Robin May

It takes (more than) two to tango

For millennia, bacteria have begged, borrowed and stolen ideas from other organisms to colonize almost every available ecological niche. In particular, pathogenic bacteria, such as Listeria monocytogenes, show an astounding ability to exploit their hosts. After hitching a ride on a dairy product, Listeria invade mammalian cells (no mean feat in itself) and then use host cell proteins to assemble an actin ‘comet tail’ at one pole of the bacterium. Tail assembly propels the bacterium through the cytoplasm and eventually into neighbouring cells, spreading the infection away from the prying eyes of the host’s immune system. As Listeria exploit mammalian mechanisms for actin assembly, they have, in turn, been exploited by researchers as powerful tools in the study of the cytoskeleton. As a result, we now have a comprehensive view of the proteins involved in comet tail assembly and the role of the different bacterial proteins in driving the infection process. Only one bacterial protein, ActA, is required for tail production, meaning that ActA must be solely accountable for the recruitment of host cell proteins that are involved in tail assembly. And yet herein lies a dilemma as ActA has been shown to interact only with the Ena/VASP family of actin-binding proteins, which are themselves useful, but not essential, for tail production.Now, Skoble et al. solve the conundrum by showing that ActA contains three domains (in addition to the Ena/VASP-binding domain), all of which contribute to tail formation1xThree regions within ActA promote Arp2/3 complex-mediated actin nucleation and Listeria monocytogenes motility. Skoble, J. et al. J. Cell Biol. 2000; 150: 527–538Crossref | PubMed | Scopus (124)See all References1. These domains operate in synergy to recruit the Arp2/3 complex (which nucleates actin filaments) and monomeric actin (the ‘raw material’ for tail formation). Intriguingly, two of the three domains show homology with regions that are present in WASP family proteins (the endogenous regulators of the Arp2/3 complex), which is indicative either of genetic ‘theft’ or of convergent evolution. In addition, ActA shows greater redundancy in its ability to trigger actin polymerization through the Arp2/3 complex than the WASP family does – another example of evolutionary ‘fine-tuning’ in response to greater selection pressure on the pathogen than on the host.With Listeria, perhaps more so than any other organism, we come close to a complete understanding of a molecular mechanism. A fully operational computer model for Listeria motility must now be within reach – a prospect that would indeed be a landmark in the field.

Phagocytosis in C. elegans: CED-1 reveals its secrets

During Caenorhabditis elegans development, exactly 131 somatic cells undergo apoptosis, their corpses being engulfed by neighbouring cells. This phenomenon provides a powerful tool to study phagocytosis as mutants that are unable to engulf cell corpses (ced mutants) are readily identifiable. So far, six separate genes required for engulfment have been identified and assigned to two, partially redundant, signal pathways; ced-2, ced-5 and ced-10 in one, ced-1, ced-6 and ced-7 in the other. With five of the six genes cloned, the identification of CED-1 by Zhou et al.1xCED-1 is a transmembrane receptor that mediates cell corpse engulfment in C. elegans. Zhou, Z et al. Cell. 2001; 104: 43–56Abstract | Full Text | Full Text PDF | PubMed | Scopus (311)See all References1 completes the list – at least for now.CED-1 is a transmembrane receptor that binds to cell corpses through its extracellular domain. Intriguingly, this binding requires the function of CED-7, a protein previously shown to resemble an ATP-binding cassette (ABC) transporter and encoded by the only ced gene that functions both in dying and in engulfing cells. The function of CED-7 remains unknown, although the authors suggest that it might present a cell-corpse ligand to CED-1.However, the intracellular domain of CED-1, which shows no overall homology to other proteins but contains an NPXY and a YXXL motif, is perhaps its most fascinating feature. Both motifs are necessary for the function of CED-1 and are partially redundant with each other. NPXY motifs bind to phosphotyrosine-binding (PTB) domains. As CED-6 contains a PTB, and is in the same functional group as CED-1, it is possible that CED-6 relays signals from CED-1, although the authors have so far failed to detect binding between the two proteins. YXXL motifs are phosphorylation sites for tyrosine kinases and are found in mammalian Fcγ-receptors, which mediate engulfment of IgG-coated particles.CED-1 is a fascinating evolutionary link – a single molecule that combines motifs found in separate proteins in higher organisms. Now the question is: how does it work?

