ESF - Network on Aspergillus and Aspergillosis

WORKSHOP : Genomics and Post-genomics in Aspergillus fumigatus

Paris, 8 – 11 June 2001

Organizers :

J.P. Latgé (France)

A. Brakhage (Germany)

G. Turner ( United Kingdom)

Local organizing committee : J.P. Latgé and the Aspergillus Unit of the Pasteur Institute

The European Science Foundation awarded a grant to hold a workshop on genomics and post-genomics of Aspergillus fumigatus which was held in June, 2001 near Paris.  This is the report from that workshop.

Programme

Workshop content

What does the Aspergillus fumigatus community expect from the Aspergillus fumigatus genome.

Dr Latgé welcomed the participants and thanked the ESF for their funding.  He overviewed the aims of the meeting which were to learn from other fungal fields especially from Saccharomyces research and to define research options and strategies, especially those which will assist in the understanding of A. fumigatus pathogenicity.  The political issues that have to be considered are the genome efforts that are going on elsewhere, such as the shotgun sequencing being done by private companies (Incyte and Elitra [the meeting subsequently learnt that Incyte have sold their sequence to Elitra]).  The clinical issues that have to be considered are that 1-2/100,000 die from aspergillosis based on autopsy data, diagnosis is often too late, the antifungal drugs in use are not efficient and its pathobiology is not well understood.  It must be remembered that, because A. fumigatus is a saprophytic fungus, it survives in diverse environments and that therefore there is probably a lot of genetic redundancy and a lot of different regulatory systems in its genome (compare with Pseudomonas aeruginosa). 

Using publication data, Dr Latgé illustrated that there is little manpower in the Aspergillus/Neurospora research fields with probably a 1/20th of the total in the Saccharomyces field.  As an example, he explained that the deletion of genes would take 80 years with the present community at a rate of 60-120 per year.  Therefore it was essential to learn what to do from the yeast communities.  For instance, should there be central facilities to serve researchers in the post-genomic era where the options might be to provide transcriptome and proteome services, and bioinformatics support? 

The genome sequence will facilitate the study of gene families, will be useful for microsatellite analysis and will enable comparisons with other genomes to find for instance genes only present in the more pathogenic Aspergillus species such as A. fumigatus and A. flavus.  Comparative studies with yeast will help identify genes present only in filamentous fungi which will be involved in polarity determination, septation, etc.  In the post-genomic era, it should be possible to address the pathogenicity of A. fumigatus.  What are the early steps in the infection process? What antigens are released which might be useful for early detection?  The questions that need to be addressed regarding transcriptome analysis are: should macro- or micro-arrays be used, which genes should be studied - are the most highly expressed the most important?  With regard to proteome analysis, what are the technical issues? How do you study membrane proteins? 

Dr Latgé finished by hoping that some of these questions would be addressed by this workshop.  Another issue that needs to be addressed according to Dr Dixon, is the need to recruit additional researchers and to convince them to study a filamentous organism.  As part of the process, the NIH will be issuing a solicitation for therapeutics and aspergillosis was listed. 

Sequencing of Aspergillus fumigatus: history and future plans

Dr Denning outlined the current status of the A. fumigatus genome sequencing project and stressed the important role that had been played of supportive people in the Wellcome Trust and NIH.  The original plan had been to employ a BAC by BAC clone approach which would have been easier for sharing between centres.  This approach was modified to a chromosome shotgun approach.  When it was clear that the libraries could not be easily constructed, a whole genome shotgun approach was adopted in March 2001. 

Dr Denning detailed the current progress of the pilot project which was taking place principally at the Sanger Centre.  A BAC library has been constructed in Paris as well as two libraries at the Sanger Centre.  These libraries were being used for physical mapping by restriction enzyme fingerprinting.  In addition, HAPPY mapping may be attempted.  BAC end-pair sequences from two of the libraries have also been generated.  The final aim of the pilot project will be to generate enough sequence around the niaD locus to enable a comparison of chromosome organisation with A. nidulans. 

Dr Denning then outlined the current state of the sequencing project, stating that funding for completion had been confirmed from both the Wellcome Trust and the Spanish government.  In addition, the Pasteur will offer assistance from internal funds.  The Spanish funding will start in September and after a review of strategy at the Sanger Centre, the A. fumigatus genome sequencing project has been moved up. The immediate aim now of the genome sequencing project is to obtain 12 x shotgun sequence in the next nine months. 

