Proteins or RNA synthesized in one cellular compartment such as the cytoplasm or the nucleus are often transported across membranes to function in another compartment. Thus, the transport process is essential for maintaining normal functions inside the cells. A recent conferenceFootnote 1, “Macromolecular transport across cellular membranes”, focused on the mechanisms of targeting and translocation of proteins or RNA to different cellular compartments, such as the bacterial periplasm, mitochondria, chloroplasts, peroxisomes, the endoplasmic reticulum (ER) and the nucleus. In this meeting, many researchers presented results using genetic, cell biological, biochemical, biophysical and structural approaches to investigate the complex transport processes. Because many basic questions (see below) in this field are structural in nature, it is particularly exciting to witness the progress made from structural research. Since space precludes a complete account of the conference, the following highlights some of the structural insights into protein transport presented at this meeting.

Meeting organization

One unique feature of the conference was that, instead of organizing the sessions in the typical manner according to the targeted locations such as the mitochondrion or ER, each session focused on one of the major questions in the transport process. What components constitute the translocation machinery (translocon)? How is it assembled and what are its structural characteristics? How are the substrates recognized and selected? What are the mechanisms of translocation, and what are the energy sources that drive transport? By presenting results that address the same basic question but in different transport systems within one session, comparisons of approaches and cross fertilization of ideas among researchers working on transport in different cellular compartments were encouraged. The lively discussions after each talk and at the poster sessions attested to the success of this approach.

Target recognition

Particular sequences encoded in transported proteins, called the targeting or signal sequences, are responsible for dictating the final locations of these proteins. One of the early steps in the transport process is the specific recognition of targeting sequences by receptors, usually located in the cytoplasm. In this meeting, several talks and posters presented characterization of the interactions between the targeting sequences and their receptors in the ER, mitochondrial, peroxisomal and the chloroplast transport systems, two of which are highlighted below.

Targeting signals for mitochondrial import are often found at the N-termini of proteins and are cleaved after transport. These targeting sequences share common physicochemical properties rather than specific sequence motifs — they are rich in positively charged residues and can adopt amphipathic helical conformations when bound to detergent micelles. Toshiyo Endo (Nagoya University) presented the NMR structure of a peptide derived from a mitochrondrial targeting sequence in complex with its receptor1 and showed that this peptide indeed forms an amphipathic α-helix in the complex, with the hydrophobic residues mediating the majority of the interactions with the receptor. Surprisingly, the positive charges in the peptide do not contribute significantly to the stability of the complex. Thus, while these charges in the targeting sequences are necessary for import, they do not appear to play important roles in the initial target recognition process.

While target recognition for mitochondrial import does not rely on sequence-specific interactions, recognition of a peroxisomal import signal does. Peroxisomal targeting sequences (PTSs) consist of specific sequence motifs. For example, the type-1 PTS (PTS1) is a C-terminal tripeptide with the sequence Ser-Lys-Leu. To illustrate the structural basis of how PTS1 might be recognized by its receptor Pex5, Steve Gould (Johns Hopkins University) presented a model of the PTS1 binding domain2 based on the structure of a domain from protein phosphatase 5. This model predicts that four Asn residues in the PTS1 binding domain play important roles in recognizing the backbone of PTS1, and that PTS1 binds in an extended conformation. These predictions will be assessed by the structural determination of the Pex5–PTS1 complex that is in progress.

Channels

A crucial component of the translocon is a membrane-spanning channel, through which macromolecules are transported. These channels are integral membrane proteins and are usually associated with regulatory components of the translocon. Therefore identification and physical characterization of these proteins are challenging tasks.

Protein import into a mitochondrion is complicated by the fact that the organelle has two membrane layers called the outer and inner membranes. Matrix (the space enclosed within the inner membrane) or integral inner membrane proteins are transported through translocation complexes that reside in the outer membrane (translocase of the outer membrane, TOM) and the inner membrane (translocase of the inner membrane, TIM) to reach their final locations. Nikolaus Pfanner (Universität Freiburg) presented evidence that purified TIM23 and TIM22, two essential components of the TIM complexes for transporting matrix and integral inner membrane proteins, respectively, exhibit characteristics of a cation channel when reconstituted into lipid bilayers. While these proteins are known to be essential for the translocase activity, these results demonstrate that they do indeed form channels in the membrane. Moreover, the cation selectivity of the TIM23 channel is consistent with the observation that the targeting sequences for the matrix proteins are positively charged.

Two other speakers presented structural characterization of the bacterial translocon, which consists of the SecY, E and G proteins (SecYEG complex). Arnold Driessen (University of Groningen) showed electron micrographs3 of reconstituted SecYEG in the presence of SecA, the ATPase component of the machinery. Single particle analysis of the images revealed a cavity at the center of the complex, supporting its role as a transport channel. Ian Collinson, a research associate in the laboratory of Tom Rapoport (Harvard Medical School/Howard Hughes Medical Institute) presented a projection map of the SecYEG complex at 9 Å resolution, which was calculated from cryo-electron microscopy data collected on two-dimensional crystals of the complex. This map clearly showed features that are consistent with individual α-helices in orientations perpendicular to the membrane. The projection map provides the first view of the arrangement of secondary structural elements within a translocon.

Sandy Simon (The Rockefeller University) presented impressive electrophysiological data on the channel forming activity of pIV, a bacteriophage-encoded protein that is required for the export of mature filamentous phage f1 from bacteria. He showed that a lipid bilayer containing purified pIV allows conductance that is characteristic of a channel through the membrane when stimulated by high voltage. While voltage gating is not likely to be a physiological response of the pIV channel, the observed behavior indicates that opening of the channel is probably regulated in some way. Interestingly, the magnitude of the conducting current is consistent with pIV forming a channel of unprecedented size (>6 nm in diameter!) in the membrane, through which a phage particle (diameter ∼7 nm) could conceivably be exported.

In addition to the physical characterization of the transport channels, several talks and posters also presented data on mechanisms of channel assembly and transport regulation. Novel transport pathways, such as those utilized by cell-penetrating peptides, may exist in parallel to the channels; these pathways were also discussed at the meeting.

Perspective

Macromolecular transport is a complex process that requires cooperation of many components, and the many different approaches used to study this process reflect its complexity. The transport mechanisms depend on the recognition and selection of substrate, as well as the architecture of the transport channels, both of which are structural in nature. This field thus provides rich ground for structural research. Judging from the work presented at this meeting, several systems are now ready for the input of structural data.