WO2001025468A1 - High-throughput screening of expressed dna libraries in filamentous fungi - Google Patents

Abstract

The invention provides a method for the expression of exogenous DNA libraries in mutant filamentous fungi. The fungi are capable of processing intron-containing eucaryotic genes, and also can carry out post-translational processing steps such as glyclosylation and protein folding. The invention provides for the use of fungi with altered morphology, which permits high-throughput screening and directed molecular evolution of expressed proteins. The same transformed fungi may be used to produce larger quantities of protein for isolation, characterization, and application testing, and may be suitable for commercial production of the protein as well.

Description

TT LE

High-Throughput Screening of Expressed DNA Libraries in Filamentous Fungi

SUMMARY OF THE TNVΕNΗON

The invention provides a method for the the expression and subsequent screening of synthetic, genomic, and cDNA libraries in filamentous fungal hosts. The system employs transformed or transfected mutant fungal strains which exhibit a morphology that minimizes or eliminates the formation of entangled mycelia. The fungal strains are also capable of efficient sporulation under submerged growth conditions, and are capable of expressing isolatable quantities of exogenous proteins for evaluation. The mutant fungal strains of the invention are particularly well-suited for high-throughput screening techniques due to their unique morphology and the very low viscosity of their cultures.

BACKGROUND OF THE INVENTION.

Naturally occurring populations of microorganisms exhibit a wide array of physiological and metabolic diversity. Due in part to difficulties in isolating and culturing many microorganisms, a vast number of potentially valuable proteins and polypeptides present in these populations have escaped identification. Indeed, it has been estimated that less than one percent of the world's organisms have been cultured to date. There remains a pressing need for new approaches to the characterization of proteins, polypeptides and metabolites from as-yet uncultivated, unidentified microorganisms, and also from known microorganisms. (The term "protein" as used hereinafter should be understood to encompass polypeptides as well.) There also remains a need for new approaches to the identification and isolation of the genes encoding these proteins, so as to enable the modification and/or production of the proteins.

One approach to this problem has been described by Short in U.S. Patents 5,958,672; 6,001,574, and 6,030,779 (the contents of which are incorporated herein by reference). In this approach, a genomic DNA library is prepared directly from an environmental sample (e.g. a soil sample), with or without making an attempt to isolate or culture any orgnisms that might be present. The genomic library is expressed in E. coli, and the expressed proteins are screened for a property or activity of interest. Short alludes to, but does not describe or enable, the use of fungal host cells in this method.

The approach as described suffers from several serious disadvantages, one of which is that is. coli does not effectively express genes having introns. Rougly 90% of the species of microorganisms in soil are eucaryotes (principally fungi), which generally do have introns in their genomic DNA. Given that there are already about 100,000 species of eumycotan fungi known, and an estimated 1,000,000 yet to be discovered (B.Kendrick, The Fifth Kingdom, Mycologue Publications 1999) the potential for protein and metabolite diversity is far higher among the fungal genomes, but the presence of introns puts most of the fungal protein and metabolite repertoire out of the reach of bacterial expression systems. Furthermore, there are many classes of enzymes (e g , secretory fungal gnin peroxidases and manganese-dependent peroxidases) that are unique to fungi, and there are many fungal enzymes (e g hgnin peroxidases, A mger lnvertase) that are glycosylated. The much greater complexity of fungal genomes, the uniqueness of many fungal proteins, and the glycosylation of many fungal proteins, all suggest that the fraction of protein and metabolite diversity in a given environmental sample that could be actually detected by bacteπal expression of genomic DNA is considerably less than 10%.

Due in part to the spread of AIDS and the rising population of organ transplant recipients, there is a growing number of immune-compromised individuals, and the number and vaπety of fungal infections has grown apace. See Infect. Med. 16:380-382, 385-386 (1999). There is a need to identify and characterize proteins from pathogenic fungi in the ongoing search for new targets for anti-fungal drugs, which requires the capability to screen DNA hbraπes deπved from fungal genomes. Again, the presence of introns in fungal genomes makes expression of genomic DNA hbraπes difficult in most currently available hosts. There has also been a πse in the prevalence of antibiotic-resistant bacteπal infections, creating a need for high-thoughput screening for new fungal metabolites having antibiotic activity.

To avoid problems with introns, it is possible to prepare a cDNA library and express it in bacteπa. However, this approach relies upon the presence of RNA transcπpts, and any genes not actively being transcπbed will not be represented in the library. Many desirable proteins are expressed only under specific conditions (e g , virulence factors in pathogenic fungi) and these conditions may not exist at the time the mRNA is harvested. Furthermore, in order to obtain sufficient RNA to prepare a cDNA library, it is necessary to culture a fair amount of the organism. For organisms in environmental samples that do not grow well in culture, or novel microorganisms for which appropπate culture conditions are unknown, sufficient RNA will not be reliably obtained. In contrast, sufficient genomic DNA can be obtained from a very small number of individual cells by PCR amplification. Clearly it is more desirable to screen a genomic DNA library if at all possible.

Also, E coli is incapable of secretion of many proteins, and thus is undesirable as a host cell for screening purposes where the screening relies upon secretion of the gene product. An additional disadvantage for E coli, and for bacteπal hosts m general, is that prokaryotes cannot provide many of the post-translational modifications required for the activity of numerous eukaryotic proteins. Glycosylation, subunit cleavage, disulfide bond formation, and proper folding of protems are examples of the post- translational processmg often required to produce an active protein. To ensure such processing one can sometimes use mammalian cells, however, mammalian cells are difficult to maintain, require expensive media, and are not generally transformed with high efficiency Such transformation systems are therefore not convenient for high-throughput screening of proteins, although efforts have been made to employ mammalian cells as hosts for cDNA library screening (Schouten et al , WO 99/64582). An approach involving fusion of transformed protoplasts with mammalian cells pπor to library screening has been described (U.S. patent 5,989,814), but expression of the protein library occurs in bacteria or yeast prior to cell fusion. There have been efforts to modify glycosylation patterns enzymatically after expression in host cells (Meynial-Salles and Combes, J. Biotechnol, 46:1-14 (1996)), but such methods must be tailored for specific products and are not suitable for expression of proteins from a DNA library.

The use of yeast as host cells solves some of the above problems, but introduces others. Yeast tend to hyper-glycosylate exogenous proteins (Bretthauer and Castellino, 1999, Biotechnol. Appl. Biochem. 30:193-200), and although yeast are capable of coping with a limited number of introns they are not generally capable of handling complex genes from higher species such as vertebrates. Even genes from filamentous fungi are usually too complex for yeast to transcribe efficiently, and this problem is compounded by differences in expression and splicing sequences between yeast and filamentous fungi (e.g., see M. Innis et al, Science (1985) 228:21-26). Despite these drawbacks, transformation and expression systems for yeast have been extensively developed, generally for use with cDNA libraries. Yeast expression systems have been developed which are used to screen for naturally secreted and membrane proteins of mammalian origin (Klein, et al, 1996 Proc. Natl. Acad. Sci. USA 93:7108-7113; Treco, U.S. patent 5,783,385), and for heterologous fungal proteins (Dalboge and Heldt-Hansen, Mol. Gen. Genet. 243:253- 260 (1994)) and mammalian proteins (Tekamp-Olson and Meryweather, U.S. patent 6,017,731). The term "yeast" as used in the context of yeast expression systems generally refers to organisms of the order Saccharomycetales, such as S. cerevisiae and Pichiapastoris. For the purposes of this disclosure, the terms "fungi" and "fungal" should be understood to refer to Basidiomycetes, Zygomycetes, Oomycetes, and Chythridiomycetes, and Ascomycetes of the class Euascomycetes, which are not of the order Saccharomycetales.

Proper intron splicing, and glycosylation, folding, and other post-translational modifications of fungal gene products would be most efficiently handled by a fungal host species, making filamentous fungi superior hosts for screening genomic DNA from soil samples. It also makes them excellent hosts for the production of fungal enzymes of commercial interest, such as proteases, cellulases, and amylases. It has also been found that filamentous fungi are capable of transcribing, translating, processing, and secreting the products of other eucaryotic genes, including mammalian genes. The latter property makes filamentous fungi attractive hosts for the production of proteins of biomedical interest. For this reason a great deal of effort has been expended on the development of fungal host systems for expression of heterologous proteins, and a number of fungal expression systems have been developed. For reviews of work in this area, see Maras et al, Glycoconjugate J., 16:99-107 (1999); Peberdy, Ada Microbiol. Immunol. Hung. 46:165- 174 (1999); Kruszewsa, Ada Biochim. Pol. 46:181-195 (1999); Archer et al, Crit. Rev. Biotechnol. 17:273- 306 (1997); and Jeenes et al, Biotech. Genet. Eng. Rev. 9:327-367 (1991).

Yelton et al, U.S. Pat. No. 4,816,405, discloses the modification of filamentous Ascomycetes to produce and secrete heterologous proteins. Buxton et al, in U.S. Pat. No. 4,885,249, and in Buxton and Radford, Mol. Gen. Genet. 196:339-344 (1984), discloses the transformation of Aspergillus niger by a DNA vector that contains a selectable marker capable of being incorporated mto the host cells. McKnight et al, U.S. patent. 4,935,349, and Boel, in U.S. patent 5,536,661, disclose methods for expressing eukaryotic genes in Aspergillus involving promoters capable of directing the expression of heterologous genes in Aspergillus and other filamentous fungi. Conneely et al, in U.S. patent 5,955,316, discloses plasmid constructs suitable for the expression and production of lactoferπn in Aspergillus. Cladosporium glucose oxidase had been expressed in Aspergillus, see U.S. patent 5,879,921.

Similar techniques have been used in Neurospora. Lambowitz, m U.S. Pat. No. 4,486,533, discloses an autonomously replicating DNA vector for filamentous fungi and its use for the introduction and expression of heterologous genes in Neurospora. Stuart et al. descπbe co-transformation of Neurospora crassa spheroplasts with mammalian genes and endogenous transcπptional regulatory elements in U.S. patent 5,695,965, and an improved strain of Neurospora having reduced levels of extracellular protease in U.S. patent 5,776,730. Vectors for transformation of Neurospora are disclosed in U.S. patent 5,834,191. Takagi et al. descπbe a transformation system for Rhizopus in U.S. patent 5,436,158. Sismega-Barroso et al. descπbe a transformation system for filamentous fungi in WO 99/51756, which employs promoters of the glutamate dehydrogenase genes from Aspergillus awamori. Dantas-Barbosa et al, in FEMS Microbiol. Lett. 169:185-90 (1998), descπbe transformation of Humicola grisea var. thermoidea to hygromycin B resistance, using either the lithium acetate method or electroporation.

Among the more successful fungal expression systems are those of Aspergillus and Trichoderma, for example as disclosed by Berka et al. in U.S. Patent 5,578,463; see also Devchand and Gwynne, J. Biotechnol. 17:3-9 (1991) and Gouka et al, Appl. Microbiol. Biotechnol. 47:1-11 (1997). Examples of transformed atrains of Myceliophthora thermophila, Acremomum alabamense, Thielavia terrestris and Sporotrichum cellulophilum are presented in WO 96/02563 and U.S. patents 5,602,004, 5,604,129 and 5,695,985, which descπbe certain drawbacks of the Aspergillus and Trichoderma systems and suggest that other fungi may be more suited to large scale protein production.

Methods for the transformation of phyla other than Ascomycetes are known in the art; see for example Munoz-Rivas et al, "Transformation of the basidiomycete, Schizophyllum commune" Mol. Gen. Genet. 205:103-106 (1986); van de Rhee et al, "Transformation of the cultivated mushroom, Agaricus bisporus, to hygromycin B resistance" Mol. Gen. Genet. 250:252-258 (1996); Arnau et al, "Integrative transformation by homologous recombination m the zygomycete Mucor circinelloides" Mol. Gen. Genet. 225:193-198 (1991); Liou et al, "Transformation of a Leu- mutant of Rhizopus niveus with the leuA gene of Mucor circinelloides" Biosci. Biotechnol. Biochem. 56:1503-1504 (1992); and Judelson et al, "Transformation of the oomycete pathogen, Phytophthora infestans" Mol. Plant Microbe Interact. 4:602- 607 (1991). See also de Groot et al, Nature Biotechnol. 16:839-842 (1998), where the bacteπum Agrobaderium tumefaciens was used to transform Agaricus bisporus. In addition to the usual methods of transformation of filamentous fungi, such as for example protoplast fusion, Chakraborty and Kapoor, Nucleic Acids Res. 18:6737 (1990) descπbe the transformation of filamentous fungi by electroporation De Groot et al , in Nature Biotechnol. 16: 839-842 (1998), descπbe Agrobaderium tumefaciens-meάmted transformation of several filamentous fungi. Biolistic introduction of DNA into fungi has been earned out; see for example Christiansen et al , Curr. Genet. 29:100-102 (1995); Durand et al , Curr Genet. 31:158-161 (1997); and Barcellos et α/ , Can J. Microbiol 44:1137-1141 (1998). The use of magnetic particles for "magneto-biolistic" transfection of cells is descπbed in U.S. patents 5,516,670 and 5,753,477, and is expected to be applicable to filamentous fungi.

It is evident that much work has been done to develop expression systems using fungi as hosts. However, the common fungal hosts are all filamentous fungi, which tend to form entangled mats of myceha and highly viscous suspension (submerged) cultures. They are not amenable to micropipetting of suspension cultures into microtiter plates, and are not easily separated into separate clones on a large scale, as would be required in a high-throughput assay system. These properties of filamentous fungi also cause some problems in the mdustπal production of enzymes in fungal host cells. For example, high viscosity and/or the local formation of dense aggregates of mycelium, leads to difficulties m agitation, aeration, and nutπent diffusion.

