a qu at ic us X -r ay N a nd G d om ai n of F fh w ith th e no nhy dr ol yz ab le 0. 20 1 2. 3 Pa dm an ab ha n et a l. 20 01 71 G TP a na lo g G M PP N P (N 1 = c ry st al fo rm 1 ) 1J PN T. a qu at ic us X -r ay N a nd G d om ai n of F fh w ith n on -h yd ro ly za bl e 0. 19 0 1. 9 Pa dm an ab ha n et a l. 20 01 71 G TP a na lo g G M PP N P (N 2 = c ry st al fo rm 2 ) 1L S1 T. a qu at ic us X -r ay A po F fh N a nd G d om ai n 0. 13 7 1. 1 R am ir ez e t a l. 20 02 72 1Q Z X S. s ol fa ta ri cu s X -r ay C om pl et e Ff h w ith ou t h el ix 8 0. 31 3 4. 0 R os en da l e t a l. 20 03 16 1O 87 T. a qu at ic us X -r ay N a nd G d om ai n of F fh w ith M gG D P 0. 19 7 2. 1 Fo ci a et a l. 20 04 20 1O K K T. a qu at ic us X -r ay N a nd G d om ai n of F fh in c om pl ex w ith N a nd G 0. 15 6 2. 0 Fo ci a et a l. 20 04 20