# 2007 The Authors Journal compilation # 2008 Blackwell Publishing Ltd doi: 10.1111/j.1600-0854.2007.00683.x Traffic 2008 9: 325���337 Blackwell Munksgaard Characterization of a Listeria monocytogenes Protein Interfering with Rab5a Carmen Alvarez-Dominguez1,*, Fidel Madrazo-Toca1,���, Lorena Fernandez-Prieto1,���, Joel �� Vandekerckhove2, Eduardo Pareja3, Raquel Tobes3, Maria Teresa Gomez-Lopez1, Elida Del Cerro-Vadillo1, Manuel Fresno4, Francisco Leyva-Cobian1 �� and Eugenio Carrasco-Mar�� ��n1,* 1Servicio de Inmunolog��a �� and Instituto de Formacion �� e Investigacion �� Marques �� de Valdecilla (IFIMAV), Hospital Universitario ������Marques �� de Valdecilla������, 39008 Santander, Spain 2Department of Biochemistry, Ghent University and Flanders Interuniversity for Biotechnology, B 9000 Ghent, Belgium 3Bioinformatics Unit, Era7 Information Technologies SL, BIC Granada CEEI, Parque Tecnologico �� de Ciencias de la Salud-Armilla, 18100 Granada, Spain 4 Centro de Biolog��a �� Molecular ������Severo Ochoa������, Universidad Autonoma �� de Madrid, 28049 Madrid, Spain *Corresponding author: Carmen Alvarez-Dominguez, calvarez@humv.es or Eugenio Carrasco-Mar��n, �� deicme@humv.es ���These authors contributed equally to this study. Listeria monocytogenes (LM) phagocytic strategy im- plies recruitment and inhibition of Rab5a. Here, we identify a Listeria protein that binds to Rab5a and is responsible for Rab5a recruitment to phagosomes and impairment of the GDP/GTP exchange activity. This protein was identified as a glyceraldehyde-3-phosphate dehydrogenase (GAPDH) from Listeria (p40 protein, Lmo 2459). The p40 protein was found within the phagoso- mal membrane. Analysis of the sequence of LM p40 protein revealed two enzymatic domains: the nicotin- amide adenine dinucleotide (NAD)-binding domain at the N-terminal and the C-terminal glycolytic domain. The putative ADP-ribosylating ability of this Listeria protein located in the N-terminal domain was examined and showed some similarities to the activity and Rab5a inhibition exerted by Pseudomonas aeruginosa ExoS onto endosome���endosome fusion. Listeria p40 caused Rab5a-specific ADP ribosylation and blocked Rab5a- exchange factor (Vps9) and GDI interaction and function, explaining the inhibition observed in Rab5a- mediated phagosome���endosome fusion. Meanwhile, ExoS impaired Rab5-early endosomal antigen 1 (EEA1) interaction and showed a wider Rab specificity. Listeria GAPDH might be the first intracellular gram-positive enzyme targeted to Rab proteins with ADP-ribosylating ability and a putative novel virulence factor. Key words: ADP-ribosylation, GDP/GTP exchange, glyceraldehyde-3-phosphate dehydrogenase, Listeria, phagocytosis, Rab5a Received 4 June 2007, revised and accepted for publica- tion 29 November 2007, uncorrected manuscript pub- lished online 7 December 2007, published online 2 January 2008 Listeria monocytogenes (LM) is a gram-positive human pathogen that remains for a relatively short time within the phagosomal compartment depending on the cell line. In macrophages, for instance, the average time of bacteria remaining inside the phagosomes is 90 min (1) thereafter, bacteria escape to the cytosol and replicate. During the time LM remains within the phagosome, it modulates the phagosomal composition by targeting Rab5a function and preventing phagosome maturation (2,3). The importance of Rab5 for LM intracellular growth and other infection steps such as entry or vacuolar escape was recently highlighted using small interfering RNA interference tech- nology (4,5). The pathogen produces membrane-active exoproteins within the phagosomes that mediate mem- brane disruption (6). Intracellular bacteria are able to interfere with vesicle trafficking regulators in order to modify the vesicles in which they reside according to the needs of the specific pathogen. In this regard, there are only a few examples of bacteria whose strategy is target small guanosine triphosphatases (GTPases) exchange activities. For instance, Legionella pneumophila protein RalF functions as a guanine nucleotide exchange factor (GEF) for the ADP ribosylation factor family of small GTPases (7). Salmonella