The influence of fungal morphology on the physical properties of the culture has been recognized, and naturally-occurring strains having more favorable morphology have been identified, as descπbed for example by Jensen and Boominathan m U.S. patent 5,695,985. Homogeneous distπbution of loose mycelium, with pronounced branchmg, was descπbed as a particularly desirable morphology. Schuster and Royer, in mtemational patent application WO 97/26330, suggest a similar method of identifymg fungal cells having more suitable morphology for mdustπal production of heterologous protems. The method compπses screenmg mutants of a parent fungal cell lme, rather than wild-type strains, to find a specific altered morphology, transforming the mutant, and assessing whether a culture of the transformed mutant produces more heterologous protem than the parent cell lme. Mutants with at least 10% greater hyphal branchmg are particulary claimed. The method is illustrated for strains of Fusanum and Aspergillus, and is suggested to be applicable to numerous other genera. The effect of branchmg frequency on culture viscosity of Aspergillus oryzae mutants was examined by Bocking et al, Biotechnol Bioeng. 65:638-648 (1999); more branched strains exhibited lower viscosity in this study. Van Wezel et al, m PCT application WO 00/00613, descπbe methods for reducing the branching and/or enhancing the fragmentation of filamentous microorganisms, whereby the viscosity of the culture is reduced. The method mvolves transforming the microorganisms with the SsgA gene of Streptomyces gnseus. The method is demonstrated in filamentous bacteπa of the order Adinomycetales, but is stated to be applicable to filamentous fungi.

Most pπor efforts m the field of filamentous fungal expression systems have been directed to the identification of strains suitable for mdustπal production of enzymes, and therefore attention has been focused on culture viscosity, stability of transformation, yield of heterologous protein per unit volume, and yield as a percentage of biomass. DNA libraries have been expressed in fungi; see for example Gems and Clutterbuck, Curr. Genet. 1993 24:520-524, where an Aspergillus nidulans library was expressed in A nidulans and Gems et al, Mol. Gen. Genet. 1994242:467-471 where a genomic library from Penicillium was expressed in Aspergillus. Neither of these reports disclosed or suggested screening the expressed proteins; it was through complementation of mutant alleles in the host that the expression of genes from the DNA library was demonstrated. The complementation method requires a specific mutant host for each exogenous protein activity one wishes to detect, and does not provide a tool for general library screening.

The cloning of an Aspergillus niger invertase gene by expression in Trichoderma reesei was described by Berges et al, Curr. Genet. 24:53-59 (1993). Using an A. niger genomic library constructed in a cosmid vector containing a selectable marker, and using as the host T. reesei (which is incapable of utilizing sucrose), an A. niger invertase gene was cloned by a sib selection procedure. Here, again, a very specific characteristic of the host was required to detect the presence of a single expressed exogenous protein, and screening of the genomic library was not disclosed or enabled.

The characteristics of a fungal host cell suitable for expression of a DNA library are different in many respects from the characteristics of hosts suitable for industrial protein manufacture. In general terms, a suitable fungal host for high-throughput screening should meet numerous criteria: The host must be transformed with high efficiency.

- The host must process intron-containing genes and carry out any necessary splicing.

- The host must post-translationally process the expressed protein so that it is produced in an active form. The host should be capable of secretion of the protein.

The host must produce the protein in high enough yield for detection by the assay.

- The host should accept a variety of expression regulatory elements, for ease of use and versatility. The host should permit the use of easily-selectable markers.

The host cell cultures should be of low viscosity.

The host should be deficient in proteases and/or be anemable to suppression of protease expression.

The host must permit screens for a wide variety of exogenous protein activities or properties.

Ideally, the host should secrete only the exogenous protein.

The hyphae in a culture of the host fungus should not be so entangled as to prevent the isolation of single clones, and should not be so entangled as to raise the viscosity to the point of preventing efficient transfer and replication in a miniaturized high throughput screening format (e.g. by micropipeting), and/or

The host should allow the efficient production of spores or other propagules under the growth conditions provided in the high throughput screen.. It would be particularly advantageous if the host also expressed enough heterologous protem to enable isolation and puπfication of the protem. A host cell with this characteπstic would make it possible to further characteπze all heterologous protems of interest merely by culturing the host cells, without time- consuming molecular biological manipulations. It would also be advantageous if the host cell were amenable to ready isolation of the heterologous DNA, so that furthe studies and modifications of the gene itself may be earned out.

In addition to these qualities of the host, the transformation system should also exhibit certrain characteπstics. The transformation frequency should be sufficiently high to generate the numbers of transformants required for meaningful screens. Ideally, expression of the exogenous protem will be mduced by a smgle inducer, by a smgle pathway, acting on a smgle promoter.

To date, no combination of host cells and transformation system has been developed that meets all of these cntena. A need therefore remains for fungal host cell and transformation systems that are capable of efficiently expressing the gene products of a DNA library, especially genomic and/or eucaryotic genomic DNA libraπes.

BRIEF DESCRIPTION OF THE INVENTION.

The present mvention takes advantage of the properties of the transformation system disclosed m mtemational patent applications PCT/NL99/00618 and PCT/EP99/202516. These applications descπbe an efficient transformation system for filamentous fungal hosts such as the genus Chrysosporium and Aspergillus sojae. These applications also disclose that mutant strams are readily prepared which retain all the advantages of the wild-type host cells, but which have partially lost their filamentous phenotype.

The present mvention employs mutant filamentous fungi which do exhibit a less pronounced filamentous phenotype and compact growth morphology, and which produce low-viscosity cultures that are suitable for the physical manipulations mvolved in high-throughput DNA library screenmg. The mvention also provides a transformation system that exhibits high yields of transformants. The mvention also provides libraπes of transformant fungi which efficiently express the protem products of heterologous cDNA inserts, and especially genomic DNA inserts. In another aspect of the mvention, the hbranes of transformed fungi may be used m screenmg for activities or properties of the heterologous protems, or in screenmg for metabolites produced by the transformed fungi as a consequence of exogenous protem activities, or m screenmg for the heterologous DNA or for RNA transcnpts denved therefrom. It will be appreciated that the present mvention also enables high-throughput screenmg for metabolites of non- transformed low-viscosity mutant strams. In yet another aspect of the mvention, the hbranes of transformed fungi may be screened for useful properties of the fungi themselves, such as for example high levels of production of a particular expressed protem. This aspect of the mvention is illustrated by a quantitative assay for the expressed protem of mterest, where the particular transformant having the most favorable combination of protem production, protem processmg, and protem secretion would be detected.

DESCRIPTION OF THE FIGURES

Figure 1 is a Western blot as descnbed m the Examples

Figure 2 is a pUT720 map

Figure 3 is a pUT970G map

Figure 4 is a pUT1064 map

Figure 5 is a pUT1065 map

Figure 6 is a pF6g map

Figure 7 is a pUTl 150 map

Figure 8 is a pUTl 152 map

Figure 9 is a pUTl 155 map

Figure 10 is a pUTl 160 map

Figure 11 is a pUTl 162 map

Figure 12 is the schematic structure of the pclA protein

Figure 13A is a photomicrograph of wildtype Aspergillus sojae

Figure 13B is a photomicrobraph of the Aspergillus sojae pclA mutant

DETAILED DESCRIPTION OF THE INVENπON.

One aspect of the present mvention is directed at a library of low- viscosity filamentous fungi compnsmg nucleic acid sequences, each nucleic acid sequence encodmg a heterologous protem, each of said nucleic acid sequences bemg operably linked to an expression regulating region and optionally a secretion signal encodmg sequence and/or a earner protem encodmg sequence. Preferably a recombinant strain accordmg to the invention will secrete the heterologous protem.

The filamentous fungi of the mvention are charactenzed by the low viscosity of the culture medium. Whereas a typical lndustπal-grade filamentous fungus will produce cultures with viscosities well over 200 centipoise (cP) and usually over 1,000 cP, and can reach 10,000 cP, the fungi of this mvention exhibit a culture viscosity of less than 200 cP, preferably less than 100 cP, more preferably less than 60 cP, and most preferably less than 10 cP after 48 or more hours of culturing m the presence of adequate nutπents under optimal or near-optimal growth conditions. The filamentous fungi of the mvention usually exhibit a morphology charactenzed by short, discrete, non-entangled hyphae or micropellets. Micropellets are slightly entangled or non-entangled collections of hyphae ansmg from a smgle clone, as distinct from pellets which are much larger and are denved from multiple entangled clones. For example, the mutant UV 18-25 Chrysosporium lucknowense strain ( viscosity < 10 cP ) and the morphologically similar mutant Trichoderma longibrachiatum X-252 strain (viscosity < 60 cP ) are characterised by the presence of short, distinct, non en tangled hyphae between 100 and 200 microns in length, and the low viscosity engineered mutant Aspergillus sojae pclA is characterized by a compact form with considerable branching and short hyphae. Whereas the low-viscosity fungi described in WO97/26330 are described as having "more extensive hyphal branching," some fungi of the present invention have equivalent or even slightly reduced hyphal branching when compared to the non-mutant strains. It appears that hyphal length plays the more dominant role in controlling the viscosity of the culture.

Among the preferred genera of filamentous fungi are the Chrysosporium, Thielavia, Neurospora, Acremonium, Tolypocladium, Scytalidium, Sporotrichum, Myceliophthora, Mucor, Aspegillus, Fusarium, Humicola, and Trichoderma, and teleomorphs thereof. More preferred are Chrysosporium, Trichoderma, Aspergillus, and Fusarium. Most preferred are Trichoderma and Chrysosporium. The genus and species of fungi can be defined by morphology consistent with that disclosed in Barnett and Hunter 1972, Illustrated Genera of Imperfect Fungi, 3rd Edition, Burgess Publishing Company. A source providing details concerning classification of fungi of the genus Chrysosporium is Van Oorschot, C.A.N. (1980) "A revision of Chrysosporium and allied genera" in Studies in Mycology No. 20, Centraal Bureau voor Schimmelcultures (CBS), Baarn, The Netherlands, pp. 1-36. According to these teachings the genus Chrysosporium falls within the family Moniliaceae which belongs to the order Hyphomycetales.

Another ready source providing information on fungal nomenclature are the Budapest treaty depositories, especially those providing online databases. The ATCC (US) provides information at http://www.atcc.org, the CBS (NE) at http://www.cbs.knaw.nl, and the VKM (RU) at http://www.bdt.org.br.bdt.msdn.vkm general. Another source is htφ://NT.ars-grin.gov/fungaldatabases. All these institutions can provide teaching on the distinguishing characteristics of fungal species. An alternate taxonomy of the Ascomycota may be found at http://www.ncbi.nlm.nih.gov/htbin- post/Taxonomy/wgetorg?mode=Undef&id=4890. According to this alterante taxonomy, the genus Chrysosporium belongs to family Onygenaceae, order Onygenales, phylum Ascomycota.

The definition of Chrysosporium includes but is not limited to these strains: C. botryoides, C. carmichaelii, C. crassitunicatum, C. europae, C. evolceannui, C. farinicola, C.fastidium, C.filiforme, C. georgiae, C. globiferum, C. globiferum var. articulatum, C. globiferum var. niveum, C. hirundo, C. hispanicum, C. holmii, C. indicum, C. inops, C. keratinophilum, C. kreiselii, C. kuzurovianum, C. lignorum, C. lobatum, C. lucknowense, C. lucknowense Garg 27K, C. medium, C. medium var. spissescens, C. mephiticum, C. merdarium, C. merdarium var. roseum, C. minor, C. pannicola, C. parvum, C. parvum var. crescens, C. pilosum, C. pseudomerdarium, C. pyriformis, C. queenslandicum, C. sigleri, C. sulfureum, C. synchronum, C. tropicum, C. undulatum, C. vallenarense, C. vespertilium, C. zonatum.

C. lucknowense forms one of the species of Chrysosporium that have raised particular interest as it has provided a natural high producer of cellulase proteins (international applications WO 98/15633, PCT/NL99/00618, and U.S. patents 5,811,381 and 6,015,707). Strams with mtemational depository accession numbers ATCC 44006, CBS 251.72, CBS 143.77, CBS 272.77, and VKM F-3500D are examples of Chrysosporium lucknowense strams. Also mcluded within the definition of Chrysosporium are strams deπved from Chrysosporium predecessors including those that have mutated either naturally or by mduced mutagenesis. The methods of the mvention, m one embodiment, employ mutants of Chrysosporium, obtained by a combination of irradiation and chemically-induced mutagenesis, that exhibit a morphology charactenzed by short discrete, non entangled hyphae, and a phenotype characterized by reduced viscosity of the fermentation medium when grown m suspension. In another embodiment, the mvention employs phenotypically similar mutants of Trichoderma. In yet another embodiment the mvention employs phenotypically similar mutants of Aspergillus sojae.

For example, VKM F-3500D ("strain CI") was mutagemsed by subjecting it to ultraviolet light to generate strain UV13-6. This strain was subsequently further mutated with N-methyl-N'-nitro-N- mtrosoguanidine to generate strain NG7C-19. The latter strain m turn was subjected to mutation by ultraviolet light resulting m strain UV18-25 (VKM F-363 ID). During this mutation process the morphological characteπstics vaπed somewhat m culture m liquid or on plates as well as under the microscope. With each successive mutagenesis the cultures showed less of the fluffy and felty appearance on plates that are descπbed as being characteπstic of Chrysosporium, until the colonies attained a flat and matted appearance. A brown pigment observed with the wild type strain m some media was also less prevalent m mutant strams. In liquid culture the mutant UV18-25 was noticeably less viscous than the wild type strain CI and the mutants UV13-6 and NG7C-19. While all strams maintained the gross microscopic characteπstics of Chrysosporium, the mycelia became narrower with each successive mutation and with UV 18-25 distinct fragmentation of the mycelia could be observed. This mycehal fragmentation is likely to be a cause of the lower viscosity associated with cultures of UV18-25. The ability of the strams to sporulate decreased with each mutagemc step. These results demonstrate that a strain may belong genetically to the genus Chrysosporium while exhibiting deviations from the traditional taxonomic (morphological) definitions.