The TIGR contribution to the sequencing of the genome of Aspergillus fumigatus

Dr Nierman introduced The Institute for Genomic Research (TIGR)’s contribution to the genome sequencing project of A. fumigatus.  Since the decision was made to switch to a whole genome shotgun approach, four production libraries have been constructed and two months of high-throughput sequencing have been performed.  A 4 x assembly is currently available and is posted on the TIGR website.  During sequencing, the aim is to obtain both ends of the clone as this is very important for assembly.  The statistics for the sequencing have been very good (e.g. average read length is 648 bp) and a comparison of the sequences from each library have not revealed any differences, which suggests that they are good quality and that E. coli is able to handle A. fumigatus DNA.  Two additional production libraries (a 50 to 60 kb BAC library and a 15 to 20 kb insert pBR vector-based library) will be constructed.  These larger insert libraries will be crucial in reducing physical gaps, since finishing is 10 times more expensive than production sequencing. 

Dr Feldblyum discussed the current 2 x shotgun assembly which was produced at the end of May.  The statistics are: 100,275 reads, 65 Mb of sequence, 14,000 assemblies which can be grouped into 2299 linked assemblies with the largest at 71 kb.  From these figures, she was able to determine that the genome is around 27 Mb in size. 

Dr Feldblyum then discussed those sequences which were present at higher redundancy than average.  These sequences included the mitochondrial DNA which could be finished with no additional sequencing.  Its size is 32,237 bp, its estimated copy number is five (though this number would be dependant on the method of DNA extraction) and it has a GC ratio of 25 % compared with 50 % for nuclear DNA.  The other sequence present at very high redundancy was the ribosomal DNA repeat, which had been misassembled to 18 kb and which contained both the 18 and 26 S rDNAs. 

Finally, Dr Feldblyum stated that discussions will be taking place with the Sanger Centre next week to work out details such as the level of shotgun sequence coverage.  In the discussion that followed, it was determined that independent assemblies will be performed by the Sanger and TIGR.  The strain which is being sequenced was described (click here for details) including the fact that it potentially contains a double stranded RNA virus.  It was not expected that this viral nucleic acid would be integrated into the genome and so specific efforts would be required to obtain its sequence. 

Genome organisation and predicted pathways of Pneumocystis carinii f. sp. carinii.

Dr Cushion gave an introduction to the biology of the ascomycete Pneumocystis carinii, mentioning such facts as there are one or two species per mammalian host, there is no ergosterol in its membrane and there is a sub-telomeric multigene family of surface antigens.  It cannot be cultured in vitro and it is still the most common opportunistic infection in AIDS even in the era of HAART.  In 1997, it was decided to sequence the 7.7 Mb genome of Pneumocystis carinii f. sp. carinii from the rat and funding was obtained from the NIH. 

Dr Cushion presented data from the EST project, which has been carried out using mRNA from a single infected rat. A computational pipeline for sifting out contaminating rat and bacterial sequences from known Pneumocystis genes and potential genes was built.  It has been possible to distinguish P. carinii sequences from rat and bacterial sequences because of the differences in GC content.  The ESTs have been built up into contigs and Blastx analysis performed on the contigs and singletons.  Apart from hits to P. carinii genes identified in the EST database (91), 1586 potential orthologs were identified from a total of 1677 potential Pneumocystis genes. Most of these had best hits to fungal sequences, especially to Schizosaccharmoyces pombe.  From these data, it has been possible to determine that P. carinii contains most of the genes in the sterol biosynthetic pathway and this was confirmed using biochemical analyses with drugs targeting specific steps. 

Dr Cushion presented data on genome organisation from the first sequenced cosmid, such as the average gene size is 1.6 kb and the number of introns ranged from one to seven. The total number of genes in the rat Pneumocystis genome was estimated to be ~3,740.  Finally she stated that the genome sequencing project was employing a whole chromosome shotgun approach using DNA purified from low melting temperature agarose CHEF gels. Due to low yield, this approach was changed to shotgun sequencing of cosmid contigs assigned to chromosomes. 

The genome sequence of Neurospora crassa.

Dr Schulte described the genomic sequencing of Neurospora crassa that is being carried out in Germany.  In his introduction, he stated that N. crassa has seven chromosomes ranging from 4.0 to 10.3 Mb with a total genome size of 43 Mb.  More than 20,000 ESTs have been generated which have been assembled into >3000 contigs. A whole genome shotgun has been carried out at the Whitehead Institute. German efforts are focused on the sequencing and analysis of linkage groups II and V. A group at the DKFZ, Heidelberg, has carried out the mapping, the sequencing was contracted out to the company MWG Biotech AG and the bioinformatics is being done at MIPS.  The physical mapping involved hybridisation mapping of cosmid and BAC clones to generate a minimal tile with gaps covered using YAC clones. The YAC clones contain the centromeres, but the telomeres are not present in any clones. 

