LETTERS Surface hydrophobin prevents immune recognition of airborne fungal spores Vishukumar Aimanianda 1 *, Jagadeesh Bayry 2,3,4 *, Silvia Bozza 5 , Olaf Kniemeyer 7 , Katia Perruccio 6 , Sri Ramulu Elluru 2,3,4 , Ce ´cile Clavaud 1 , Sophie Paris 1 , Axel A. Brakhage 7 , Srini V. Kaveri 2,3,4 , Luigina Romani 5 & Jean-Paul Latge ´ 1 The air we breathe is filled with thousands of fungal spores (conidia) per cubic metre, which in certain composting environ- ments can easily exceed 10 9 per cubic metre. They originate from more than a hundred fungal species belonging mainly to the genera Cladosporium, Penicillium, Alternaria and Aspergillus 1–4 . Although these conidia contain many antigens and allergens 5–7 , it is not known why airborne fungal microflora do not activate the host innate immune cells continuously and do not induce detrimental inflammatory responses following their inhalation. Here we show that the surface layer on the dormant conidia masks their re- cognition by the immune system and hence prevents immune response. To explore this, we used several fungal members of the airborne microflora, including the human opportunistic fungal pathogen Aspergillus fumigatus, in in vitro assays with dendritic cells and alveolar macrophages and in in vivo murine experiments. In A. fumigatus, this surface ‘rodlet layer’ is composed of hydro- phobic RodA protein covalently bound to the conidial cell wall through glycosylphosphatidylinositol-remnants. RodA extracted from conidia of A. fumigatus was immunologically inert and did not induce dendritic cell or alveolar macrophage maturation and activation, and failed to activate helper T-cell immune responses in vivo. The removal of this surface ‘rodlet/hydrophobin layer’ either chemically (using hydrofluoric acid), genetically (DrodA mutant) or biologically (germination) resulted in conidial morphotypes inducing immune activation. All these observations show that the hydrophobic rodlet layer on the conidial cell surface immuno- logically silences airborne moulds. Dendritic cells are the sentinels of the immune system controlling fungal immunity 8 . Germinating A. fumigatus conidia induced signifi- cant expression of co-stimulatory molecules (CD80, CD86, CD40 and CD83) and antigen-presenting molecule human leukocyte antigen DR (HLA-DR) on human dendritic cells, and induced the secretion of inflammatory and anti-inflammatory cytokines, indicating that meta- bolically active germinating conidia provide maturation-associated signals to dendritic cells. On the other hand, dormant conidia did not modify the expression of surface molecules or the secretion of cytokines (Supplementary Fig. 2). These results, in agreement with previous studies 9–11 , suggest that dormant conidia, in contrast to germinated conidia, are immunologically inert. However, dormant conidia contain many immunogenic molecules 12 so that on cell wall disruption, the intracellular material of dormant conidia could activate dendritic cells (Supplementary Fig. 3). We then attempted to dissect the reasons for the immunologically inert nature of the dormant conidia. A. fumigatus dormant conidia are covered by a rodlet layer, a thin coating of regularly arranged RodA hydrophobins 13 . The presence of a glycosylphosphatidylinosi- tol (GPI)-anchoring sequence discovered during analysis of the rodA gene (Afu5g09580; Supplementary Fig. 4a) indicates that RodA is covalently bound to the cell-wall polysaccharides 14 . Accordingly, hydrofluoric acid treatment that cleaves phosphodiester bonds of GPI anchors/remnants and releases GPI proteins bound to the cell wall was performed 15 . The hydrofluoric acid extract that was com- pletely water soluble accounted for 1.7% of conidial dry weight and resolved into three bands on SDS–PAGE (Fig. 1a). Mass spectro- metry (MS) and MS/MS analysis showed that these proteins with an apparent mass of 32, 16 and 14.5 kDa on SDS–PAGE corre- sponded to the dimeric form of the native RodA, native RodA (con- sistent with its theoretical mass) and partially degraded or processed RodA (RodA*) (Supplementary Table 1), respectively. We could not observe any RodA released into the culture supernatant when conidia were germinated in various culture media (Fig. 1b, also see ref. 16), indicating the complete degradation of the pre-existing rodlet layer during germination. These observations indicate that the RodA of the outer rodlet layer, covalently bound to the cell wall of the dormant conidia, is degraded during germination, exposing the underlying immunogenic cell wall components usually masked by this rodlet layer. On the basis of the ability of different morphotypes to induce dend- ritic cell activation, we surmised that the rodlet layer on the dormant conidia imparts immunological inertness. To examine this, 5 3 10 5 human dendritic cells were treated with 0.33 mg of RodA (concentra- tion corresponding to 5 3 10 5 conidia). Interestingly, RodA did not induce maturation of dendritic cells (Fig. 1c, d). Even at higher con- centrations of RodA (up to 1 mg), there were no changes in the dend- ritic cell phenotype. Also, RodA neither induced nor altered the basal level of dendritic cell cytokines (Fig. 1e). Thus, the results are remini- scent of interaction of dendritic cells with dormant conidia (Sup- plementary Fig. 2). Moreover and in contrast to 18-kDa ribonuclease (Aspf1, encoded by the gene Afu5g02330), one of the most immuno- genic proteins of A. fumigatus, RodA was unable to stimulate lympho- proliferation (Fig. 1f, g) or to activate Aspergillus-specific human CD4 1 T-cell clones for cytokine production (Fig. 1h). We verified that the lack of activation by RodA was not due to the hydrofluoric acid treatment because this treatment did not alter the immunogeni- city of other Aspergillus proteins such as Aspf1 (Supplementary Fig. 5). Further, to confirm that RodA does not impart tolerogenic properties and immunological unresponsiveness to dendritic cells on encounter with other immunogenic molecules, we treated dendritic cells with a *These authors contributed equally to this work. 1 Unite ´ des Aspergillus, Institut Pasteur, Paris F-75015, France. 2 INSERM, U 872, 3 Centre de Recherche des Cordeliers, Universite ´ Pierre et Marie Curie–Paris 6, UMR S 872, 4 Universite ´ Paris Descartes, UMR S 872, Paris F-75006, France. 5 Department of Experimental Medicine and Biochemical Sciences, 6 Clinical Immunology, Department of Clinical and Experimental Medicine, University of Perugia, Perugia 06122, Italy. 7 Department of Molecular and Applied Microbiology, Leibniz-Institute for Natural Product Research and Infection Biology (HKI) and Friedrich Schiller University, 07745 Jena, Germany. Vol 460 | 27 August 2009 | doi:10.1038/nature08264 1117 Macmillan Publishers Limited. All rights reserved ©2009