In particular the anamorph form of Chrysosporium has been found to be suited for the screenmg application accordmg to the mvention. The metabolism of the anamorph renders it extremely suitable for a high degree of expression. A teleomorph should also be suitable as the genetic make-up of the anamorphs and teleomorphs is identical. The difference between anamorph and teleomorph is that one is the asexual state and the other is the sexual state; the two states exhibit different morphology under certam conditions.

Another example embodies genetically engeneered mutant strams of Aspergillus sojae. In one of these mutants a specific endoprotease encodmg gene was disrupted. This resulted m a compact growth phenotype exhibiting enhanced branchmg and short hyphae, and the formation of micropellets m submerged cultivation. Moreover, the Aspergillus sojae referred to m this application displays efficient sporulation under specific submerged cultivation conditions, which is a further advantage for its use m a high throughput screenmg system.

It is preferable to use non-toxigenic and non-pathogenic fungal strams, of which a number are known m the art, as this will reduce nsks to the operators and will simplify the overall screenmg process. In a preferred embodiment the fungi will also be protease deficient, so as to minimize degradation of the exogenous protems, and or amenable to suppression of protease production. The use of protease defidient strams as expression hosts is well known, see for example PCT application W096/29391. Protease deficient strams may be produced by screenmg of mutants, or the protease gene(s) may be "knocked out" or otherwise inactivated by methods known m the art. It may be desirable to inactivate other genes m the host filamentous fungus, such as for example those encodmg cellulases and other heavily secreted protems, m order to minimize interference m the assay by host protems. The genes encodmg secreted proteins may be deleted or mutated, or alternatively genes controlling the mduction system or other pathways mvolved in the expession of unwanted protems may be modified m such a way as to reduce such expression. Where an endogenous promoter is employed in the vectors of the mvention (see below), it may be especially desirable to inactivate genes for other protems under control of the same mducer Fungi amenable to suppression of protease secretion are those where protease expression is under the control of a regulatory element that responds to environmental conditions, such that these conditions (e.g., ammo acid concentration) can be manipulated to minimize protease production.

Preferably a homologous expression-regulating region enablmg high expression m the selected host is employed. High expression-regulating regions deπved from a heterologous host, such as from Trichoderma or Aspergillus, are well known m the art, can also be used. By way of example, and not limitation, examples of protems known to be expressed m large quantities and thus providing suitable expression regulating sequences for use m the present mvention are hydrophobm, protease, amylase, xylanase, pectinase, esterase, beta-galactosidase, cellulase (e g. endo-glucanase, cellobiohydrolase) and polygalacturonase.

An expression-regulating region compπses a promoter sequence operably linked to a nucleic acid sequence encodmg the protem to be expressed. The promoter is linked such that the positioning vis-a-vis the initiation codon of the sequence to be expressed allows expression. The promoter sequence can be constitutive but preferably is inducible. Use of an inducible promoter and appropπate mduction media favors expression of genes operably linked to the promoter. Any expression regulating sequence from a homologous species, or from a heterologous strain capable of permitting expression of a protem, is envisaged. The expression regulating sequence is suitably a fungal expression-regulating region, e g. an ascomycete regulating region. Suitably the ascomycete expression regulating region is a regulating region from any of the following genera: Aspergillus, Trichoderma, Chrysosporium.Humicola, Neurospora, Tolypocladium, Fusarium, Penicillium, Talaromyces, or alternative sexual forms thereof such as Emericela and Hypocrea. The cellobiohydrolase promoter from Trichoderma; alcohol dehydrogenase A, alcohol dehydrogenase R, glutamate dehydrogenase, TAKA amylase, glucoamylase, and glyceraldehyde phosphate dehydrogenase promoters from Aspergillus; phosphoglycerate and cross-pathway control promoters of Neurospora; lipase and aspartic proteinase promoter of Rhizomucor miehei; beta-galactosidase promoter of Penicillium canescens;anά cellobiohydrolase, endoglucanase, xylanase, glyceraldehyde-3 -phosphate dehydrogenase A, and protease promoters from Chrysosporium are representative examples. An expression regulating sequence from the same genus as the host strain is preferable, as it is likely to be specifically adapted to the host.

Natural expression-regulating sequences from strains of Chrysosporium which express proteins in extremely large amounts, are particularly preferred. Examples of such strains have been deposited in accordance with the Budapest Treaty with the All Russian Collection (VKM) depository institute in Moscow. Wild type CI strain has the number VKM F-3500 D, deposit date 29-08-1996, CI UV13-6 mutant was deposited with number VKM F-3632 D, and deposit date 02-09-1998, CI NG7C-19 mutant was deposited with number VKM F-3633 D and deposit date 02-09-1998 and CI UV18-25 mutant was deposited with number VKM F-3631 D and deposit date 02-09-1998. These strains are also preferred as sources for the generation of low-viscosity mutants; indeed the VKM F-3631 D strain already exhibits the necessary low viscosity phenotype. A low-viscosity mutant Trichoderma strain, designated X-252, was obtained after two rounds of irradiation of Trichoderma longibrachiatum 18.2KK, which in turn was derived by mutation of the QM 9414 strain of T. longibrachiatum (ATCC 26921). In yet another embodiment the invention employs phenotypically similar mutants of Aspergillus sojae.

Preferably, where the host is a Chrysosporium, a Chrysosporium promoter sequence is applied to ensure good recognition thereof by the host. Certain heterologous expression-regulating sequences also work as efficiently in Chrysosporium as native Chrysosporium sequences. This allows well-known constructs and vectors to be used in transformation of Chrysosporium, and offers numerous other possibilities for constructing vectors enabling good rates of transformation and expression in this host. For example standard Aspergillus transformation techniques can be used as described for example by Christiansen et al in Bio/Technology 6: 1419-1422 (1988). Other documents providing details of Aspergillus transformation vectors, e.g. US patents 4,816,405, 5,198,345, 5,503,991, 5,364,770, 5,705,358, 5,728,547, and 5,578,463, EP-B-215.594 (also for Trichoderma) and their contents are incorporated by reference. As extremely high expression rates for cellulase have been observed in Chrysosporium strains, the expression regulating regions of cellulase genes are particularly preferred.

A nucleic acid construct will preferably comprise a nucleic acid expression regulatory region from Chrysosporium, more preferably from Chrysosporium lucknowense or a derivative thereof, operably linked to a nucleic acid sequence encoding a protein to be expressed. Particularly preferred nucleic acid constructs will comprise an expression regulatory region from Chrysosporium associated with cellulase or xylanase expression, preferably cellobiohydrolase expression, most preferably expression of the 55 kDa cellobiohydrolase (CBH1) described in Table A. Genes containing the promoter sequences of two Chrysosporium endoglucanases (C1-EG6 and C1-EG5; SEQ ID NO: 1 and SEQ ID NO:2, respectively) are provided by way of example. As additional examples, the Chrysosporium promoter sequences of hydrophobin, protease, amylase, xylanase, esterase, pectinase, beta-galactosidase, cellulase (e.g. endoglucanase, cellobiohydrolase) and polygalacturonase are also considered to fall within the scope of the invention.

Any of the promoters or regulatory regions of expression of enzymes disclosed in Table A, for example, can be suitably employed. The nucleic acid sequences of these promoters and regulatory regions can readily be obtained from a Chrysosporium strain. Methods by which promoter sequences can be determined are numerous and well known in the art. Promoter sequences are generally found immediately preceding the ATG start codon at the beginning of the relevant gene. For example, promoter sequences can be identified by deleting sequences upstream of the relevant gene, using recombinant DNA techniques, and examining the effects of these deletions on expression of the gene. Also, for example, promoter sequences can often be inferred by comparing the sequence of regions upstream of the relevant gene with concensus promoter sequences.

For example, the promoter sequences of CI endoglucanases were identified in this manner (see PCT/NL99/00618) by cloning the corresponding genes. Preferred promoters according to the invention are the 55 kDa cellobiohydrolase (CBH1), glyceraldehyde-3-phosphate dehydrogenase A, and the 30 kDa xylanase (XylF) promoters from Chrysosporium, as these enzymes are expressed at high level by their own promoters. The promoters of the carbohydrate-degrading enzymes of Chrysosporium lucknowense in particular, especially C. lucknowense GARG 27K, can advantageously be used for expressing libraries of proteins in other fungal host organisms.

Particular embodiments of nucleic acid sequences according to the invention are known for Chrysosporium, Aspergillus and Trichoderma. Promoters for Chrysosporium are described in PCT/NL99/00618. The prior art provides a number of expression regulating regions for use in Aspergillus, e.g. U.S. patents 4,935,349; 5,198,345; 5,252,726; 5,705,358; and 5,965,384; and PCT application WO93/07277. Expression in Trichoderma is disclosed in U.S. patent 6,022,725. The contents of these patents are hereby incorporated by reference in their entirety.

The hydrophobin gene is a fungal gene that is highly expressed. It is thus suggested that the promoter sequence of a hydrophobin gene, preferably from Chrysosporium, may be suitably applied as expression regulating sequence in a suitable embodiment of the invention. Trichoderma reesei and Trichoderma harzianum gene sequences for hydrophobin have been disclosed for example in the prior art as well as a gene sequence for Aspergillus fumigatus and Aspergillus nidulans and the relevant sequence information is hereby incorporated by reference (Nakari-Setala et al, Eur. J. Biochem. 1996, 235:248-255; Parta et al , Infect. Immun 1994 62:4389-4395; Munoz et al , Curr. Genet. 1997, 32:225-230; and Stringer et al , Mol Microbiol. 1995 16.33-44). Usmg this sequence information a person skilled m the art can obtain the expression regulating sequences of Chrysosporium hydrophobin genes without undue expenmentation following standard techniques such as those suggested above. A recombmant Chrysosporium strain accordmg to the mvention can compπse a hydrophobm-regulating region operably linked to the sequence encodmg the heterologous protem.

An expression regulating sequence can also additionally compπse an enhancer or silencer. These are also well known m the pπor art and are usually located some distance away from the promoter. The expression regulating sequences can also compπse promoters with activator bmdmg sites and repressor bmdmg sites. In some cases such sites may also be modified to eliminate this type of regulation. For example, filamentous fungal promoters in which creA sites are present have been descπbed. The creA sites can be mutated to ensure the glucose repression normally resulting from the presence of the non-mutated sites is eliminated. Use of such a promoter enables production of the library of protems encoded by the nucleic acid sequences regulated by the promoter m the presence of glucose. The method is exemplified in WO 94/13820 and WO 97/09438. These promoters can be used either with or without their creA sites. Mutants in which the creA sites have been mutated can be used as expression regulating sequences in a recombmant strain accordmg to the mvention and the library of nucleic acid sequences it regulates can then be expressed m the presence of glucose. Such Chrysosporium promoters ensure derepression in an analogous manner to that illustrated m WO 97/09438. The identity of creA sites is known from the pπor art. Alternatively, it is possible to apply a promoter with CreA bmdmg sites that have not been mutated m a host strain with a mutation elsewhere in the repression system e.g. m the creA gene itself, so that the strain can, notwithstanding the presence of creA bmdmg sites, produce the library of protems m the presence of glucose.

Terminator sequences are also expression-regulating sequences and these are operably linked to the 3' termini of the sequences to be expressed. A vanety of known fungal terminators are likely to be functional in the host strams of the mvention. Examples are the A. nidulans trpC terminator, A. niger alpha- glucosidase terminator, A. niger glucoamylase terminator, Mucor miehei carboxyl protease terminator (see US 5,578,463), and the Trichoderma reesei cellobiohydrolase terminator. Chrysosporium terminator sequences, e g the EG6 terminator, will of course function well m Chrysosporium.

A suitable transformation vector for use accordmg to the mvention may optionally have the exogenous nucleic acid sequences to be expressed operably linked to a sequence encodmg a signal sequence. A signal sequence is an ammo acid sequence which, when operably linked to the ammo acid sequence of an expressed protem, enables secretion of the protem from the host organism. Such a signal sequence may be one associated with a heterologous protem or it may be one native to the host. The nucleic acid sequence encodmg the signal sequence must be positioned in frame to permit translation of the signal sequence and the heterologous proteins. Signal sequences will be particularly preferred where the invention is being used in conjunction with molecular evolution, and a single, secreted exogenous protein is being evolved.

It will be understood that it is less advanatageous to incorporate a signal sequence in a vector that is to be used to express a library, as this will decrease the probability of expressing the protein of interest. Since genomic libraries are prepared by randomly shearing the DNA and cloning into a vector, the probability that one would obtain an in frame fusion of a gene in the library to the signal sequence is low. Also, even where an in-frame fusion has been obtained, the chosen signal sequence may not work with all genes. For these reasons it is preferable not to employ a signal sequence when screening a genomic DNA library, but rather to screen for the activity of intracelllular exogenous protein. Analysis of the activity of intracellular proteins may be accomplished by pretreating the transformant library with enzymes that convert the fungal cells to protoplasts, followed by lysis. The procedure has been described by van Zeyl et al, J. Biotechnol. 59:221-224 (1997). This procedure has been applied to Chrysosporium to allow colony PCR from Chrysosporium transformants grown in microtiter plates.

Any signal sequence capable of permitting secretion of a protein from a Chrysosporium strain is envisaged. Such a signal sequence is preferably a fungal signal sequence, more preferably an Ascomycete signal sequence. Suitable signal sequences can be derived from eucaryotes generally, preferably from yeasts or from any of the following genera of fungi: Aspergillus, Trichoderma, Chrysosporium, Pichia, Neurospora, Rhizomucor, Hansenula, Humicola, Mucor, Tolypocladium, Fusarium, Penicillium, Saccharomyces, Talaromyces or alternative sexual forms thereof such as Emericella and Hypocrea. Signal sequences that are particularly useful are those natively associated with cellobiohydrolase, endoglucanase, beta-galactosidase, xylanase, pectinase, esterase, hydrophobin, protease or amylase. Examples include amylase or glucoamylase of Aspergillus or Humicola, TAKA amylase of Aspergillus oryzae, α-amylase of Aspergillus niger, carboxyl peptidase of Mucor (US 5,578,463), a lipase or proteinase from Rhizomucor miehei, cellobiohydrolase of Trichoderma, beta-galactosidase of Penicillium canescens CBH1 from Chrysosporium, and the alpha mating factor of Saccharomyces.