Dr Schulte then described the sequence. The programme FGENESH has been trained, using data on N. crassa codon usage and splice sites, and used to locate ORFs automatically in 14 Mb of sequence - 4700 ORFs have been identified. Manual identification has been performed on 7.5 Mb of sequence and 2404 ORFs located.  An ORF has been found every 3.1 kb which, when extrapolated to the entire genome, gives 13,000 ORFs. The average gene length is 1.6 kb and the average intron size is 105 bp. Protein sequences have been analysed using the PEDANT system and grouped into 6 classes, where class 1 is a known protein and class 6 is a hypothetical protein.  For classes 1 to 3, it is possible to assign a function and these classes represent 42 % of the proteins. Thirty three percent of the predicted proteins have no hits to any other organisms.  A comparison of the hits showed that 1/5th were not present in the yeasts, Saccharomyces cerevisiae and S. pombe, 1/4 were present only in the yeasts and 1/4 were present in both prokaryotes and eukaryotes. The MIPS functional classification system is being applied to the predicted proteins.

The Candida albicans genome

Dr Scherer introduced the Candida albicans genome sequencing project which was carried out at Sanford Genome Technology Center by Ron Davis. Candida albicans has a 16 Mb haploid genome with repeated DNA in large blocks. However, Candida albicans is a diploid organism which complicates the assembly where the distinction has to be made between genes or alleles. The 10.4 x whole genome shotgun has been assembled into 1213 contigs >2 kb where the largest contig is 282 kb and the total size is 17.4 Mb (which is greater than the haploid content). The sequence is currently being assembled into diploid chromosomes.  Most of the genes do not contain an intron.  The current estimate is that there are ~ 6500 genes in the haploid set with 65 - 70 % matching S. cerevisiae and 25 - 30 % having no database match. 

Genes were selected for detailed analysis on the basis that they had a significant database hit but no S. cerevisiae counterpart or that the gene was more closely related to another species rather than S. cerevisiae. These genes could be important in Candida pathogenesis.  588 genes were identified.  For instance, 11 lipases and other enzymes involved in lipid and fatty acid catabolism are not present in yeast.  A similar situation occurs in amino acid and ketone body catabolism. Dr Scherer stated that these catabolic pathways might indicate its source of energy in its natural environment.  The anabolic pathways are similar in Candida and S. cerevisiae with one difference where C. albicans has an additional metabolic capability for the synthesis of cysteine. There are also differences in regulation and in the cell cycle.  For instance, in the anaphase promoting complex, only three out of ten genes are the same as those in S. cerevisiae. Other differences pointed out by Dr Scherer included the presence of three chloride ion channels, two of which are similar to mammalian proteins and one of which is similar to a yeast protein and a calmodulin which is more like that from N. crassa than yeast. Finally Dr Scherer hoped that the genome sequence would encourage researchers to diversify their research interests.

Comparative genomics of yeasts reveal basic mechanisms of eukaryotic evolution

Prof Dujon described a comparative genomic study carried out on 14 different yeast species of the Hemiascomycete group. Thirteen species were selected based on their evolutionary distance from S. cerevisiae, including several Saccharomyces and Kluyveromyces species, Candida tropicalis and other more distantly related species. The sequencing was done at Genoscope using long paired reads of average length 910 bp from 2 to 4 kb insert libraries: two genes on average should therefore be identified. Some species were sequenced to 0.2 x genome coverage and some to 0.4 times. 

Prof Dujon stated that 18,145 genes were identified which were homologous to 4800 S. cerevisiae genes and 593 genes were identified which were not present in S. cerevisiae.  In the most closely related species to S. cerevisiae, the average identity at the amino acid level is 80 % whereas, with the other Saccharomyces species, the average amino acid identity is 60 %. This comparative study enabled a re-evaluation of the total number of genes in S. cerevisiae and the new total is 5651 which includes 50 genes not previously predicted.  Most of these 50 are less than 100 codons long and often have one or two introns.  In answer to a question, Prof Dujon stressed that these small genes do not code for structural RNAs, that they were bordered by a start and stop codon and that everything had to be in frame.  Two of them in addition had a phenotype when deleted. 

A third of the genome of S. cerevisiae consists of yeast specific genes and these genes fall into such functional classes as cell wall metabolism, cell cycle and morphology, and RNA metabolism.  A comparison of pairs of genes showed that the average loss of synteny varied from 2 % in the most closely related species to S. cerevisiae to 90 % in the most distantly related species. Prof Dujon proposed that a possible explanation was segmental duplication followed by gene loss. A consequence of this mechanism would result in the creation of gene families where the numbers would vary between species - 609 S. cerevisiae genes have been expanded in the other yeast species. In conclusion, Prof Dujon stated that comparative genomic sequencing is cheap. However, a good reference species is required and the other species must be reasonably closely related.

Analysis of protein families

Prof Attwood introduced the use of protein family databases by stating that their propose is to assist in assigning function from primary sequence data.  Protein family or pattern databases can be based on single motifs such as PROSITE or multiple motifs such as PRINTS, PROFILES or Pfam.  Single motif databases can be useful in finding very distantly related family members. PROSITE, for instance, uses a regular expression; however, single entries can have more false positives than true positives.  Multiple motif databases can be used to identify proteins which contain only some of the domains and to define families within superfamilies. Those that include gaps are good for finding proteins which vary in amino acid length. 