Ready-mix protein cookery

Making a protein from scratch takes time. Not only must you produce messenger RNA and (usually) process it to remove introns, but you also have to export it to the cytoplasm and have it translated by a willing ribosome. Thus, one tends to think of a change in gene expression as being a relatively leisurely way of responding to an external signal. This, however, is not always true, as Li et al. describe [1xModulation of an RNA-binding protein by abscisic-acid-activated protein kinase. Li, J. et al. Nature. 2002; 418: 793–797CrossRef | PubMed | Scopus (112)See all References][1].Like ready-to-bake muffin mix, evolution has come up with a masterful shortcut to take the hassle out of protein cookery. And it seems that the first step is to stop throwing your mix away. Li et al. took abscisic-acid-activated protein kinase (AAPK) and screened a bean expression library for binding partners. This fishing expedition caught the humbly-named AAPK-interacting protein 1 (AKIP1)-a protein that is homologous to a class of single-stranded RNA-binding proteins termed hnRNPs (heterogeneous nuclear RNA-binding proteins). GFP-tagged AKIP1 localizes throughout the nucleus but, surprisingly, redistributes to nuclear speckles upon treatment with abscisic acid (ABA). This response to ABA is likely to be mediated by AAPK as this kinase can phosphorylate AKIP1 in vitro, but only if it has been isolated from cells treated with ABA.ABA is produced by plants in response to water shortage, making it likely that downstream targets of ABA are involved in drought resistance. Consequently, Li et al. asked whether AKIP1 could be involved in binding to RNA encoding dehydrins – stress-regulated proteins that stabilize enzymes and membranes. Sure enough, dehydrin mRNA could be bound by phosphorylated AKIP1 but not by inactive AKIP1. The authors point out that environmental stress has long been known to alter RNA longevity in plants. This suggests that RNA for proteins such as the dehydrins is transcribed constitutively, but degraded continuously before translation. During water shortage, ABA production would stimulate the phosphorylation of AKIP1 by AAPK. Active AKIP1 would then bind to dehydrin mRNA, stabilizing it and allowing translation to take place.This work demonstrates once again that posttranscriptional events are a major means of regulating gene expression, as well as hinting at another tool to add to the molecular biologists toolkit. A quick squirt of ABA, and minutes later your transgenic Christmas tree starts to glow in the dark…patent application form, anyone?

Phg1p: Dictyostelium's first ‘fork’ for phagocytosis finger-food

As all those who have tried eating olives with chopsticks know, a good grip is an essential precursor to a fine meal. Strange, then, that Dictyostelium discoideum – a veritable glutton for slippery meals – has kept its molecular cutlery well hidden for so long. For most phagocytes, numerous phagocytic receptors have already been identified and matched to target ligands, ranging from immunoglobulins and serum proteins to complex sugars and lipids. Although Dictyostelium has an eclectic diet, encompassing varied bacteria and even latex beads, the receptor(s) it uses to bind to these moving meals have so far evaded identification.Cornillon et al. now make the first inroad with the characterization of Phg1p, a prime candidate for a phagocytic receptor in this organism 1xPhg1p: A nine-transmembrane protein superfamily member involved in Dictyostelium adhesion and phagocytosis. Cornillon, S. et al. J. Biol. Chem. 2000; 275: 34287–34292Crossref | PubMedSee all References1. A 642-residue protein, Phg1p appears to have the unusual feature of nine transmembrane domains. The phg1 mutant shows a dramatic reduction in the internalization of latex beads and Escherichia coli, although uptake of another bacterium, Klebsiella aerogenes, is affected less severely. Interestingly, changing the surrounding medium to a phosphate buffer partially overcomes the phagocytic defect, suggesting that Dictyostelium relies on a variety of receptors to engulf its prey. The authors propose that some of these (unidentified) receptors may be lectin-based and thus rendered useless in normal growth medium (HL5) by the saturating concentrations of maltose. In phosphate buffer, this receptor (or receptors) can compensate for the absence of Phg1p in the phg1 mutant.Most intriguing is the discovery of genes homologous to PHG1 in a variety of species, ranging from human to yeast and plants. Given that many of these organisms do not produce phagocytic cells, it appears likely that these proteins perform diverse roles. In support of this, Phg1p appears also to be involved in substrate adhesion in Dictyostelium. With this major opening into the field, we can look forward to learning much more about the molecular tools that cells use to grab ‘lunch on the go’.

Seeing the light

If you want to explain an astronomical observation, consider starting with a physical picture of what you think is going on.

It's not how old you are, but how old your T-cells feel

For a biological phenomenon that affects all of us, the process of aging remains surprisingly poorly understood, mainly because of its intrinsic complexity. One of its most obvious features is a significant reduction in one’s ability to fight infection, yet the biological basis for this decline is far from clear. Now Sakata-Kaneko et al.1xAltered Th1/Th2 commitment in human CD4+ T cells with ageing. Sakata-Kaneko, S. et al. Clin. Exp. Immunol. 2000; 120: 267–273Crossref | PubMed | Scopus (67)See all References1 have examined one facet of this process at a molecular level – and come up with a few surprises.The group isolated CD4+ T cells (Th, or ‘helper’ T cells) from two groups of volunteers, one young and one old. These were then analysed for cytokine synthesis or expression of surface (CD) markers. The CD4+ T cell population can be divided into two subsets, Th1 and Th2. Th1 produce cytokines such as interleukin 2 (IL-2) and interferon γ (IFN-γ) and play a central role in regulating cell-mediated immunity (CMI) against intracellular pathogens and tumours. Th2, on the other hand, produce IL-4, -5 and -10 and are primarily concerned with humoral immunity to extracellular parasites.As elderly patients often show defective CMI, Sakata-Kaneko and colleagues predicted an age-related reduction in the Th1 population. In fact, their work clearly demonstrated the opposite. The older patients synthesized more IFN-γ and had a higher proportion of Th1 cells than the younger group. Intriguingly, CD4+ T cells from old patients showed increased IL-2 secretion (a Th1 response) before, as well as after, stimulation. The high basal (pre-stimulation) level of secretion is indicative of a large proportion of previously activated Th cells – so-called ‘memory cells’.Given the enhanced Th1 population in old subjects, why is it that the same patients often show a defect in CMI, a Th1 response? In a lucid discussion, the authors present several explanations. Foremost among them is the concept that, although old patients have more Th1 cells, they might be unable to amplify individual antigen-specific cells in response to antigen exposure, leading to a broad but ineffective response. This might be amplified by a defective effector phase – other groups have already shown age-related deficiencies in natural killer and cytotoxic T-lymphocyte function.This reductionist approach to a complex problem has proven both informative and encouraging. Perhaps growing old won’t turn out to be such a complicated business after all.