Alternatively the signal sequence can be from an amylase or subtilisin gene of a strain of Bacillus. A signal sequence from the same genus as the host strain is extremely suitable as it is most likely to be specifically adapted to the specific host thus preferably the signal sequence is a signal sequence of Chrysosporium. Chrysosporium strains CI, UV13-6, NG7C-19 and UV18-25 secrete proteins in extremely large amounts, and signal sequences from these strains are of particular interest. Signal sequences from filamentous fungi and yeast , are useful, as are signal sequences of non-fungal origin.

A transformed recombinant host fungus according to any of the embodiments of the invention can further comprise a selectable marker. Such a selectable marker permits selection of transformed or transfected cells. A selectable marker often encodes a gene product providing a specific type of resistance foreign to the non-transformed strain. This can be resistance to heavy metals, antibiotics or biocides in general. Prototrophy is also a useful selectable marker of the non-antibiotic vaπety. Auxotrophic markers generate nutntional deficiencies m the host cells, and genes correcting those deficiencies can be used for selection. Examples of commonly used resistance and auxotrophic selection markers are amdS (acetamidase), hph (hygromycin phosphotransferase), pyrG (orotidιne-5 '-phosphate decarboxylase), trpC (anthranilate synthase), argB (ornithme carbamoyltransferase), sC (sulphate adenyltransferase), bar (phosphmothncm acetyltransferase), maD (nitrate reductase), Sh-ble (bleomycin-phleomycin resistance), mutant acetolactate synthase (sulfonylurea resistance), and neomycm phosphotransferase (aminoglycoside resistance). Selection can be earned out by cotransformation where the selection marker is on a separate vector or where the selection marker is on the same nucleic acid fragment as the protem-encoding sequence for the heterologous protem.

A further improvement of the transformation frequency is exemplified by the use of the AMA1 replicator sequence for Aspergillus niger (Verdoes et al , Gene 146:159-165 (1994)). This sequence results in a 10- to 100-fold mcrease m the transformation frequency m a number of different filamentous fungi. Moreover, the introduced DNA is retained autonomously in the fungal cells without integration mto the fungal genome m a multiple copy fashion. These two aspects are expected to be beneficial for its use m the high throughput screenmg method of the present invention, as the non-inbtegrative state excludes vanations in the level of gene expression between different transformants. Moreover, as the introduced DNA is not reco bmed mto the host DNA, no unwanted mutations m the host genome will occur.

As used herein the term "heterologous protem" is a protem or polypeptide not normally expressed and secreted by the host strain used for expression accordmg to the mvention. A heterologous protem may be of prokayotic ongm, or it may be denved from a fungus, plant, msect, or higher animal such as mammals. For pharmaceutical screenmg purposes quite often a preference will exist for human protems, thus a preferred embodiment will be a host wherem the DNA library is of human ongm. Such embodiments are therefore also considered suitable examples of the mvention.

A further embodiment of the mvention mcludes the construction and screenmg of fungal mutant libraπes, and fungal mutant libraπes prepared by the methods disclosed herem. The libraπes may be obtained by transformation of the fungal hosts accordmg to this mvention with any means of lntegrative transformation, usmg methods known to those skilled m the art. This library of fungi based on the preferred host strams may be handled and screened for desired properties or activities of exogenous protems m miniatunzed and/or high-throughput format screenmg methods. By property or activity of mterest is meant any physical, physicochemical, chemical, biological, or catalytic property, or any improvement, mcrease, or decrease m such a property, associated with an exogenous protem of a library member. The library may also be screened for a property or activity associated with a metabolite produced as a result of the presence of exogenous and/or endogenous protems The library may also be screened for fungi producmg increased or decreased quantities of such protem or metabolites.

In another aspect of this mvention, the library of transformed fungi may be screened for the presence of fungal metabolites having desirable properties. It is anticipated that multiple genes or gene clusters may be transferred to the host cells of the mvention, and that non-protem products generated by the action of the encoded enzymes may be generated m the host cells. For example, it has been shown that DNA encodmg the protems necessary for production of lovastatin can be transferred to Aspergillus oryzae (U.S. patent 5,362,638; see also U.S. patent 5,849,541).

The heterologous DNA may be genomic DNA or cDNA, prepared from biological specimens by methods well known in the art. The biological specimen may be an environmental sample (for example, soil, compost, forest litter, seawater, or fresh water), or an extracted, filtered, or centπfuged or otherwise concentrated sample therefrom. Mixed cultures of microorganisms denved from environmental samples may be employed as well. The biological sample may also be denved from any smgle species of organism, such as a cultured microorganism, or plant, msect, or other animal such as a mammal. In addition, the heterologous DNA may be synthetic or semi-synthetic, for example random DNA sequences or DNA compnsmg naturally-occurπng segments which have been shuffled, mutated, or otherwise altered. An example of a semi-synthetic nucleic library is found m Wagner et al, WO 00/0632. DNA from environmental samples (or mixed cultures denved therefrom) will be advantageous for the discovery of novel protems, while the use of DNA from a smgle species will be advantageous m that (1) an appropπate vector may be more judiciously chosen, and (2) the practitioner will be directed to related or similar species for further screenmg if a protem of mterest is identified.

The vectors of the mvention can compπse a promoter sequence denved from a gene encodmg an enzyme, preferably a secreted enzyme. Examples of suitable enzymes from which promoter sequences may be taken are the carbohydrate-degrading enzymes (e.g., cellulases, xylanases, mannanases, mannosidases, pectinases, amylases, e g. glucoamylases , α-amylases, α- and β-galactosidases, α- and β- glucosidases , β-glucanases, chitmases, chitanases), proteases (endoproteases, ammo-proteases, ammo-and carboxy-peptidases), other hydrolases (hpases, esterases, phytases), oxidoreductases (catalases, glucose- oxidases) and transferases (transglycosylases, transglutaminases, isomerases and mvertases). Several examples from Chrysosporium lucknowense are presented in Table A. Table A: Characteristics of selected enzymes from Chrysosporium lucknowense

Notes: molecular weights by MALDI; all others by SDS PAGE xyl = xylanase endo = endoglucanase gal = galactosidase glue = glucosidase CBN = cellbiohydrolase PGU = polygalacturonase

It has been found that Chrysosporium mutants can be made that have reduced expression of protease, thus making them even more suitable for the production of proteinaceous products, especially if the proteinaceous product is sensitive to protease activity. Thus the invention my also employ a mutant Chrysosporium strain which produces less protease than non-mutant Chrysosporium strain, for example less than C. lucknowense strain CI (VKM F-3500 D). In particular the protease acitivity (other than any selective protease intended to cleave a secreted fusion protem) of such strains is less than half the amount, more preferably less than 30% of the amount, and most preferably less than about 10% the amount produced by the CI strain. The decreased protease activity can be measured by known methods, such as by measuring the halo formed on skim milk plates or by bovine serum albumin (BSA) degradation.

Compared to traditional fungal hosts, transformation, expression and secretion rates are exceedingly high when using a Chrysosporium strain exhibiting the mycelial morphology of strain UV18- 25, i.e. short, non-entangled mycelia. Thus a recombmant strain accordmg to the mvention will preferably exhibit such morphology. The invention however also covers non-recombinant strams or otherwise engmeered strams of fungi exhibiting this novel and mventive characteπstic. Another attractive embodiment of the mvention also covers a recombmant Chrysosporium strain exhibiting a viscosity below that of strain NG7C-19, preferably below that of UV 18-25 under corresponding or identical fermenter conditions. We have determined that the viscosity of a culture of UV18-25 is below 10 cP as opposed to that of previously known Trichoderma reesei be g of the order 200-600 cP, and with that of traditional Aspergillus niger bemg of the order 1500-2000 cP under optimal culture conditions during the middle to late stages of fermentation. Accordmgly the mvention may employ any engmeered or mutant filamentous fungus exhibiting this low-viscosity charactersistic, such as the Chrysospoπum UV18-25 (VKM F-3631D) strain, the Tπchoderma X 252 strain, or A. sojae pclA (denved from ATCC 11906).

The fluidity of filamentous fungal cultures can vary over a wide range, from nearly solid to a free- flowing liquid. Viscosity can readily be quantitated by Brookfield rotational viscometry, use of kinematic viscosity tubes, falling ball viscometer or cup type viscometer. Fermentation broths are non-Newtonian fluids, and the apparent viscosity will be dependent to some extent upon the shear rate (Goudar et al, Appl. Microbiol. Biotechnol. 51:310-315 (1999)). This effect is however much less pronounced for the low- viscosity cultures employed m the present mvention.

The use of such low viscosity cultures m the screenmg of an expression library accordmg to the method of the mvention is highly advantageous. The screenmg of DNA hbranes expressed m filamentous fungi has heretofore been limited to relatively slow and labonous methods. In general, once fungi have been transformed (and the transformants optionally selected for), it has been necessary to prepare spores or conidia, or to mechanically disrupt the mycelia, m order to disperse the library of transformed fungi mto individual organisms. This dispersal is necessary so that the separated organisms can be cultured mto clonal colonies or cultures. The spores, conidia, or mycehal fragments are then diluted and "plated out" m standard culture dishes, and the individual colonies are inspected for color, alterations to the substrate, or other detectable indication of the presence of the protem activity or property bemg sought. In another approach, secreted protems are blotted from the colonies onto a membrane, and the membrane is probed or examined for an indication of the presence of the protem activity or property of interest. Use of membranes has proved useful where proteolytic degradation of exogenous protem is a problem, Asgeirsdottir et al , Appl Environ. Microbiol 1999, 65:2250-2252. Such procedures have not proven amenable to automation, and as a result high-throughput screenmg of fungally-expressed protems has not been accomplished with conventional filamentous fungi For purposes of this disclosure, high-throughput screenmg refers to any partially or fully automated method that is capable of evaluating the protem expression of about 10,000 or more transformants per day. The automated high-throughput screenmg of a library of transformed fungi accordmg to the present mvention, accordmgly, may be earned out in a number of known ways. Methods that are known to be applicable to bactena or yeast may m general be applied to the low- viscosity fungi of the present mvention. This is made possible by the low- viscosity phenotype, a consequence of the relatively untangled morphology of the hyphae of the mutant fungi employed. In essence, the mutant fungi behave very much like mdividual bactena or yeast durmg the mechanical manipulations mvolved m automated high-throughput screenmg. This is m contrast to wild-type fungi, and most industπally-adapted fungi as well, which produce highly entangled mycelia which do not permit the ready separation of the mdividual organisms from one another.

For example, a dilute suspension of transformed fungi accordmg to the present mvention may be ahquotted out through a mechanical micropipette mto the wells of a 96-well microplate. It is anticipated that liquid-handling apparatus capable of pipetting mto 384- or 1536- well microplates can also be adapted to the task of automated dispersal of the organisms mto microplates. The concentration of the suspended organisms can be adjusted as desired to control the average number of organisms per well. It will be appreciated that where multiple mdividual organisms are ahquotted mto wells, the identification of the desired protem activity or property in that well will be followed by dilution of the contents of the well and culturing the organisms present mto mdividual clonal colonies or cultures. In this manner the throughput of the system may be increased, at the cost of the need for subsequent resolution of the contents of each well

In an alternative embodiment, a cell sorter may be interposed in the fluid path, which is capable of directing the flow of the culture to the wells of the microplate upon the detection of an organism in the detector cell. This embodiment permits the reasonably accurate dispensation of one organism per well.

In yet another embodiment, colonies growing on solid media can be picked by a robotic colony picker, and the organisms transferred by the robot to the wells of a microtiter plate. Well-separated colonies will give nse to smgle clones m each well.

The dispersed organisms are then permitted to grow mto clonal cultures m the microplate wells. Inducers, nutnents, etc. may be added as desired by the automated fluid dispensmg system. The system may also be used to add any reagents required to enable the detection of the protem activity or property of interest. For example, colorogemc or fluorogenic substrates can be added so as to permit the spectroscopic or fluorometπc detection of an enzyme activity. The low viscosity of the cultures m the wells of a microtiter plate permits the rapid diffusion of such reagents mto the culture, greatly enhancing the sensitivity and reliability of the assay. Diffusion of oxygen and nutπents is also greatly enhanced, facilitating rapid growth and maximal expression and secretion of exogenous peptides. Certam assays, such as the scintillation proximity assay, rely on the diffusion of soluble components so as to arπve at an equi bπum state; agam the low viscosity of the fungal cultures of the present mvention makes this high throughput assay possible. Fmally, m a highly automated system it will be desirable to automatically aspirate or pipette clonal cultures of mterest from their wells m the microtiter plate, and the low viscosity of the cultures will make this possible. All of the above operations would be difficult or impossible given the viscosity of traditional filamenous fungal cultures, especially cultures growmg in the unstirred, shear-free conditions of a microtiter plate well.

In another emodiment, smgle cells are passed through a microfluidic apparatus, and the property or activity of mterest is detected optically (Wada et al , WO 99/67639) Low viscosity is essential to the operation of a microfluidics device, and cultures of the low- viscosity mutant fungi of the present mvention are expected to be amenable to microfluidic manipulation.

Another class of high-thoughput screens is by photometnc analysis, by digital imaging spectroscopy, of large numbers of mdividual colonies growmg on a solid substrate. See for example Youvan et al , 1994, Meth Enzymol 246:732-748. In this method, changes m the overall absorption or emission spectra of specialized reagents are indicative of the presence of a heterologous protem activity or property of mterest. The ready dispersal of mdividual organisms attendant upon the use of low- viscosity mutants also enables the use of filamentous fungi m this method The tendency for colonies of the mutant fungi of the mvention to exhibit less lateral growth, and to produce smooth, compact, and well-defined colonies, is also advantageous m such a screenmg system. Furthermore, the supenor expression and secretion charactenstics of fungi as compared to bactena provide greater quantities of protem for spectral analysis.