Prof Attwood then illustrated the use of pattern databases with examples where the annotation given to the sequence was based on the best Blast or Fasta hit - there was no match to the expected protein family. The use of these search programmes is limited and problems can occur because of hits to domains rather than entire proteins and because the annotation of the hits is incorrect. 

Some examples of protein family databases are PROSITE with 1034 entries; PROFILES with 300 (good for superfamilies); PRINTS with 1500; Pfam which is good for divergent domains and superfamilies and BLOCKS which is using information from PROSITE, Pfam and PRINTS to create multiple motifs. Some of these databases have been incorporated into a combined database called InterPro which contains over 3000 entries.  In conclusion, Prof Attwood stated that it was important to understand the method and that no single database is best: none of them is complete and their contents do not entirely overlap.

The BioKnowledge^TM Library: an integrated resource for fungal protein information

Dr Costanzo introduced the BioKnowledge Library of species-specific protein databases which contain sequence and functional genomic data as well as comprehensive reviews of the research literature. These reviews are carried out by PhD qualified curators and overviewed by editors. Each protein has its own page with a descriptive title line. A controlled vocabulary is used to describe function, cellular localisation, processing, etc. Free text annotations under >30 possible headings are provided with links to the relevant references.  Each property of a protein is marked to indicate whether it was experimentally determined or predicted. 

The model organism databases are complete and MycoPathPD™ will be available soon.   MycoPathPD collects all sequence and literature information for the major fungal pathogens of humans: A. fumigatus, A. flavus and A. niger;  P. carinii; Cryptococcus neoformans; Histoplasma capsulatum; Coccidioides immitis; Blastomyces dermatidis; and several Candida species, including C. albicans and C. glabrata. This database contains approximately 11,000 protein reports and >2,800 references were reviewed. Two new cellular roles, adhesion and virulence, have been added in order to capture information relating to interactions between pathogenic fungi and the host. 

Dr Costanzo finished by presenting the data available for A. fumigatus. For instance, 210 papers have been reviewed and 120 proteins have been annotated, which includes 83 derived from sequence, 37 that have been analysed biochemically and 66 that have been experimentally characterised. Mutant phenotypes are given if the gene has been mutated and, as might be expected with the study of a pathogenic organism, there is a bias towards proteins that have been studied for a possible role in pathogenicity and that are extracellular.

Molecular mechanisms of multidrug resistance and the reversal of clinical antifungal drug resistance.

Dr Kuchler introduced the subject of drug resistance and reviewed the role of ABC transporters (ABCTs) in the multidrug resistance phenotype. The yeast S. cerevisiae is used as a model system and for instance, Pdr5 transports over 300 hydrophobic chemicals.  Mutants that are deleted in the PDR5 and SNQ2 genes, are hypersusceptible to many chemicals. Cdr1 and Cdr2 from C. albicans are overexpressed in fluconazole resistant clinical isolates and their expression is induced by antifungals and steroids. Dr Kuchler proposed therefore that one mechanism to overcome antifungal resistance is to inhibit ABCTs.  For instance, FK506 inhibits Pdr5 and GP31 and GP382 reverse P-glycoprotein mediated drug resistance. They have developed an in vivo fluorescence assay using rhodamine 6-G and shown that these inhibitors prevent efflux by Pdr5 and Cdr1.

Genomic studies of Aspergillus fumigatus: genomic tag sequencing and transcriptome analysis.

Dr Glaser started his talk by describing the BAC library that had been constructed in Paris for the Aspergillus fumigatus genome sequencing project. He stated that 9000 clones with an insert size range of 70 to 120 kb, have been stored and so far, 1056 clones have been end-sequenced giving 1937 sequences. 

He then continued with the genomic tag and transcriptome analyses that have been carried out at the Pasteur. Chromosomal DNA was nebulised to create a shotgun library containing 0.6 to 1.1 kb inserts.  9900 clones have been sequenced and 33 % of the sequences have significant Blastx hits. These tags have been used to select probes for creating macroarrays.  Repetitive sequences, such as ribosomal and mitochondrial DNA and the retrotransposon Afut1, were removed. Universal primers were used to amplify the inserts and the products were checked on agarose gels with a 1/10th being re-sequenced. Nylon membranes were used and tested using genomic DNA. To probe with labelled cDNA, 20 to 30 µg of total RNA is required to obtain a reasonable signal.  A final macro-array has been created containing 5600 probes. 

