Many fungal species produce pigments such as melanin, either from L-3,4-dihydroxyphenylalanine (the DOPA-melanin pathway, which is more frequently encountered in Basidiomycetes) or from 1,8-dihydroxynaphthalene (the DHN-melanin pathway, usually found in Ascomycetes and relative Deuteromycetes) . The genes and enzymes involved in these metabolic pathways have been known for many years, but the two types of melanin were only recently related to virulence in phytopathogenic or human pathogenic fungi [12–14]. For example, DHN-melanin provides the rigidity of appressoria, which allow the fungus to penetrate plant leaves, in Magnaporthe grisea, the agent responsible for rice blast , and in Colletotrichum lagenarium, responsible for cucurbits disease . The role of melanin in virulence is less well defined in human pathogens such as Cryptococcus neoformans , Paracoccidioides brasiliensis , Exophiala dermatitidis  and Sporothrix schenckii . It has been demonstrated that this pigment protects the fungal cells especially from reactive oxygen species produced by the host immune defences. Brakhage  and Kwon-Chung  demonstrated the importance of melanin for A. fumigatus. They generated white mutants either by UV mutagenesis, or by targeted mutagenesis. These mutants produced white colonies and had mutations in the PKSP (= ALB1) gene, encoding a polyketide synthase required for conidial pigmentation. They were less virulent than their parent wild-type strains in murine models of disseminated aspergillosis, probably due to an increased susceptibility of their conidia to phagocytosis and reactive oxygen species. However, virulence in mice was not affected by the disruption of the ABR2 gene which is involved in a later step of the melanin pathway .
Mutation in the PKSP (ALB1) gene also led to morphological changes of the conidia. Indeed, SEM showed that these pigmentless mutants produced smooth-walled conidia, whereas the conidia of A. fumigatus have typically a rough surface covered with echinulations . The study of mutant isolates of clinical or environmental origin, with defective melanin biosynthesis pathways, suggests that the pigment also plays an indirect role in virulence of A. fumigatus.
Sequencing of the genes involved in the melanin biosynthesis pathway showed a genetic defect in the early steps of this pathway for our isolates. This was consistent with the changes in colony colour observed for reference strains grown in the presence of specific DHN-melanin inhibitors. Two distinct mutations in the ALB1 gene were detected for IHEM 2508 and 9860 isolates, leading to the production of white powdery colonies; whereas the genetic defect was localised in the ARP2 gene for isolate IHEM 15998, producing brown, powdery colonies. As expected, SEM examination of conidial suspensions from our pigmentless isolates showed a smooth surface. However, a lack of ornamentation was also observed on the conidial surface for the brownish isolate, as well as in reference strains cultivated in the presence of pyroquilon, an inhibitor of the hydroxynaphtalene reductase.
Results from flow cytometry experiments confirmed previous work which suggested that the laminin receptors were located on the ornamentations of the conidial wall. Scanning or transmission electron microscopy, showed that labelling was associated mainly with protrusions of the cell wall [21, 22]. The marked decrease in laminin binding receptors to the surface of conidia of mutant isolates compared to reference strains, together with the smooth-walled appearance of these conidia, strengthens our previous conclusions. Previous work  also suggested the presence of at least two distinct adherence systems on the conidial surface in A. fumigatus: 1) the recognition of fibronectin from its tripeptide sequence Arg-Gly-Asp by two fungal polypeptides of 23 and 30 kDa, and 2) the binding of laminin and fibrinogen by a 72-kDa sialic acid-specific lectin located on the ornamentations of the conidial wall . Our current results also support this hypothesis, showing a slight increase in the fibronectin binding capaCity of mutant isolates compared with reference strains, together with a marked decrease in the binding of laminin to the conidial surface.
The physical properties of the surface of the conidia were also investigated, as they may contribute to host tissue adherence by bringing interacting surfaces closer and mediating their dehydration. We showed that blockage of the melanin biosynthesis pathway resulted in a marked decrease in the electronegative charge of the conidia, a charge which may be due to ionization of free amine and carboxylic acid groups of some surface proteins. A marked decrease in CSH was also observed for conidia of mutant isolates when compared to reference strains, which was consistent with the increased wettability of the colonies. This result suggests that blockage of the melanin pathway also led to the lack of some hydrophobic components on the conidial surface. The defect in melanin in A. fumigatus mutant isolates could also contribute to the marked loss of adherence properties of their conidia , as melanins are hydrophobic molecules and have a negative charge. Youngchim et al.  localised melanin in the electron dense outer layer of the cell wall which surrounds the conidia by TEM examination of A. fumigatus conidia before and after treatment with enzymes and hot acid. Nevertheless, the precise physico-chemical nature of melanin is not well defined and relationships between melanin and other components of the conidial wall, particularly polysaccharides, remain to be clarified [25, 26].
Among the components of the conidial wall are small proteins called hydrophobins which have been described in a large variety of filamentous fungi including A. fumigatus . Hydrophobins share some common properties. These moderately hydrophobic proteins are secreted into the environment by the fungus and they remain in a soluble form when the fungus is cultivated in a liquid medium. However, at an air-liquid interface (e.g. when the fungus is grown on a solid medium), they assemble in about 10-nm thick rodlets organised in bundles or fascicles on the conidial surface, forming a hydrophobic rodlet layer which may be visualised by AFM.AFM examination of the conidial surface showed that this rodlet layer was lacking in mutant isolates whereas typical rodlets were seen on conidia of the tested reference strain. Immunofluorescence or flow cytometry using specific anti-hydrophobin antibodies should be performed to determine whether or not hydrophobins are totally lacking at the conidial surface or simply not organised into a rodlet layer.
Conidia of A. fumigatus may germinate on contact with water. Previous studies showed major changes in the ultrastructure of the conidial wall during the first stage (swelling) of germination. In addition to a marked increase in cell size and the vacuolisation of the cytoplasm, TEM examination of swollen conidia showed changes in the cell wall which became thinner, probably due to the progressive detachment of the outermost cell wall layer . Conidia of mutant isolates and of reference strains were also examined by SEM and AFM using laminin-coated glass coverslips applied to the centre of sporulating cultures. These experiments confirmed the smooth surface of the conidia of mutant isolates and showed the lack of rodlets at their surface. However, this study was conducted on clinical or environmental isolates with defective DHN-melanin pathways and no isogenic wild-type isolates were available as controls, so other mutations, besides those identified in the melanin pathway may have been responsible for phenotypic changes other than colony colour. Nevertheless, the role of melanin in the organisation of the conidial wall was established, because cultivation of reference strains in a medium containing DHN-inhibitors including pyroquilon led to smooth-walled conidia devoid of the outermost electron-dense layer.