An automated microorganism handlmg tool is descnbed m Japanese patent application publication number 11-304666 This device is capable of the transfer of microdroplets contammg mdividual cells, and it is anticipated that the fungal strams of the present mvention, by virtue of their morphology, will be amenable to micromanipulation of mdividual clones with this device.

An automated microbiological high-throughput screenmg system is descπbed m Beydon et al, J Biomol. Screening 5.13-21 (2000). The robotic system is capable of transferring droplets with a volume of 400 nl to agar plates, and processmg 10,000 screenmg points per hour, and has been used to conduct yeast two-hybnd screens. It is anticipated that the fungal hosts of the present mvention will be as amenable as yeast to high-throughput screenmg with systems of this type.

The development of high throughput screens in general is discussed by Jayawickreme and Kost, Curr Opin Biotechnol 8.629-634 (1997) A high throughput screen for rarely transcnbed differentially expressed genes is descnbed in von Stem et al , Nucleic Acids Res. 35: 2598-2602 (1997)

The Chrysosporium strain UV18-25 and the Trichoderma strain X 252 illustrate vaπous aspects of the mvention exceedmgly well. The mvention however may employ other mutant or otherwise engmeered strams of filamentous fungi that exhibit low viscocity m culture. The specific morphology of the fungi may not be cπtical; the present inventors have observed short, non-entangled mycelia m these two strams but other morphologies, such as close and extensive hyphal branchmg, may also lead to reduced viscosity. Fungal strams accordmg to the mvention are preferred if they exhibit optimal growth conditions at neutral pH and temperatures of 25-43°C. Such screenmg conditions are advantageous for maintaining the activity of exogenous protems, m particular those susceptible to degradation or lnactivation at acidic pH. Most mammalian protems, and human protems in particular, have evolved to function at physiological pH and temperature, and screenmg for the normal activity of a human enzyme is best earned out under those conditions. Protems mtended for therapeutic use will have to function under such conditions, which also makes these the preferred screenmg conditions. Chrysosporium strams exhibit precisely this charactenstic, growmg well at neutral pH and 35-40 °C, while other commonly employed fungal host species (e.g. Aspergillus and Trichoderma) grow best at acidic pH and may be less suitable for this reason

Another application of the method of the present invention is m the process of "directed evolution," wherem novel protein-encoding DNA sequences are generated, the encoded protems are expressed m a host cell, and those seqences encodmg protems having a desired charactenstic are mutated and expressed agam. The process is repeated for a number of cycles until a protem with the desired charactenstics is obtained Gene shuffling, protem engineering, error-prone PCR, site-directed mutagenesis, and combmatonal and random mutagenesis are examples of processes through which novel DNA sequences encodmg exogenous protems can be generated. U.S. patents 5,223,409, 5,780,279 and 5,770,356 provide teaching of directed evolution. See also Kuchner and Arnold, Trends in Biotechnology, 15:523-530 (1997); Schmidt-Dannert and Arnold, Trends in Biotech , 17.135-136 (1999); Arnold and Volkov, Curr Opin. Chem Biol , 3-54-59 (1999); Zhao et al , Manual of Industrial Microbiology and Biotechnology, 2^nd Ed., (Dema and Davies, eds.) pp. 597-604, ASM Press, Washington DC, 1999; Arnold and Wintrode, Encyclopedia ofBwprocess Technology Fermentation, Biocatalysis, and Bioseparation, (Flickinger and Drew, eds.) pp. 971-987, John Wiley & Sons, New York, 1999; and Minshull and Stemmer, Curr Opin Chem Biol. 3:284-290.

An application of combmatonal mutagenesis is disclosed m Hu et al , Biochemistry 1998 37:10006-10015. US 5,763,192 descnbes a process for obtaining novel protem-encodmg DNA sequences by stochastically generating synthetic sequences, introducing them mto a host, and selecting host cells with the desired charactenstic. Methods for effecting artificial gene recombination (DNA shuffling) include random priming recombination (Z. Shao, et al , Nucleic Acids Res., 26:681-683 (1998)), the staggered extension process (H. Zhao et al , Nature Biotech , 16.258-262 (1998)), and heteroduplex recombination (A. Volkov et al , Nucleic Acids Re , 27:e 18 ( 1999)) Error-prone PCR is yet another approach (Song md R ee, Appl Environ Microbiol. 66-890-894 (2000))

There are two widely-practiced methods of carrying out the selection step m a directed evolution process. In one method, the protem activity of mterest is somehow made essential to the survival of the host cells. For example, if the activity desired is a cellulase active at pH 8, a cellulase gene could be mutated and mtroduced mto the host cells. The transformants are grown with cellulose as the sole carbon source, and the pH raised gradually until only a few survivors remain. The mutanted cellulase gene from the survivors, which presumably encodes a cellulase active at relatively high pH, is subjected to another round of mutation, and the process is repeated until transformants that can grow on cellulose at pH 8 are obtained. Thermostable vaπants of enzymes can likewise be evolved, by cycles of gene mutation and high- temperature culturing of host cells (Liao et al, Proc. Natl. Acad. Sci. USA 83:576-580 (1986); Giver et al, Proc. Natl. Acad. Sci. USA. 95:12809-12813 (1998).

The chief advantage of this method is the massively parallel nature of the "survival of the fittest" selection step. Millions, or billions, of unsuccessful mutations are simultaneously eliminated from consideration without the need to evaluate them individually. However, it is not always possible to link an enzyme activity of mterest to the survival of the host. Where the desired protem property is selective bmdmg to a target of mterest, making the bmdmg property essential to survival is especially difficult. Furthermore, survival under forced conditions such as high temperature or extreme pH is likely to be dependent upon multiple factors, and a desirable mutation will not be selected for and will be lost if the host cell is unable to survive for reasons unrelated to the properties of the mutant protem.

An alternative to the massively parallel "survival of the fittest" approach is seπal screenmg. In this approach, mdividual transformants are screened by traditional methods, such as observation of cleared or colored zones around colonies growmg on mdicator media, coloπmetπc or fluorometπc enzyme assays, lmmunoassays, bmdmg assays, etc. See for example Joo et al, Nature 399:670-673 (1999), where a cytochrome P450 monooxygenase not requirmg NADH as a cofactor was evolved by cycles of mutation and screening; May et al, Nature Biotech. 18:317-320 (2000), where a hydantomase of reversed stereoselectivity was evolved m a similar fashion; and Miyazaki et al, J. Mol. Biol. 297:1015-1026 (2000), where a thermostable subtilism was evolved.

The screenmg approach has clear advantages over a simple "survival screen," especially if it can be earned out m a high-throughput manner that approaches the throughput of the massively parallel "survival screen" technique. For example, a degree of parallelism has been mtroduced by employmg such measures as digital imaging of the transformed organisms (Joo et al, Chemistry & Biology, 6:699-706 (1999)) or digital spectroscopic evaluation of colonies (Youvan et al, 1994, Meth. Enzymol. 246:732-748). Serial assays can be automated by the use of cell sorting (Fu et al , Nature Biotech , 17:1109-1111 (1999)). A well-established approach to high-thorughput screening involves the automated evaluation of expressed proteins in microtiter plates, using commercially available plate readers, and the method of the present invention is well-suited to the application of this mode of high-throughput screening to directed evolution.

In this embodiment of the invention, a gene encoding a protein of interest is mutated by any known method of generating a plurality of mutants, the mutant protein-encoding DNA is introduced by means of a suitable expression vector into a low-viscosity filamentous fungal host according to the present invention, and the transformants are optionally selected for and cultured. The host cells are then dispersed as described previously into the wells of a microtiter plate, or otherwise spatially separated into resolvable locations, so as to provide individual mono-clonal cultures (or poly-clonal cultures having fewer than about 100 diferent clones). The cells are preferably dispersed into the wells of a micro-titer plate. The protein encodede by the mutant DNA is preferably secreted into the medium in the wells of the microtiter plates. Each of the dispersed cultures is screened for the protein activity of interest, and those most strongly exhibiting the desired property are selected. The gene encoding the protein of interest in the selected cultures is mutated again, the mutant DNA is again introduced into the low- viscosity fungal host, and the transformants are re-screened. The mutating and re-screening process is repeated until the value of the property of interest reaches a desired level.

It will be readily appreciated by those skilled m the art that a protem that appears to be of mterest based upon the screenmg assay will not necessaπly have all the other properties required for commercial utility. For example, the possession of enzymatic activity, however high the specific activity, will not mdicate that the mutant enzyme has the requisite thermal or pH stability, or detergent or protease resistance, or non-immunogenicity, or other property that might be desirable or necessary m a commercially viable product. The pnor art approaches to screenmg have not provided a solution to this need, m that the host organisms (bacteπa and yeast) were not adapted to the production of isolable quantities of protem. The mutant filamentous fungi of the present mvention, on the other hand, are excellent overproducers and secretors of exogenous protems, especially when employed with the vectors disclosed herem. Sufficient protem may be isolated not only for purposes of charactenzation, but for evaluation m application tnals. Indeed, the strams used m screenmg may be suitable for mdustπal production as well, since they possess desirable production properties such as low viscosity and high expression rates.

Accordmgly, m a preferred embodiment of the present mvention, the method further compπses culturing a clonal colony or culture identified accordmg to the method of the mvention, under conditions permitting expression and secretion of the exogenous library protem (or a precursor thereof),and recovermg the subsequently produced protem to obtain the protem of mterest. Expression and secretion of a library protem may be facilitated by creating an m-frame fusion of the cloned gene with the gene for a heterologous protein (or a fragment thereof) with its corresponding signal sequence, or with the signal sequence from a third protem, all operably linked to an expression regulating sequence. By this approach a fusion protem is created that contains heterologous ammo acid sequences upstream of the library protem. Subsequently, this fusion precursor protem may be isolated and recovered usmg punfaction techniques known m the art. The method may optionally compnse subjecting the secreted fusion protem precursor to a cleavage step to generate the library protem of mterest. The cleavage step can be earned out with Kex-2, a Kex-2 like protease, or another selective protease, when the vector is engmeered so that a protease cleavage site links a well-secreted protem earner and the protem of mterest.

The ready availability of mutant protem, directly from the screening host organism, has not previously been possible with pπor art screenmg hosts. The present mvention thus provides an advantage, m that the mutant protems deemed of mterest based upon the high-throughput screen can be isolated m sufficient quantities (milligrams) for further charactenzation and even larger quantities (grams to kilograms) for application tnals. This particular embodiment of the mvention thus permits the practitioner to select mutant protems for the next round of directed evolution based upon any number of desirable properties, and not merely upon the one property detected m the high-throughput screen. The more stringent selection cntena made possible by the present mvention should lead to a more efficient and cost-effective directed evolution process.

The method of production of a recombmant mutant filamentous fungal strain accordmg to the mvention compnses introducing a library of DNA sequences compnsmg nucleic acid sequences encodmg heterologous protems mto a low- viscosity mutant filamentous fungus accordmg to the mvention, the nucleic acid sequences bemg operably linked to an expression regulating region. The introduction of the DNA sequences may be earned out m any manner known per se for transforming filamentous fungi. Those skilled m the art will appreciate that there are several well-established methods, such as CaCl₂-polyethylene glycol stimulated DNA uptake by fungal protoplasts (Johnstone et al, EMBOJ., 1985, 4:1307-1311). A protoplast transformation method is descnbed m the examples. Alternative protoplast or spheroplast transformation methods are known and can be used as have been descπbed m the pπor art for other filamentous fungi. Vectors suitable for multicopy integration of heterologous DNA mto the funal genome are well-known; see for example Giuseppm et al , WO 91/00920. The use of autonomously replicating plasmids has long been known as an efficient transformation tool for fungi (Gems et al, Gene 1991 98:61 - 67; Verdoes et al , Gene 1994 146:159-165; Aleksenko and Clutterbuck, Fungal Genetics Biol. 1997 21:373-387; Aleksenko et al, Mol. Gen Genet. 1996253:242-246). Details of such methods can be found m many of the cited references and are thus incorporated by reference.

Exemplary methods accordmg to the mvention, compnsmg usmg a low-viscosity mutant strain of Chrysosporium or A. sojae as starting mateπal for introduction of vectors carrying heterologous DNA, are presented below.

EXAMPLES

VISCOSITY DETERMINATIONS

The following operating parameter data ranges have been determined for fungal fermentations usmg five different fungal organisms. The five fungal organisms compared were strams of Aspergillus niger, Trichoderma longibrachiatum 18.2KK (formerly T. reesei), Trichoderma longibrachiatum X 252 Chrysosporium lucknowense strain UV18-25, and Aspergillus sojae pclA. Viscosity of a fungal culture vanes during the course of a fermentation, and vanes with nutnent concentration. For the measurements reported here, medium contammg between 20 and 100 g/1 of a carbohydrate carbon source (e g., cellulose, lactose, sucrose, xylose, glucose, and the like) is moculated with the fungus, and the culture allowed to proceed through a "growth phase" durmg which the carbon source is consumed. Shake flask cultures are shaken at 200 rpm, while one-liter fermentation vessels are stirred with an impeller at 500-1000 rpm. Maximal viscosity typically occurs at or close to the end of the growth phase. At this time the culture is switched to a fed batch mode, wherem a carbon source is fed to the culture at a rate such that the concentration of the carbon source does not nse above about 0.5 g 1. A feed rate of between 1 and 3 g/l/hr is typical. Viscosity:

Viscosity is determined on a Brookfield LVF viscometer usmg the small sample adapter and spmdle number 31, operated at 30°C. A fresh sample of fermentation broth (10 ml) is placed m the small sample spmdle. The spmdle speed is adjusted to give a readmg m the range 10-80. After four minutes a readmg is taken from the viscometer scale. The readmg is multiplied by the factor given below to get the viscosity m centipoise (cP).