Dr Glaser finished his talk by describing a focussed array containing 95 probes consisting of known and predicted genes involved in cell wall synthesis. Two mutants were studied. Both had been disrupted in two genes important for cell morphology. In comparison with a wild type strain, they were able to observe reproducible changes in expression. Finally in answer to questions, Dr Glaser stated that they re-used the membranes a maximum of three times and considered a greater than two fold difference in signal as significant. 

Transcription profiling of organisms with unsequenced genomes.

In place of Dr Schuren who was unable to attend the meeting, Prof van den Hondel presented this talk. He introduced transcriptome analysis, describing glass micro-arrays and the use of two dye direct comparisons.  He then described how it is possible to create arrays without knowing the complete genomic sequence of an organism. Various types of libraries can be used to obtain clones containing one gene per insert. For instance, genomic libraries with suitably sized inserts can be spotted directly and it is possible to calculate the theoretical number of clones required to obtain a given level of genome coverage. It is possible also to use cDNA libraries. Macro-arrays can be created directly by growing colonies on the filter followed by lysis. Alternatively, the PCR products of inserts can be spotted onto glass slides.  The use of glass slides has the advantage that quantification is simple, but has the disadvantage of being expensive and time consuming. Slides can only be used once.  PCR also has its own problems and will require optimisation. Prof van den Hondel stated that it took six months to sort out their problems with reproducibility, as well as having a faulty batch of glass slides. Finally, he recommended the use of glass slides since, even though they were more expensive, the data were better. 

There followed a discussion about what sort of micro-array facilities were required. Every laboratory does not need one and the use of a centralised facility would resolve problems with reproducibility. The example of the BBSRC in the UK was given where they had set up facilities for both Drosophila and yeast. The NIH are also looking into setting up such a facility.  It was felt that the meeting might need to look to the EU for funding and that companies like Affymetrix might become involved.

Exploring phenotype with the Saccharomyces cerevisiae genome set of gene disruptions.

Prof Bussey demonstrated the use of the S. cerevisiae genome set of gene disruption mutants to screen a phenotype. His laboratory has been interested in the K1 killer toxin produced by a double stranded RNA virus which causes cell lysis by promoting lethal cation loss. Using classical genetics, they were able to isolate 10 resistant mutants over 15 years. 

The disruption collection consists of haploid, homozygous diploid and heterozygous diploid gene knock-out mutants. The heterozygous diploid mutants are disrupted in essential genes and can be used to screen for haplo-insufficiency. Mutants can be checked by PCR since each knock-out is associated with a unique sequence (barcode).  Twelve to fifteen people took 1.5 years to knock out all the genes and one technician in his lab did 600. 

They screened 5500 mutants for altered killer toxin phenotype and identified >200 genes. They were able to identify hypersusceptible as well as resistant mutants, which would not have been possible using classical approaches. The affected genes fell into various classes such as being involved in secretion, cell wall synthesis and stress response. Prof Bussey described in detail one type of resistant mutant, which contained more polymer in their cell wall. These mutants were defective in daughter cell production.

Proteome studies in Aspergillus fumigatus: the example of GPI-anchored proteins.

Dr Bruneau started his talk by describing a family of glycosylphosphatidylinositol (GPI)-anchored proteins (Gel), which are involved in the elongation of glucan chains in the cell wall.  He then stated that the aim of this study was to identify additional GPI-anchored proteins. GPI-anchored proteins were released into the supernatant from a membrane preparation by the action of endogenous phosphatidylinositol-phospholipase C, purified through liquid chromatography, separated on 2-D gels and identified by immunoblot using a specific antibody. Bands were cut out, trypsinised and analysed using MALDI-TOF mass spectrometry. However, it was not possible to identify all proteins by the size of their peptides, since there are few Aspergillus proteins in the public databases and the proteins present in the private databases are mainly incomplete. Amino acid sequences therefore had to be determined using Edman degradation and mass spectrometry. Dr Bruneau stated that they were able to identify all the spots and illustrated some examples. For instance, the major spot was an acid phosphatase, Gel1 and a new member of the family, Gel4, were identified, as well as other proteins homologous to yeast proteins involved in cell wall biosynthesis.  They also used 2-D analysis to compare the protein profiles of wild-type and mutant cells.  The major difference in a mutant disrupted in two chitin synthase genes, was the loss of the acid phosphatase, whereas in a gel1 and gel2 double deletant, most proteins were present at very reduced levels. Dr Bruneau concluded that their hypothesis is that some of the identified proteins would be involved in cell wall biosynthesis.

Proteomics for transmembrane proteins: a practical approach.

Dr Gibrat introduced the subject of transmembrane proteins which comprise about 30 % of yeast and other eukaryotic proteomes. In analyses of the Arabidopsis plasma membrane using 2-D gels, it has been possible to identify 200 of the 700 spots - however none of these proteins are integral. The difficulty with analysing integral membrane proteins using 2-D gels occurs during the iso-electrofocusing (IEF) stage where they aggregate at their pI and so can not be separated during the SDS-PAGE step.  