Dimers are forever

The dissection of cellular signalling pathways demands ever more precise tools, as research moves from the ‘who’ to the ‘how’ of signalling. It is now clear that many signalling molecules have more than one downstream target, and the coordination of these different ‘arms’ is a major area for investigation.Ortega and colleagues have taken an elegant approach to this problem by using monoclonal antibodies to investigate mast cell responses following activation of the FcϵRI receptor1xLyn dissociation from phosphorylated FcϵRI subunits: a new regulatory step in the FcϵRI signalling cascade revealed by studies of FcϵRI dimer signalling activity. Ortega, E. et al. J. Immunol. 1999; 162: 176–185See all References1. Previous work has demonstrated that activation can be triggered by receptor dimerization and that dimers can be induced by application of antibodies that bind to a single receptor with each Fab domain. In an earlier paper2xOrtega, E., Schweitzer-Stenner, R., and Pecht, I. EMBO J. 1988; 7: 4101–4109See all References2, Ortega et al. described the use of three such antibodies – F4, J17 and H10 – which activate mast cells to varying extents. These authors now show that antibody H10 cannot induce secretion, Ca2+ mobilization or cytoskeletal rearrangement in the RBL-2H3 mast cell line because it activates only one arm of the signal pathway.During normal activation, the FcϵRI receptor associates with the tyrosine kinases Lyn and Syk. The former is involved in receptor tail phosphorylation, and thus signal amplification, whereas the latter phosphorylates numerous downstream targets, including phospholipase Cγ, phosphoinositide 3-kinase and Grb2. Although all three antibodies can crosslink receptors normally and trigger Lyn activation, H10 differs from F4 and J17 in being unable to activate Syk. Unexpectedly, this is not due to different clustering abilities amongst the antibodies but instead occurs because H10-induced crosslinking prevents Lyn from dissociating from the receptor following activation. This presumably impedes Syk from binding to the receptor and becoming activated.This paper clearly demonstrates the need to consider dissociation, as well as association, of signalling intermediates. Fortunately, it also illustrates one attractive way in which this might be achieved.

Going local: rethinking Rho behaviour at the membrane

Radical changes in biology are often so subtle that they pass unnoticed. Cell signalling is one area that has been in need of a concept change for some time – until recently, we all believed that our favourite signalling protein was multifunctional, and that rival candidates were mere imposters! Slowly, though, realization is dawning that the key to understanding signalling lies with localization. Protein X might phosphorylate histones in vitro, but, if X is normally lysosomal in distribution, then it probably is not a histone kinase.One of the great signalling spaghetti junctions is the Rho family of small GTPases, whose numerous members appear to regulate a bewildering plethora of cellular phenomena. Fortunately, some radical concept-changing has been going on here too.It has been known for some time that Rho proteins are recruited to the plasma membrane during activation. Now, however, one group has added an extraordinary dimension to this. Using a combination of immunoblotting and electron-microscopy techniques, Michaely et al.1xPolarized distribution of endogenous Rac1 and RhoA at the cell surface. Michaely, P.A. et al. J. Biol. Chem. 1999; 274: 21430–21436CrossRef | PubMed | Scopus (133)See all References1 have shown that endogenous RhoA and Rac1 are highly abundant in caveolae, and that this enrichment is increased dramatically upon stimulation with platelet-derived growth factor (PDGF) or by ‘loading’ these GTPases with GTPγS (which forces them into an active state).Intriguingly, C3 toxin (which ribosylates and inactivates Rho) does not affect recruitment or targeting of GTPγS-loaded RhoA to caveolae but, rather, prevents it from being retained at this site. Since ribosylated RhoA is inactive, it is tempting to suggest that retention in caveolae is therefore an essential element of Rho activation. Perhaps caveolae act to restrict diffusion along the membrane, pushing the quantity of ‘local’ signal above some critical threshold.The investigation of signalling from a ‘where’ rather than a ‘how’ perspective is a refreshing approach. Perhaps this is one subtle change that we should all adopt at an unsubtle pace.