Spmdle Speed Multiplication Factor

6 50

12 25

30 10

60 5

The final viscosity was measured at fermentation end:

Strain Fmal viscosity, cP (mean ± s.d.)

T. longibrachiatum 18.2KK (297 ± 173)

A. niger 1,500 - 2,000

T longibrachiatum X-252 < 60

C. lucknowense UV18-25 < 10

A. sojae pclA n.d.

EXAMPLES OF TRANSFORMATION COMPARING CHRYSOSPORIUM, TRICHODERMA AND TOLYPOCLADIUM GEODES

Two untransformed Chrysosporium CI strams and one Trichoderma reesei reference strain were tested on two media (Gs pH 6,8 and Pndham agar, PA, pH 6.8). To test the antibiotic resistance level spores were collected from 7 day old PDA plates. Selective plates were mcubated at 32°C and scored after 2, 4 and 5 days. The C-l strams NG7C-19 and UV18-25 clearly had a low basal resistance level both to phleomycm and hygromycin, comparable to that for a reference T reesei laboratory strain. This is a clear indication these standard fungal selectable markers can be used m Chrysosporium strams Problems with other standard fungal selectable markers are not expected.

Selection of Sh-ble (phleomycin-resistance) transformed Chrysosporium strams was successfully earned out at 50 μg/ml This was also the selection level used for T reesei thus showing that differential selection can be easily achieved m Chrysosporium. The same comments are valid for strams transformed for hygromycin resistance at a level of 150 μg/ml.

The protoplast transformation technique was used on Chrysosporium based on the most generally applied fungal transformation technology All spores from one 90mm PDA plate were recovered in 8ml ICl and transferred mto a shake flask of 50ml ICl medium for incubation for 15 hours at 35°C and 200 rpm. After this the culture was centπfuged, the pellet was washed in MnP, brought back mto solution m 10ml MnP and lOmg/ml Caylase C₃ and mcubated for 30 mmutes at 35°C with agitation (150 rpm).

The solution was filtered and the filtrate was subjected to centnfugation for 10 mmutes at 3500 rpm. The pellet was washed with 10 ml MnPCa²⁺. This was centnfuged for 10 mmutes at 25°C Then 50 microlitres of cold MPC was added The mixture was kept on ice for 30 mmutes whereupon 2.5 ml PMC was added. After 15 mmutes at room temperature 500 microlitres of the treated protoplasts were mixed to 3 ml of MnR Soft and immediately plated out on a MnR plate contammg phleomycm or hygromycin as selection agent. After incubation for five days at 30°C transformants were analysed (clones become visible after 48 hours). Transformation efficiency was determined usmg 10 μg of reference plasmid pAN8-l. The results are presented m the following Table D.

Table D: Transformation efficiency

(usmg 10 μg of reference plasmid pAN8-l)

The results show that the Chrysosporium transformant viability is supenor to that of Trichoderma. The transformabihty of the strams is comparable and thus the number of transformants obtained in one expenment lies 4 times higher for Chrysosporium than for T reesei. Thus the Chrysosporium transformation system not only equals the commonly used T. reesei system, but even outperforms it. This improvement can prove especially useful for vectors that are less transformation efficient than pAN8-l.

A number of other transformation and expression plasmids were constructed with homologous Chrysosporium protein encoding sequences and also with heterologous protein encoding sequences for use in transformation experiments with Chrysosporium. The vector maps are provided in Figures 6-11.

The homologous protein to be expressed was selected from the group of cellulases produced by Chrysosporium and consisted of endoglucanase 6 which belongs to family 6 (MW 43 kDa) and the heterologous protein was endoglucanase 3 which belongs to family 12 (MW 25 kDa) of Penicillium. pF6g comprises Chrysosporium endoglucanase 6 promoter fragment linked to endoglucanase 6 signal sequence in frame with the endoglucanase 6 open reading frame followed by the endoglucanase 6 terminator sequence. Transformant selection is carried out by using cotransformation with a selectable vector. pUTl 150 comprises Trichoderma reesei cellobiohydrolase promoter linked to endoglucanase 6 signal sequence in frame with the endoglucanase 6 open reading frame followed by the T. reesei cellobiohydrolase terminator sequence. In addition this vector carries a second expression cassette with a selection marker i.e. the phleomycin resistance gene (Sh-ble gene). pUTl 152 comprises Aspergillus nidulans glyceraldehyde-3-phosphate dehydrogenase A promoter linked to endoglucanase 6 signal sequence in frame with the endoglucanase 6 open reading frame followed by the A. nidulans anthranilate synthase (trpC) terminator sequence. In addition this vector carries a second expression cassette with a selection marker, i.e. the phleomycin resistance gene (Sh-ble gene). pUTl 155 comprises A. nidulans glyceraldehyde-3-phosphate dehydrogenase A promoter linked to Trichoderma reesei cellobiohydrolase signal sequence in frame with the carrier protein Sh-ble which in turn is linked in frame to the endoglucanase 6 open reading frame followed by the A. nidulans trpC terminator sequence. This vector uses the technology of the carrier protein fused to the protein of interest which is known to very much improve the secretion of the protein of interest. pUTl 160 comprises Aspergillus nidulans glyceraldehyde-3 -phosphate dehydrogenase A promoter linked to Trichoderma reesei cellobiohydrolase signal sequence in frame with the carrier protein Sh-ble which in tum is linked in frame to the endoglucanase 3 open reading frame of Penicillium followed by the A. nidulans trpC terminator sequence. pUTl 162 comprises Trichoderma reesei cellobiohydrolase promoter linked to endoglucanase 3 signal sequence in frame with the endoglucanase 3 open reading frame of Penicillium followed by the T. reesei cellobiohydrolase terminator sequence. In addition this vector carries a second expression cassette with the phleomycin resistance gene (Sh-ble gene) asa selection marker.

It will be apparent to those skilled in the art that a sample of genomic or cDNA can be readily sheared or digested into protein-encoding fragments, and the fragments ligated into vectors such as those lllustrated herem so as to produce a library of expression vectors. It will be further apparent that methods employmg co-transfection are applicable, and that autonomously replicating vectors or integrating vectors may be employed to transfect filamentous fungi with such a library of vectors.

Table E: Comparative transformations

Table E shows the results of transformation of both Chrysosporium UV18-25 and Tolypocladium geodes. The transformation protocol used is descπbed in the section for heterologous transformation.

HETEROLOGOUS AND HOMOLOGOUS EXPRESSION OF CHRYSOSPORIUM TRANSFORMANTS CI strams (NG7C-19 and or UV18-25) were tested for their ability to secrete vanous heterologous protems: a bactenal protein (Streptoalloteichus hindustanus phleomycm-resistance protem, Sh-ble), a fungal protem (Trichoderma reesei xylanase II, XYN2) and a human protem (the human lysozyme, HLZ).

The details of the process are as follows:

[1] CI secretion of Streptoalloteichus hindustanus phleomycm-resistance protem (Sh-ble).

CI strams NG7C-19 and UV18-25 were transformed by the plasmid pUT720 (ref. 1). This vector presents the following fungal expression cassette:

- Aspergillus nidulans glyceraldehyde-3 -phosphate dehydrogenase (gpdA) promoter (ref. 2)

- A synthetic Trichoderma reesei cellobiohydrolase I (cbhl) signal sequence (refs 1, 3)

- Streptoalloteichus hindustanus phleomycm-resistance gene Sh-ble (ref. 4)

- Aspergillus nidulans tryptophan-synthase (trpC) terminator (ref. 5)

The vector also carnes the beta-lactamase gene (bla) and E. coli replication ongm from plasmid pUC18 (Ref 6). The detailed plasmid map is provided in figure 2

CI protoplasts were transformed accordmg to Durand et al. (ref.7) adapted to CI (media & solutions composition is given elsewhere). All spores from one 90mm PDA plate of untransformed CI stram were recovered m 8ml ICl and transferred mto a shake flask with 50ml ICl medium for mcubation 15 hours at 35°C and 150 rpm. Thereupon, the culture was spun down, the pellet washed m MnP, resolved in 10ml MnP + lOmg/ml Caylase C₃, and mcubated 30 mm at 35°C with agitation (150 rpm) The solution was filtered and the filtrate was centnfuged 10 mm at 3500 rpm. The pellet was washed with 10ml MnPCa²⁺. This was spun down 10mm at 3500 rpm and the pellet was taken up mto 1ml MnPCa²⁺. lOμg of pUT720 DNA were added to 200μl of protoplast solution and mcubated 10mm at room temperature (ca. 20°C). Then, 50μl of cold MPC was added. The mixture was kept on ice for 30mm whereupon 2.5ml PMC was added. After 15mm at room temperature 500μl of the treated protoplasts were mixed to 3ml of MnR Soft and immediately plated out on a MnR plate contammg phleomycm (50μg/ml at pH6.5) as selection agent. After 5 days mcubation at 30°C, transformants were analysed (clones start to be visible after 48 hours).

The Sh-ble production of CI transformants (phleomycin-resistant clones) was analysed as follows: Primary transformants were toothpicked to GS+phleomycin (5 μg/ml) plates and grown for 5 days at 32°C for resistance venfication. Each validated resistant clone was subcloned onto GS plates. Two subclones per transformant were used to moculate PDA plates m order to get spores for liquid culture mitiation. The liquid cultures m ICl were grown 5 days at 27°C (shaking 200 rpm). Then, the cultures were centnfuged (5000g, 10mm.) and 500μl of supernatant were collected. From these samples, the protems were precipitated with TCA and resuspended m Western Sample Buffer to 4 mg/ml of total protems (Lowry method, Ref. 8). lOμl (about 40μg of total protems) were loaded on a 12% acrylamide/SDS gel and run (Mmi Trans-Blot™ system, BioRad Laboratones). Western blotting was conducted accordmg to BioRad instructions (Schleicher & Schull 0.2μm membrane) usmg rabbit anti-Sh-ble antiserum (Societe Cayla, Tolouse FR, Catalog #ANTI-0010) as pπmary antibody. The results are shown m Figure 1 and Table F.

Table F: Sh-ble estimated production levels m CI

These data show that:

1) The heterologous transcnption/translation signals from pUT720 are functional m Chrysosporium.

2) The heterologous signal sequence of pUT720 is functional m Chrysosporium.

3) Chrysosporium can be used a host for the secretion of heterologous bactenal protems.

[2] CI secretion of human lysozyme (HLZ).

CI strams NG7C-19 and UV18-25 were transformed by the plasmid pUT970G (ref. 9) This vector presents the following fungal expression cassette:

- Aspergillus nidulans glyceraldehyde-3 -phosphate dehydrogenase (gpdA) promoter (ref.2)

- A synthetic Trichoderma reesei cellobiohydrolase I (cbhl) signal sequence (refs. 1, 3)

- Streptoalloteichus hindustanus phleomycm-resistance gene Sh-ble 4 used as carπer-protem (ref.

10)

■ Aspergillus niger glucoamylase (glaAl) hmge domam cloned from plasmid pAN56-2 (refs. 11 ,

12)

- A linker peptide (LGERK) featuring a KEX2-lιke protease cleavage site (ref. 1)

- A synthetic human lysozyme gene (hlz) (ref. 10)

- Aspergillus nidulans tryptophan-synthase (trpC) terminator (ref. 5)

The vector also carπes the beta-lactamase gene (bla) and E coli replication ongm from plasmid pUCl 8 6 The detailed plasmid map is provided m Figure 3.

CI protoplasts were transformed with plasmid pUT970G following the same procedure already descnbed m example 1. The fusion protem (Sh-ble :: GAM hmge :: HLZ) is functional with respect to the phleomycm-resistance thus allowing easy selection of the CI transformants. Moreover, the level of phleomycm resistance correlates roughly with the level of hlz expression.

The HLZ production of CI transformants (phleomycm-resistant clones) was analysed by lysozyme-activity assay as follow: Primary transformants were toothpicked to GS+phleomycm (5 μg/ml) plates (resistance venfication) and also on LYSO plates (HLZ activity detection by clearing zone visualisation (refs. 1, 10). Plates were grown for 5 days at 32°C. Each validated clone was subcloned onto LYSO plates. Two subclones per transformant were used to inoculate PDA plates m order to get spores for liquid culture mitiation The liquid cultures in ICl were grown 5 days at 27°C (shaking 180 rpm). Then, the cultures were centnfuged (5000g, lOmin.). From these samples, lysozyme activity was measured accordmg to Mδrsky et al. (ref. 13)

Table G: Active HLZ production levels in CI

These data show that:

1) Points 1 & 2 from example 1 are confirmed.

2) Sh-ble is functional m Chrysosporium as resistance marker.

3) Sh-ble is functional m Chrysosporium as earner protem.

4) The KEX2-hke protease cleavage site is functional m Chrysosporium (otherwise HLZ would not be active).

5) Chrysosporium can be used as host for the secretion of heterologous mammalian protems.

[3] CI secretion of Tπchoderma reesei xylanase II (XYN2).

CI stram UV18-25 was transformed by the plasmids pUT1064 and pUT1065. pUT1064 presents the two following fungal expression cassettes: The first cassette allows the selection of phleomycin-resistant transformants:

- Neurospora crassa cross-pathway control gene 1 (cpc-T) promoter (ref. 14)

- Streptoalloteichus hindustanus phleomycm-resistance gene Sh-ble (ref. 4)

- Aspergillus nidulans tryptophan-synthase (trpC) terminator (ref. 5) The second cassette is the xylanase production cassette:

-T reesei stram TR2 cbhl promoter (ref. 15)

-T reesei stram TR2 xyn2 gene (mcludmg its signal sequence) (ref. 16) -T reesei stram TR2 cbhl terminator (ref. 15)

The vector also carnes an E coli replication ongm from plasmid pUC19 (ref. 6) The plasmid detailed map is provided m figure 4. pUT1065 presents the following fungal expression cassette:

- A nidulans glyceraldehyde-3 -phosphate dehydrogenase (gpdA) promoter (ref 2)

- A synthetic T reesei cellobiohydrolase I (cbhl) signal sequence (refs. 1, 3)

- S hindustanus phleomycm-resistance gene Sh-ble 4 used as carner-protem (ref. 10)

- A linker peptide (SGERK) featuring a KEX2-hke protease cleavage site (ref. 1)

- T reesei stram TR2xyn2 gene (without signal sequence) (ref. 16)

- A nidulans tryptophan-synthase (trpC) terminator (ref. 5)

The vector also carnes the beta-lactamase gene (bla) and an E. coli replication ongm from plasmid pUC 18 (Ref. 6). The plasmid detailed map is provided in Figure 5.