Dr Gibrat then described the technical approaches they have used to improve the analysis of transmembrane proteins, which involved maintaining the proteins in a native state in micelles and in the liquid phase. Peripheral proteins are removed from the interfaces of membrane vesicles using chaotropic ions and divalent chelators. In addition, incubation for several hours after vesicle permeabilisation allows for the release of peripheral protein entrapped in the lumen. Ion exchange chromatography, which uses the same separation basis as IEF, was tested to separate solubilised integral proteins. In parallel, they were able to use a nitrate transport assay to identify an efflux carrier associated with one of the eluted fractions. The protein is one of a 50 member family, which includes known nitrate and oligopeptide influx carriers. 

Because they were obtaining bands containing more than one protein after ion-exchange, they added an initial gel filtration step. Using the combination of gel filtration, ion-exchange (2D-FPLC) and SDS-PAGE followed by MALDI-TOF mass spectrometry, they have been able to identify 100 proteins from a plasma membrane fraction: 10 % are known and include ion channels, and 70 % are predicted to contain from 1 to 15 < style='font-size:12.0pt; font-family:Symbol'>a< -helices. Finally, Dr Gibrat concluded his talk by stating that they are attempting to increase throughput by automating the chromatography and trypsin digestion steps.

What have we learned using proteome analysis on Saccharomyces cerevisiae

Dr Fey gave an overview of the lessons that have been learnt for 2-D gel electrophoresis and mass spectrometry.  The basic lesson is that you will have to be much more careful than ever before. The use of deletion mutants is preferable and should be compared with the parental strain containing the selectable marker. Growth conditions need to be reproducible and cultures done in triplicate. Cell lysis is carried out in 30 secs so that there is no proteolysis. 

Dr Fey then described the various methods for detecting protein on the 2D gels, dismissing silver and Coomassee staining. Flourescent dyes can be used, though one dye binds to cysteine residues and so will not detect 5 to 10 % of proteins and the other dye is very expensive. The best approach is to use radiolabelled methionine in combination with a phosphoimager to obtain a wide detection range for quantification. Narrow pH range (zoom) gels should be used, for instance 4.5 to 5.5, though only pH 6 to 9 is available for basic separations.  Image analysis of the gels is very labour intensive with manual correction taking hours and acts as the bottleneck in the process. 

For mass spectrometry, Dr Fey stated that they used gels that were not stained or fixed and that they cut out the spots manually, since robots were not as accurate. After trypsin digestion, the molecular weight of the peptides is determined and compared with the databases. Usually 2/3 peptides will be enough for identification. A dedicated person is required to collect all the data. Finally, Dr Fey stated that in their study of deletion mutants, 50 % showed detectable effects with normally 5 to 20 spots being altered. In conclusion, he said that because proteome analysis is slow and expensive, and because you are dealing with a dynamic system (the cell), it is important to get it right, otherwise it is a waste of time.

From gene to function in Saccharomyces cerevisiae.

Dr Rodriguez-Peña described the EUROFAN project to assign function to ORFs of unknown function in the S. cerevisiae genome.  In 1996, 49 % of the ORFs had an unknown function - 758 ORFs were deleted as part of EUROFAN I. At least 80 % of the genes were deleted in two different strain backgrounds. In EUROFAN part II, five consortia were set up, some of which were involved in phenotypic analysis and others in the generation of tools. Functional analysis was divided into nodes related to particular functions of the cell and Dr Rodriguez-Peña described their work on cell wall synthesis and morphogenesis. Four rapid screens were developed which allowed them identify 145 mutants  out of 620, positive in at least one screen. Secondary screens were then used which measured specific cell wall-related activities and altered drug sensitivities. The mutants studied fell into several classes, including, as expected, those involved in cell wall assembly, as well as transcription factors and proteins involved in signal transduction. Dr Rodriguez-Peña concluded by stating that, as a consequence of these types of study, the number of unknowns in yeast had fallen to 31 %.

In the proceeding discussion about the generation of mutants in Aspergillus, the following points were made: the generation of mutant banks should be professionally done in one or two labs; a system needs to be developed to delete complete genes and the genes need to be tagged, so that they can be studied in mixed infections in animal models.

Functional complementation of yeast and other fungal mutants

Prof Sánchez introduced the cell cycle and the regulatory network at the end of the cycle - the mitosis exit network (MEN). Mutants in this network arrest at the same point and lyse, indicating that the genes are essential. Prof Sánchez then described aspects of the mitosis exit network in more detail, a group of proteins recently organized into a pathway. 

