Foreword: bacterial homologues of eukaryotic membrane proteins

Bacterial membrane proteins have played a major role in elucidating the structure and function of this important class of proteins. The thematic reviews in this issue of Molecular Membrane Biology focus on key insights provided by those structural analyses of bacterial membrane proteins that extend our understanding of their mammalian homologues. However, what also emerges from this collection of reviews are the notable exceptions where the mammalian structure has driven our understanding of particular families of membrane proteins.

The fluid mosaic model of biological membranes proposed in 1972 [1] depicted membrane proteins as transmembranous blobs, but by 1975, the first indication that the transmembrane domain of bacteriorhodopsin was made up of hydrophobic and amphipathic α-helices was published [2]. A detailed description of the structure of transmembrane helices did not appear until 1985 when the structure of the photosynthetic reaction centre of Rhodopseudomonas viridis became available [3]. At that time the likelihood of obtaining an X-ray structure of a protein of medical relevance (i.e., a mammalian channel, transporter or receptor) seemed remote, but the elucidation of the structures of the mitochondrial complexes in the late 1990s (see [4] for review) and the mammalian calcium pump in 2000 [5] showed that there was nothing inherently intractable about the crystallization of mammalian proteins. It seemed that, provided enough material could be obtained, structures would follow.

In parallel with these approaches the various genome projects were beginning to identify increasing numbers of bacterial homologues of mammalian membrane proteins. Rod MacKinnon had the foresight and courage to attempt to crystallise the potassium channel from S. lividans [6] giving invaluable insight into the mechanism of potassium selectivity by mammalian channels. Crystallizing bacterial homologues has now produced structures of channels, transporters and receptors and provided insight into structures of mammalian counterparts. Furthermore, because bacterial proteins can be altered by directed mutagenesis, and produced in quantity, this has opened up new possibilities for investigating structure and function; which brings us to the thematic review papers in this issue.

Granseth et al. begin with an analysis of why it is currently appropriate to pursue the study of mammalian membrane proteins indirectly through the use of bacterial homologues and also to consider how far the data from prokaryotic membrane protein structures can be usefully extrapolated to provide insights into the structures of their mammalian homologues.

Lemieux has reviewed the contribution that molecular modeling has made to our understanding of the structure and function of the major facilitator superfamily. Four bacterial homologue structures are available and the author has evaluated molecular modelling as a way of providing insight into structure activity of the mammalian homologues for which no high resolution structural information exists. Modelling the mammalian homologues as well as a bacterial homologue of known structure, for validation purposes, reveals that experimental evidence needs to be incorporated into the modelling to arrive at a meaningful structure.

The structures of the CLC ‘chloride channel’ family discussed by Matulef and Maduke show what can be revealed by investigating prokaryotic structures. What was thought to be a family of channels turns out to contain a significant number of Cl⁻/H⁺ antiporters. This led to a reappraisal of the mammalian homologues, originally assigned as channels. It was generally held that transporters and channels would be quite different structures, but this family of transporters/channels is challenging that view.

The funnel shaped structure of the Mg^2 + channel CorA from Thermatoga maritime discussed here by Papp-Wallace and Maguire is the first divalent cation channel structure to be resolved at high resolution. Its structure is quite unlike any transporter or channel resolved so far and the structures appear to be retained across the prokaryote/eukaryote divide. The closest related mammalian homologue of CorA is Mrs2p, a mitochondrial Mg^2 + channel and the CorA structure has provided insights into the structure and function of a family of proteins involved in the transport of multivalent metal ions including Mg^2 + and Al^3 +.

Two recent structures of the ammonia transport family AmtB from E. coli and Amt-1 from Archaeoglobus fulgidus have produced the first glimpses into the function of their mammalian counterparts the Rhesus proteins. Andrade and Einsle have reviewed the recent studies of structure and function of this family of proteins as well as discussing the bioenergetic issues raised, relating to proposed mechanisms of ammonia translocation.

A re-evaluation of the rate of water transport across biological membranes revealed the requirement for water channels, discussed here by Fu and Lu. The first structure to be resolved was that of AQP1 erythrocyte membranes. Several hundred homologues have been identified in mammals, plants and bacteria and prokaryotic homologues have been crystallized and their structures elucidated. This review shows how the structural and sequence information, from a range of sources, has been used to elucidate how these water carrying channels are able to promote the rapid transport of strings of water molecules across the membrane while preventing the passage of protons.

Bramkamp, Altendorf and Greie discuss the structure of the P-type ATPases a disparate family of transporters. The family includes the first mammalian transporter to be resolved at high resolution, the sarcoplasmic reticulum calcium pump, as well as the multi-subunit family member from E. coli, KdpFABC, in which the various functional domains are to be found in separate subunits. In the case of the P-type ATPases much of the insight into the function of these transporters comes from studies of the mammalian structure, but for the present the prokaryotes still provide the best opportunity for studying structure function relationships through mutagenesis.

The homologous structures Sec61 and SecY, the membrane complexes responsible for the insertion of membrane proteins into endoplasmic reticulum and the bacterial plasma membrane respectively, are examined by Gold et al. The review also deals with the structure/function of SecA, responsible for driving post-translational membrane insertion in bacteria, but which has no mammalian counterpart, though a homologue involved in thylakoid protein insertion has been identified in plants.

The studies discussed in these reviews are providing a wealth of information concerning the molecular mechanisms of membrane transport and channel function. The major goal though is the elucidation of the structure and function of the 8,000 or so membrane proteins identified in the human genome. At the time when the first crystal structures of membrane proteins were appearing [3] this idea would have been inconceivable but in 2007 there is a new optimism. However, James Bowie has estimated that, at the predicted rate of progress, it will take at least three decades to obtain sufficient representative structures to be confident enough to assign structures to 90% of membrane proteins [7]. What would accelerate progress is an efficient platform for the expression of mammalian proteins. Perhaps a more thorough understanding of the differences between bacterial and mammalian membrane protein insertion, initiated by the structural information available from SecY and the study of its homologue Sec61 will help in this regard.