CI protoplasts were transformed with plasmid pUT1064 or pUT1065 following the same procedure already descnbed m example 1 The fusion protem m plasmid pUT1065 (Sh-ble :. XYN2) is functional with respect to the phleomycm-resistance thus allowmg easy selection of the CI transformants. Moreover, the level of phleomycm resistance correlates roughly with the level of xyn2 expression. In pUT1064, xyn2 was cloned with its own signal sequence.

The xylanase production of CI transformants (phleomycin-resistant clones) was analysed by xylanase-activity assay as follow: Pnmary transformants were toothpicked to GS+phleomycm (5 μg/ml) plates (resistance venfication) and also on XYLAN plates (Ref. 17), where xylanase activity is detected by observation of a clearmg zone. Plates were grown for 5 days at 32°C. Each validated clone was subcloned onto XYLAN plates. Two subclones per transformant were used to moculate PDA plates m order to get spores for liquid culture inoculation. The liquid cultures m IC1+ 5g/l K⁺ Phtalate were grown 5 days at 27°C (shaking 180 rpm) Then, the cultures were centnfuged (5000g, 10 mm.). From these samples, xylanase activity was measured by DNS Technique according to Miller et al. (ref. 18) Table H: Active XYN2 production levels m CI (best producers)

These data show that:

1) Pomts 1 to 4 from example 2 are confirmed.

2) CI can be used as host for the secretion of heterologous fungal protems.

[4] Table I shows the results for the plasmids with which transformation of UVl 8-25was earned out. The Table shows expression levels for endoglucanase and cellobiohydrolase usmg heterologous expression regulating sequences and signal sequences and also with homologous expression regulating sequences and signal sequences. The details of the vanous plasmids can be denved elsewhere in the descnption and from the figures. The production occurs at alkaline pH at a temperature of 35°C.

Table I: Expression data of transformed UV18-25 strain (% relative to parent UV18-25 strain)

Culture conditions (shake flask): 88h, 35°C, 230 rpm

CONSTRUCTION OF AN Aspergillus sojae GENE LIBRARY

Genomic DNA of A sojae was isolated from protoplasts obtained from ATCC 11906 usmg a previously descπbed protocol (Punt, van den Hondel, Methods Enzymol 216:447-457 (1992)). After isolation DNA was extracted from the protoplasts usmg the protocol descπbed by Kolar et al, Gene 62:127-34 (1988). Subsequently the DNA was partially digested with Mbol to result in a DNA fragment of an average size of 30-50 kb.

Vector pAOpyrGcosarpl, which was used for the construction of the gene library was constructed by hgation of a 3 kb BamHI-Hmdll fragment from pANsCosl (Osiewacz, Curr Genet. 26:87-90 (1994)) and a 3.2 kb Acc65I-HmdIII fragment from pA04.2 (De Ruiter- Jacobs et al, Curr. Genet. 16-159-63 (1989)) Acc65I-BamHI digested pHELPl (Gems et al , Gene 1991 98:61-67). This cosmid vector carnes the A oryzae pyrG selection marker and is self-replicating in filamentous fungi.

Mbol digested genomic DNA was hgated to BamHI-digested pAOpyrGcosarpl, and the hgation mixture was packaged into phage particles usmg the Stratagene Supercosl vector kit (Stratagene Inc., La Jolla CA) resulted in a total of ca. 30,000 individual clones, representing an approximate 30-fold representation of the A. sojae genome. Stocks (in 15% glycerol) of pools of the resulting clones were stored at -80°C for later use.

HIGH FREQUENCY TRANSFORMATION

An A. sojae ATCC 11906 pyrG mutant was selected as a fluoroorotic acid-resistant deπvative from ATCC 11906, as descnbed m PCT/EU99/202516.3. This stram, A. sojae ATCC 11906pyrG, was transformed with two vectors carrymg the A. niger pyrG gene. One vector pAB4-l (van Hartingsveldt et al, Mol. Gen. Genet. 206:71-75 (1987)) carnes only the pyrG gene, whereas pAB4-arpl (Verdoes et al , Gene 146:159-165 (1994)) carnes the pyrG gene and the A. nidulans AMA1 sequence. Transformation of ATCC 11906pyrG results m 5-10 transformants per microgram DNA from pAB4-l, whereas with pAB4-arpl frequency were at least 10-100 fold higher. Phenotypic analysis of the transformants revealed that the pyrG phenotype of the pAB4-arpl transformants was maintained only under continuous selection, whereas the pAB4-l transformants were stable with and without selection for the pyrG phenotype. These results confirm autonomous replication of the mtroduced plasmid DNA m pAB4-arpl transformants. Similar results were obtained with alternative fungal transformation vectors carrymg the AMA1 sequence or denvatives thereof, e.g. pAOpyrGcosarpl.

CONSTRUCTION OF A FUNGAL TRANSFORMANT LIBRARY

A. sojae ATCC11906pyrG or relevant mutants, m particular compact morphology mutants thereof, was transformed with an A. sojae gene library based on transformation vector pAOpyrGcosarp 1. This vector results in a high frequency of transformants with freely replicating vector copies. Fungal protoplasts were treated as descπbed m Punt and van den Hondel, Methods Enzymol 216:447-457 (1992) with DNA from a cosmid library carrymg genomic fungal DNA clones from A. sojae or Chrysosporium and senal dilutions of the transformed protoplasts were plated on selective agar plates to determine the transformation frequency obtained. The remammg protoplasts were regenerated m selective medium for a few hours and stored at 4°C. Based on the results obtained for the transformation frequency (which dependmg of the experiment will reach values up to several thousand transformants per microgram of cosmid library DNA), limiting dilutions of the regenerated protoplasts were plated m microtiter plates of 96, 248, or alternative well format, resulting m one transformed protoplast per well. Plates were mcubated at 35°C to form fungal biomass. The resulting transformant library is used for further expenments.

A similar strategy was used for the construction of a collection of fungal transformants carrying mutant alleles of Chrysosporium CBH1 or any other protein of mterest which were generated by mutagenesis, gene shuffling or gene-evolution approaches. DEVELOPMENT OF COMPACT GROWTH MORPHOLOGY MUTANTS

Vanous patent applications teach that morphological mutants can be isolated by vanous ways of screening. WO9602653 and WO9726330 describe non-defined mutants exhibiting compact morphology. It was found that a proprotem processing mutant of A. sojae had an unexpected aberrant growth phenotype (hyper-branchmg) while no detπmental effect on protem production was observed. Culture expenments with this strain revealed a very compact growth phenotype with micropellets. The observed charactenstics were not only present in A. sojae but other mutated fungi as well, e.g. A. Niger. (1) Construction of an A. niger proprotein processing mutant

To clone the proprotem convertase encoding gene from A. niger, PCR was used. Based on the comparison of vanous proprotem convertase genes from vanous yeast species and higher eukaryotes, different PCR pnmers were designed (SEQ. ID Nos 4-12) which are degenerated, respectively, 4, 2, 2, 512, 1152, 4608, 2048 and 49152 times. From the amplification usmg pnmers PE4 and PE6, two individual clones were obtained of which the encoded protem sequence did show significant homology to the S. cerevisiae KEX2 sequence (SEQ. ID No. 13). These clones were used for further expenments.

Based on the observed homology to other proprotein convertase genes of the cloned PCR fragment, the corresponding A niger gene was designated pclA (from/?roprotem-convertase-/ιke). Southern analysis of genomic digests of A. niger revealed that the pclA gene was a single copy gene with no closely related genes in the A. niger genome, as even at heterologous hybridisation conditions (50°C; washes at 6xSSC) no additional hybndisation signals were evident. A first screening of an EMBL3 genomic library of A. niger N401 (van Hartmgsveldt et al, Mol. Gen. Genet. 206:71-75 (1987)) did not result in any positively hybπdismg plaques although about 10-20 genome equivalents were screened. In a second screening a full length genomic copy of the pclA gene was isolated from an A. niger N400 genomic library m EMBL4 (Goosen et al, Curr. Genet. 11:499-503 (1987)).

Of the 8 hybridising plaques which were obtained after screening 5-10 genome equivalents, 6 were still positive after a first rescreenmg. All these 6 clones most likely earned a full copy of the pclA gene, as m all clones (as was observed for the genomic DNA) with the PCR fragment two hybndismg EcoRV fragments of 3 and 4 kb were present (Note that the PCR fragment (SEQ. ID No. 13) contains an EcoRV restnction site). Based on compaπson of the size of other proprotem convertases, together these fragments will contain the complete pclA gene with 5' and 3' flanking sequences. The two EcoRV fragments and an overlapping 5 kb EcoRI fragment were subcloned for further charactensation. A detailed restnction map of the DNA fragment carrying the pel A gene is given in figure 7.

Based on the restnction map given m figure 7 the complete DNA sequence of the pclA gene was determined from the EcoRI and EcoRV subclones (SEQ. ID No. 14). Analysis of the obtained sequence revealed an open reading frame with considerable similaπty to that of the S. cerevisiae KEX2 gene and other proprotein convertases. Based on further companson two putative mtron sequences (SEQ. ID No. 14, from position 1838 to 1889 and from 2132 to 2181) were identified m the codmg region. Subsequent PCR analysis with pnmers flanking the putative mtrons, on a pEMBLyex based A. niger cDNA library revealed that only the most 5' of these two sequences represented an actual mtron. The general structure of the encoded PclA protein was clearly similar to that of other proprotein convertases (SEQ. ID No. 15 and figure 8). The overall similarity of the PclA protem with the other proprotein convertases was about 50% (figure 9).

To demonstrate that the cloned pclA gene is a functional gene encodmg a functional protem, the construction of strains devoid of the pclA gene was attempted Therefore, pPCLlA, apclA deletion vector, in which a large part of the pclA coding region was replaced for the A oryzaepyrG selection marker, was generated. Subsequently, from this vector the 5 kb EcoRI insert fragment was used for transformation of various A niger strams. From these transformations (based on pyrG selection) numerous transformants were obtained. Interestingly, a fraction of the transformants (varying from 1- 50%) displayed a very distinct aberrant phenotype (figure 10). Southern analysis of several wildtype and aberrant transformants revealed that these aberrant transformants which displayed a severely restπcted growth phenotype, had lost the pclA gene All strams displaying wild-type growth were shown to carry a copy of the replacement fragment integrated adjacent to the wild-type pclA gene or at a non-homologous position. (2) Construction of an A sojae proprotein processing mutant

To construct the corresponding mutant in A sojae, functional complementation of the low-viscosity mutant of A niger was carried out by transformation of an A niger pclA mutant with the A sojae ATCC 11906 cosmid library. From the resulting complemented A niger transformants, genomic cosmid clones were isolated, which comprised the A sojae protein processing protease pclA. Partial sequence analysis of the isolated sequences SEQ. ID Nos 16, 17 and 18 confirmed the cloning of the A sojae pclA gene. Figure 11 shows the compaπson of the amino acid sequence of the A niger and A sojae pclA encoded protems, with those of proprotein processing enzymes from vanous yeast species. Based on the cloned A sojae pclA sequences a gene replacement vector was generated following an approach similar to that descπbed elsewhere m our examples usmg the reusable pyrG selection marker descπbed m PCT EU99/202516. In addition, a gene disruption vector was constructed carrying the pyrG selection marker and 5' and 3' truncated fragment from the A sojae pclA gene Both the gene replacement and gene disruption vector were used to generate pclA mutants m ATCC 11906 and ATCC 11906 deπvatives. Culture expenments with some of the resulting transformants revealed improved morphological charactenstics, in particular compact growth morphology and micropellets.. (3) Isolation of alternative A sojae compact growth mutants

Transformation of A sojae ATCC 11906 and denvatives may be earned out with linear DNA fragments carrymg a fungal selection marker. If no specific replicating sequences are provided transformants obtained using this procedure carry the introduced DNA integrated mto the genome of the host stram. As the introduced selection marker is from heterologous ongm (A niger) only heterologous recombination will occur, leadmg to a collection of transformants carrying the marker DNA at vanous positions in the genome. This integration is prone to result in disruption of endogenous A sojae sequences, thus resulting in a collection of A sojae mutant strams. This is exemplified by the analysis of a large collection of transformants obtained from A sojae ATCC l l906alpApyrG using a DNA fragment with the A niger pyrG selection marker. In total several thousand transformants were analysed and from these 5-10 showed a morphologically aberrant phenotype. Amoung these several had a phenotype comparable to the pclA mutants. Similar as descnbed for the clonmg of the A. sojae pclA gene, the gene corresponding to the mutation could be isolated from the A. sojae gene library by complementation of the morphological phenotype. Based on the cloned gene the corresponding gene disruption deletion mutants can be generated. INDUCTION OF SPORULATION IN SUBMERGED FERMENTATION

Many fungi, such as Aspergillus sojae, do not show sporulation under submerged fermentation. Here we descnbe a previously unknown approach to obtain sporulation under these conditions. A sojae ATCC 11906 and in particular compact growth morphology mutants thereof were grown is a synthetic growth medium supplemented with Yeast extract. Under these conditions rapid accumulation of biomass is occunng in static and agitated cultures. However, no sporulation occurs in the culture fluid. As similar growth medium with the addition of 0.6 g/kg EDTA results in considerable yields of spores reaching up to 10⁹ spores per ml culture fluid after cubation of 2-4 days at 35°C STANDARD MEDIUM (+/- EDTA): g/kg medium

KH₂P0₄ 2.5

NH₄C1 7.2

MgSO₄-7H₂0 0.7

CaCl₂-2H₂0 0.2

Yeast Extract 20

ZnS0₄ 7H₂0 0.015

CoCl₂-6H₂0 0.005

CuSO₄-5H₂0 0.016

FeSO₄-7H₂0 0.040

H₃B0₄ 0.005

KI 0.003

MnCl₂ 2H20 0.012

Na₂Mo0₄ 2H20 0.003

EDTA (0.6 or 0.0) PH adjusted to 5.5 with NaOH/H₃P0₄ EXPRESSION AND SECRETION OF GREEN FLUORESCENT PROTEIN IN A. sojae

As an example of a versatile and easily screenable reporter protein we have attempted to produce in A. sojae GFP (Green Fluorescent Protein) from the jelly fish Aequoria victoria. Vectors carrying GFP (pGPDGFP) or Glucoamylase-GFP fusion genes (pGPDGLAGFP, derivatives of the vectors described by Gordon et al, Microbiology 146(Pt 2):415-426 (2000), in which the glaA promoter was replaced for the constitutively expressed A. nidulans gpdA promoter) were introduced into A. sojae by cotransformation using either the pyrG or amdS selection marker. Expression resulted in brightly fluorescent A. sojae transformants for both vector types. Based on the observed fluorescence in transformant colonies (both vectors) and the subsequent analysis of fluoresence of culture supernatants from transformants carrying Glucoamylase-GFP fusion protein, the production of fluorescently active GFP was confirmed.