typhimurium SopE protein is another example of an intracellular bacterial factor target small GTPase exchange factor for the Rho/Rab family (8). Recently, our group has described the intracellular traffick- ing strategy of LM to delay Rab5a exchange activity (9). However, the molecular mechanism and the bacterial protein(s) participating in this interference are currently unknown. We initiated this study with the hypothesis that LM may present Rab5a-binding proteins responsible for this GTPase interference. In this report, we characterize a LM-Rab5a-binding protein (Lmo 2459) participating in both activities: Rab5a recruitment and inactivation. Results and Discussion A crude LM extract mimicked Rab5a recruitment and exchange delay of pathogenic LM Listeria monocytogenes phagosomal interference is based on Rab5a recruitment and retention in the phagosomes and delay of Rab5a GDP/GTP exchange activity. However, this strategy of pathogenic LM was not observed with heat-killed LM (HKLM) (2,9), suggesting that the putative LM inhibitory factor(s) was heat sensitive and likely has www.traffic.dk 325

a protein enzymatic nature because it required bacterial viability. In this regard, here, we show that a live LM crude extract (LM extract) mimicked both inhibitory actions: Rab5a recruitment and delay of the exchange activity. Rab5a recruitment was examined using recombinant prenylated Rab5a: guanidine dissociation inhibitor (GDI) complexes and control phagosomes, containing HKLM but Rab depleted, because these control phagosomes did not interfere with Rab5a function (2). This protocol was selected for analyzing the effect of LM extract proteins in Rab5a recruitment because it was previously validated to transport Rab5a to subcellular membranes (10,11). The presence of the LM extract increased 15-fold the binding of Rab5a to control phagosomes (right panel in Figure 1A). Heat inacti- vation of this bacterial extract (hi lanes) or elimination of LM- Rab5a-binding proteins (5-nb lanes for Rab5a not bound in Figure 1A) restored the normal binding pattern of Rab5a to control phagosomes. Similarly, the LM extract delayed recombinant Rab5a exchange activity (left panel with tri- angles in Figure 1B), and removal of Rab5a-binding proteins from the LM extract restored the normal exchange kinetics (left panel with open quadrangles in Figure 1B). On the other hand, the LM extract caused no effect with any treat- ment on Rab4 binding to control phagosomes (Figure 1A, left panels), or Rab4 exchange activity (left panel with open quadrangles in Figure 1C), a closely related GTPase sharing effectors and intracellular localization (12). The GTP hydro- lysis activity of both Rabs remained also unchangeable (right Figure 1: Inhibition of Rab5a function with a Listeria extract. A) Isolated ���control phagosomes��� were used for binding of recombinant prenylated Rab5a (right lanes) or Rab4 (left lanes) proteins in the presence (��) or absence ( ) of a LM extract (extract lanes). The 5-nb lanes correspond to samples treated with a LM extract lacking Rab5a-binding proteins, hi lanes with a heat-treated LM extract and treatment lanes with an untreated LM extract. B) Recombinant Rab5a protein was preloaded with [3H]-GDP for exchange reactions (left graphic), and reactions performed in the presence or absence of the following reagents: none or control (circles), LM extract labelled as extract (triangles) or a LM extract lacking Rab5a-binding proteins labelled as Rab5a not bound (quadrangles). Right graphic corresponds to recombinant Rab5a protein preloaded with [g32P]-GTP for hydrolysis reactions performed in the presence or absence of same reagents as for the exchange reactions. C) Recombinant Rab4 protein was preloaded with [3H-GDP] for exchange reaction (left graphic) and with [g32P]- GTP for hydrolysis reaction (right graphic) as in panel B. Both reactions were performed in the presence or absence of the following reagents: none or control (circles), LM extract labelled as extract (white quadrangles) or a LM extract lacking Rab5a-binding proteins labelled as Rab5a not bound (black quadrangles). 326 Traffic 2008 9: 325���337 Alvarez-Dominguez et al.