This pathway starts with the protein encoded by the TEM1 gene and is modulated by Lte1 and Bub2; this pathway also includes a group of kinases encoded by CDC15 (a protein kinase associated with the spindle pole body (SPB)), DBF2/DBF20 and CDC5, phosphatases such as CDC14, and other proteins, whose function is not very clear, such as Spo12 and Mbo1. Most of these genes are essential, which means that they are attractive targets for the design of anti-fungal drugs.

Prof Sánchez proposed the existence of a new checkpoint at the end of mitosis and mentioned that double mutants in cdc15 and cdc10 (a septin) are not affected by the delay characteristic of the checkpoint . He demonstrated using a GFP fusion, that Cdc15 is localised at the SPB and that this localisation depends on Cnm67 and Cdc14 function. Finally, orthologous genes including C. albicans and A. fumigatus CDC15 and CDC14 have been cloned by either complementation or PCR.

A collection of mutants covering all genes of Saccharomyces cerevisiae

Dr Rose described the EUROSCARF collection centre, which acts as an archive for yeast functional analysis. The centre is funded by handling fees and contains over 26,000 strains which have between them over 5900 ORFs deleted. This centre and the ATCC contain the complete set of the world-wide yeast genome deletion project, which was created by 18 labs and co-ordinated by Ron Davis. Dr Rose described some of the details involved in constructing these mutants, which included adding tags at the 5' and 3' ends, and having 45 bp of homology at the ends for recombination. The procedures were set up to be done in microtitre plates. Most of the ORFs over 100 codons long have been done, but it has not been possible to delete the gene families present at the telomeres because of problems with homologous recombination.  As a consequence, 3 % of the ORFs are not available. Dr Rose stated that 18 % of the genes are essential and that therefore these mutants have had to be constructed as heterozygous diploids. These diploids still have a phenotype. Mutants in non-essential genes (~ 5000) have been constructed in haploid MATa, haploid MAT style='font-size:12.0pt;font-family:Symbol'>a< and homozygous diploid backgrounds. Because there are between 1000 and 1500 labs working on S. cerevisiae, there are problems with the assignment of gene names and so systematic names are used.  Sets are offered in 96 well microtitre plates, containing gaps which act as a contamination control.

Analysis of gene function in Aspergillus nidulans by using promoter exchange.

Prof Turner introduced the various tools used for functional analysis in A. nidulans, including using transformations to disrupt and delete genes and to clone unknown genes by complementation. No active transposons have been found in A. nidulans, but a heterologous system using the transposon Impala from Fusarium has recently been developed as a way of introducing random mutations into the genome. Prof Turner also described the use of REMI to generate mutations, but this procedure can be messy since it can cause re-arrangements, which will result in problems with using sexual crosses to establish if a single locus is involved. 

Prof Turner then discussed the particular problem of studying essential genes. It is possible to inactivate these genes by making use of heterokaryons or asexual diploids, but the most useful approach is to exchange the promoter with one that can be regulated. Prof. Turner described their research using the alcA promoter in the study of the essential gene cotA, which encodes a serine/threonine protein kinase important for hyphal branching. This promoter is repressed in the presence of glucose and they obtained better repression using YEPD plus glucose rather than minimal medium. No mRNA was detectable and they obtained very swollen spores containing many nuclei. Although this promoter is very strong, no detectable differences were observed between the mutant and wild-type when the gene was expressed. 

During the discussion, the point was raised that having access to several different promoters would be preferable. There are also problems with overexpressing proteins in a pathogenic organism, though Prof Brakhage pointed out that they have constructed a weakened Paba ^- and Ura^- A. fumigatus strain for such proposes. 

Generation of gene knock-outs and use of reporter genes in Aspergillus fumigatus

Prof Brakhage introduced the various transformation strategies that have been employed in A. fumigatus. Dominant selection, such as hygromycin resistance, has mainly been used, though prototrophic selection using auxotrophic mutants is also possible. Transformation usually results in ectopic integration, especially if electroporation is used and this has been exploited in the generation of a bank of signature tagged mutants. A gene disruption system has been developed for A. nidulans by d’Enfert, which involves the use of a PCR cassette, lambda Red recombinase and a cosmid containing the gene of interest. They were unable to get the system to work with an A. fumigatus cosmid [this system has subsequently been shown to work in A. fumigatus- click here for protocol details]. 

Prof Brakhage then described the use of reporter genes to study gene expression. Green fluorescent protein and ß-galactosidase have both been used. The gene fusion can either be integrated at the homologous locus, which is important to avoid positional effects, or at the pyrG locus using a specific construct in a pyrG mutant background - this option is useful for studying promoter deletions.  Prof Brakhage presented their work on the regulation of the polyketide synthase gene involved in pigment biosynthesis (pksP). Using a construct consisting of 500 bp of promoter and enhanced GFP, they were able to demonstrate expression in conidia and conidiophores and in hyphae during growth in the lung.  Expression in the hypha was confirmed using a lacZ construct, as it is easier to assay ß-galactosidase activity quantitatively.