Appendix to the Examples: Transformation media:

Mandels Base: MnP Medium ;

KH₂P0₄ 2.0 g/1 Mandels Base with

(NH₄)₂S0₄ 1.4 g/1 Peptone 1 g/1

MgS0₄-7H₂0 0.3 g/1 MES 2 g/1

CaCl₂ 0.3 g/1 Sucrose 100 g/1

Oligoelements 1.0 ml/1 Adjust pH to 5

MnR MnP CA²⁺ :

MnP+sucrose 130 g/1 MnP Medium +

Yeast extract 2.5 g/1 CaCl₂2H₂0 50 mM

Glucose 2.5 g/1 Adjust pH to 6.5

Agar 15 g/1

MnR Soft : MnR with only 7.5 g/1 of agar.

MPC :

CaCl₂ 50 mM pH 5.8

MOPS 10 mM

PEG 40%

For selection and culture

GS :

Glucose 10 g/1

Biosoyase 5 g/l [Merieux]

Agar 15 g 1 pH should be 6.8 PDA :

Potato Dextrose Agar ^■ (Difco) 39 g/1 pH should be 5.5

MPG :

Mandels Base with

K.Phtalate 5 g/1

Glucose 30 g/1

Yeast extract 5 g/1

ICl

0.5 g/L K2HP04

0.15 g/L MgSO4-7H20

0.05 g/L KC1

0.007 g/L FeS04-7H20

1 g/L Yeast extract (ohly KAT)

10 g/L Peptone or Pharmamedia

10 g/L lactose

10 g/L glucose pH 7.0

The regeneration media (MnR) supplemented with 50 μg/ml phleomycm or 100-150 μg/ml hygromycm is used to select transformants. GS medium, supplemented with 5 μg/ml phleomycm is used to confirm antibiotic resistance.

PDA is a complete medium for fast growth and good sporulation. Liquid media are moculated with l/20th of spore suspension (all spores from one 90mm PDA plate m 5ml 0.1% Tween). Such cultures are grown at 27°C m shake flasks (200 rpm).

References

(The contents of the following, and all patents and references cited heremabove, are incorporated herein by reference):

1. Calmels T.P., Martin F., Durand H., and Tiraby G. (1991) Proteolytic events in the processing of secreted proteins in fungi. J Biotechnol 17(l):51-66.

2. Punt P.J., Dingemanse M.A., Jacobs-Meysing B.J., Pouwels P.H., and van den Hondel CA. (1988) Isolation and characterization of the glyceraldehyde-3-phosphate dehydrogenase gene of Aspergillus nidulans. Gene 69(1).49-57.

3. Shoemaker S., Schweickart V., Ladner M., Gelfand D., Kwok S., Myambo K., and Innis M (1983) Molecular cloning of exo-cellobiohydrolase I derived from Trichoderma reesei strain L27. Bio/Technology Oct.:691-696. 4. Drocourt D., Calmels T., Reynes J.P., Baron M., and Tiraby G. (1990) Cassettes of the Streptoalloteichus hindustanus ble gene for transformation of lower and higher eukaryotes to phleomycin resistance. Nucleic Acids Res 18(13):4009.

5. Mullaney E.J., Hamer J.E., Roberti K.A., Yelton M.M., and Timberlake W.E. (1985) Primary structure of the trpC gene from Aspergillus nidulans. Mol Gen Genet 199(l):37-45.

6. Yanisch-Perron C, Vieira J., and Messing J. (1987) Improved Ml 3 phage cloning vectors and host strains: nucleotide sequences of the M13mpl8 andpUC19 vectors. Gene 33:103-119.

7. Durand H., Baron M., Calmels T., and Tiraby G. (1988) Classical and molecular genetics applied to Trichoderma reesei for the selection of improved cellulolytic industrial strains, in Biochemistry and genetics oj cellulose degradation, J.P. Aubert, Editor. Academic Press, pp. 135-151.

8. Lowry O.H., Rosebrough N.J., Farr A.L., and Randall R.J. (1951) Protein measurements with the folin phenol reagent. J. Biol. Chem 193, 265-275.

9. Parriche M., Bousson J.C., Baron M., and Tiraby G. Development of heterologous protein secretion systems in filamentous fungi, in 3rd European Conference on Fungal Genetics. 1996. Mϋnster, Germany.

10. Baron M., Tiraby G., Calmels T., Parriche M., and Durand H. (1992) Efficient secretion of human lysozyme fused to the Sh-ble phleomycin resistance protein by the fungus Tolypocladium geodes. J. Biotechnol. 24(3):253-266.

11. Jeenes D.J., Marczinke B., MacKenzie D.A., and Archer D.B. (1993) A truncated glucoamylase gene fusion for heterologous protein secretion from Aspergillus niger. FEMS Microbiol. Lett. 107(2-3):267-271.

12. Stone P.J., Makoff A.J., Parish J.H., and Radford A. (1993) Cloning and sequence-analysis of the glucoamylase gene of neurospora-crassa. Cuπent Genetics 24(3):205-211.

13. Mδrsky P. (1983) Turbidimetric determination of lysozyme with Micrococcus lysodeikticus cells: Reexamination of reaction conditions. Analytical Biochem. 128:77-85.

14. Paluh J.L., Orbach M.J., Legerton T.L., and Yanofsky C. (1988) 772e cross-pathway control gene of Neurospora crassa, cpc-1, encodes a protein similar to GCN4 of yeast and the DNA-binding domain of the oncogene v-jun-encoded protein. Proc. Natl. Acad. Sci. U S A 85(11):3728-32.

15. Nakari T., Onnela M.L., Ilmen M., Nevalainen K., and Penttila M. (1994) Fungal promoters active in the presence of glucose, International patent application WO 94/04673

16. Torronen A., Mach R.L., Messner R., Gonzalez R., Kalkkinen N., Harkki A., and Kubicek CP. (1992) 77je two major xylanases from Trichoderma reesei: characterization of both enzymes and genes. Biotechnology (N Y) 10(11):1461-5.

17. Farkas V. (1985) Novel media for detection of microbiol producers of cellulase and xylanase. FEMS Microbiol. Letters 28:137-140.

18. Miller G.L. (1959) Use of dinitrosalicylic acid reagent for determination of reducing sugar. Anal. Chem. 31:426-428. 9. Punt P.J., Mattern I.E., van den Hondel C.A.M.J.J. (1988) A vector for Aspergillus transformation conferring phleomycin resistance. Fungal Genetics Newsletter 35, 25-30.

Claims

1. A method of expressing a plurality of proteins encoded by a library of DNA vectors, wherein the library of vectors comprises a plurality of different vectors, each different vector comprising a different protein-encoding nucleic acid sequence, said nucleic acid sequence being operably linked to an expression-regulating region and optionally a secretion signal encoding sequence, comprising the steps of: a) providing a mutant filamentous fungus having a phenotype characterized by a culture viscosity, when cultured in suspension, of less than 200 cP at the end of fermentation when grown with adequate nutrients under optimal or near-optimal conditions; b) stably transforming said mutant filamentous fungus with said library of DNA vectors so as to introduce into each of a plurality of individual fungi at least one heterologous protein-encoding nucleic acid sequence; and c) culturing the transformed mutant filamentous fungi under conditions conducive to expression of the heterologous proteins encoded by the heterologous protein-encoding nucleic acid sequences.

2. A method of screening a plurality of proteins for an activity or property of interest, comprising the steps of: a) stably transforming a mutant filamentous fungus with a library of DNA vectors, by the method of claim 1 ; b) separating a plurality of individual transformed mutant filamentous fungi from one another; c) optionally culturing into a clonal culture or clonal colony the individual transformed mutant filamentous fungi, under conditions conducive to expression of the heterologous proteins encoded by the heterologous protein-encoding nucleic acid sequences; and d) screening each individual organism, clonal culture, or clonal colony for the activity or property of interest.

3. The method of claim 2, wherein the screening step is carried out by high-thoughput screening.

4. A method of optimizing a protein's activity or property of interest, comprising the steps of: a) providing a library of vectors which comprise DNA sequences encoding mutant forms of the protein; b) stably transforming a mutant filamentous fungus with the library of vectors, by the method of claim 1 ; c) separating a plurality of individual transformed mutant filamentous fungi from one another; d) optionally culturing into a clonal culture or clonal colony the individual transformed mutant filamentous fungi, under conditions conducive to expression of the heterologous proteins encoded by the heterologous protein-encoding nucleic acid sequences; e) screening each individual organism, clonal culture, or clonal colony for an expressed protein having the activity or property of interest; f) isolating one or more individual organisms, clonal cultures, or clonal colonies that express protein exhibiting the activity or property of interest; g) mutating the DNA from the isolated individual organisms, clonal cultures, or clonal colonies that encodes the protein exhibiting the activity or property of interest; h) providing a library of vectors which comprise the mutated DNA sequences obtained in step (g); and h) repeating steps (b) through (g), until the property or activity of interest either reaches a desirable level or no longer improves.

5. The method of claim 4, further comprising between steps (f) and (g) the steps of

(i) culturing one or more of the individual organisms, clonal cultures, or clonal colonies isolated in step (f);

(ii) isolating the expressed protein exhibiting the activity or property of interest; and

(iii) evaluating the isolated protein for desirable properties.

6. The method of any one of claims 1-5, wherein the fungus has a phenotype characterized by a culture viscosity, when cultured in suspension, of less than 100 cP at the end of fermentation when grown with adequate nutrients under optimal or near-optimal conditions.

7. The method of any one of claims 1-5, wherein the fungus has a phenotype characterized by culture viscosity, when cultured in suspension, of less than 60 cP at the end of fermentation when grown with adequate nutrients under optimal or near-optimal conditions.

8. The method of any one of claims 1-5, wherein the fungus has a phenotype characterized by a culture viscosity, when cultured in suspension, of less than 10 cP at the end of fermentation when grown with adequate nutrients under optimal or near-optimal conditions.

9. The method of any one of claims 1-5, wherein the vectors comprise a fungal signal sequence.

10. The method of claim 9, wherein the fungal signal sequence is the signal sequence of a fungal gene encoding a protein selected from the group consisting of cellulase, β-galactosidase, xylanase, pectinase, esterase, protease, amylase, polygalacturonase and hydrophobin.

11. The method of any one of claims 1-5, wherein the vectors comprise a nucleotide sequence encoding a selectable marker.

12. The method of any one of claims 1-5, wherein the vectors comprise an expression- regulating region region operably linked to the protein-encoding nucleic acid sequence.

13. The method of claim 12, wherein the expression regulating region comprises is an inducible promoter.

14. The method of any one of claims 1-5, wherein the fungus is of the class Euascomycetes.

15. The method of claim 14 wherein the fungus is of the sub-class Plectomycetes.

16. The method of claim 14 wherein the fungus is of the sub-class Pyrenomycetes.

17. The method of claim 15 wherein the fungus is of the order Onygenales.

18. The method of claim 15 wherein the fungus is of the order Eurotiales.

19. The method of claim 18 wherein the fungus is of the family Trichocomaceae.

20. The method of claim 17 wherein the fungus is of the family Onygenaceae.

21. The method of any one of claims 1-5, wherein the fungus is of the division Ascomycota, with the proviso that it is not of the order Saccharomycetales.

22. The method of any one of claims 1-5, wherein the fungus is of a genus selected from the group consisting of : Aspergillus, Trichoderma, Chrysosporium, Neurospora, Rhizomucor, Hansenula, Humicola, Mucor, Tolypocladium, Fusarium, Penicillium, Talaromyces, Emericella and Hypocrea.

23.The method of claim 22 wherein the fungus is selected from the group consisting of Aspergillus, Fusarium, Chrysosporium, and Trichoderma.

24.The method of claim 23, wherein the fungus is Chrysosporium strain UV18-25 having accession number VKM F-3631 D.

25. The method of claim 22, wherein the fungus is Trichoderma longibrachiatum strain X-252.

26. The method of claim 22, wherein the fungus is Aspergillus sojae strain pclA.

25. A method for obtaining multi-gram quantities of a protein having an activity or property of interest, comprising optimizing the activity or property of interest by the method of claim 4, culturing on a large scale an individual organism, clonal culture, or clonal colony isolated in the final step (f), and isolating the expressed protein from the culture.