Insertional mutagenesis in Aspergillus fumigatus

Dr Brookman described the various approaches that they have attempted to generate gene knockout libraries using insertional mutagenesis. One approach is to use transposons. No active transposons have been found in A. fumigatus and so they used the EZ:TN system based on the bacterial Tn5 mosaic ends. DNA carrying an auxotrophic or drug resistance marker is placed between the Tn5 ends and transformation is carried out in the presence of the transposase. The use of the A. fumigatus pyrG gene was more efficient than using the A. nidulans gene. Transformation was carried out by electroporating swollen spores. Forty six percent of the transformants were normal, 6 % highly branched and 35 % poor spore formers. Of the eight normal transformants investigated so far, all had an insertion at a different site and five were in ORFs. 

To investigate essential genes, Dr Brookman described their use of the parasexual cycle and diploids. Parental strains cnx^- pyrG^- / niaD^- pyrG^- which also contain spore colour markers, were grown on nitrate-containing plates and spores isolated from the green frontier. One in 10⁵ spores were diploid and could be used to create a bank of gene knockout mutants. These mutants were generated by simple insertional mutagenesis. After haploidisation, the selection of those transformants unable to produce haploid progeny enables the identification of strains containing a disrupted essential gene. 

In the proceeding discussion, the following questions regarding transformation in A. fumigatus were raised for which answers need to be sought: what is the minimum length of homology required for recombination; which selection marker should be used - markers such as pyrG or niaD can be used for multiple deletions; what types of regulatable promoters are required and should deletions be carried out in a haploid or diploid strain?

Heterologous expression of fungal secreted and GPI-anchored proteins using Pichia pastoris.

Dr Monod introduced his talk by explaining the need for a good heterologous expression system that provides substantial amounts of protein to enable functional studies to be performed. The system they used involved Pichia pastoris and commercially available vectors which utilised the alcohol oxidase 1 promoter - this strong promoter can result in 50 % of the cellular mRNA being produced from the construct after induction. This system has been used to produce recombinant proteins from Aspergillus and Candida. 

Dr Monod then gave an example of the use of this system. Candida albicans contains ten aspartic proteinases and the secreted proteinases 1 to 6 have been expressed in Pichia. This has permitted an investigation of their pH optima and enabled antibodies to be raised, which have been used to study expression of the proteinases after Candida has been phagocytosed.  The use of the Pichia expression system has enabled an investigation into the prosequence of these proteinases and Dr Monod stated that only 12 amino acids are necessary and sufficient for proper secretion and that there is evidence that the pro-molecule acts between proteins in the maturation process. 

Tagging the allergen repertoire of Aspergillus fumigatus by advanced cloning technologies.

Dr Crameri introduced his talk by stating that complex allergenic sources such as moulds produce many IgE binding proteins. Sera from patients recognise different patterns of bands on protein gels.  He described the experimental system that they have set up involving phage surface-display cDNA libraries. Several rounds of selection with IgE from allergic patients permitted the cloning of several allergens. These proteins were tested against the sera from various patient groups. They were able to identify two intracellular allergens that are only detected by ABPA patient sera and that were not detected in patients sensitised to Aspergillus or in asthmatics. 

Dr Crameri then described their use of high-throughput screening to identify all the allergens present in A. fumigatus. A selectively enriched phage-display library was analysed using robotics to plate and array enriched phagemids. Eighty one IgE binding proteins were identified. Several allergenic sources have been studied in this manner and they have been able to identify the same proteins in different species. One such example is manganese superoxide dismutase.

Progress in heterologous gene expression and extracellular protein production by Aspergillus species and other Ascomycetes/Basidiomycetes production strains.

Prof van den Hondel described the various approaches used to improve the production of homologous and heterologous proteins in filamentous fungal species. The classical approach to strain improvement has used mutagenesis followed by selection. For instance, a neutral cellulase producing strain of Chrysosporium lucknowensi was obtained after UV treatment, which produced 200 times the amount of cellulase. Production can be increased by introducing multiple copies of the gene into the strain and finally five to ten times more protein can be produced during fermentation. They have produced protease deficient mutants in order to limit degradation of the expressed protein.  The promoters that they use for expression include the constitutive promoter gpdA and the inducible promoters alcA, glaA which is induced by starch, and exlA which is induced by xylose and xylan. Because these inducible promoters are highly efficient, some expression occurs during repression. The alcA promoter has the disadvantage that the use of methanol is a safety concern in large scale production.  In general, more than one host is tried and genetic engineering can be combined with classical mutagenesis and selection.  Finally, they have shown that the correct folding of heterologous proteins is very important since otherwise they are degraded. 

Michael J. Anderson, Jean-Paul Latgé September 2001