Kluyveromyces lactisand the 14th Workshop ‘Biology ofKluyveromyces’

This annual workshop is one of the few yeast meetings dedicated to a non-conventional species. Its 14th meeting was held on July 6–8, 2001, at the Orsay campus of the University of Paris XI. Originally, the workshop was created in 1988 as an instrument for the network of European laboratories involved in the European Biotechnology programmes. With its modest size (generally oscillating around 50–80 participants), the meeting is conveniently held at different European cities. Although it receives a small ad hoc contribution from a few companies, the cost is mostly covered by the participants. On some occasions the workshop took the form of a satellite meeting parasiting large yeast conferences like ICYGMB, because this allows the scientists of other continents to participate without additional costs. In the 1960s, H.O. Halvorson (Wisconsin) began to use Kluyveromyces lactis to study β-glucosidase regulation. Since then, this organism always had its advocates, but remained in the shadow of the prestigious Saccharomyces cerevisiae. It is only after 1980 that K. lactis research became more visible, probably due to the interest of three unrelated topics: (i) the killer system based on linear DNA plasmids, (ii) the lactose metabolism and (iii) the secretion of recombinant proteins. At present an increasing number of people working on S. cerevisiae try to extend their thematic enquiries to other yeasts, but the choice is limited. K. lactis is practically the only alternative that can offer all the possibilities of a genetic approach, the other ‘genetic yeasts’, Schizosaccharomyces pombe and Yarrowia lipolytica, being too distant from S. cerevisiae. Rather than providing a report dealing only with the last meeting (which was much troubled by an airline strike), I may give below a quick summary, admittedly partial and biased, of the topics which have often been discussed in the Kluyveromyces workshops over the last 2–3 years. This system is based on the plasmid couple pGKL1/pGKL2. It has two interesting features. First, linear DNA plasmids were a novelty for the yeast kingdom in 1981 (N. Gunge, Kumamoto). Later, it turned out that linear DNA plasmids were widely spread among many yeast genera (H. Fukuhara, Orsay). Second, the killer toxin, encoded by the plasmid pGKL1, arrests the growth of the target cell at G1 phase. Also interesting is the fact that they have a protein-primed mode of replication adopted by adenoviruses. Most, but not all, of the plasmid genes have been identified for their possible functions. The latest in the list, as reported in a recent workshop, are the gene coding for an RNA capping enzyme (F. Meinhardt, Münster) and the gene encoding a single-stranded DNA binding protein (R. Schaffrath, Halle, and P.A. Meacock, Leicester). Since these plasmids replicate in the cytoplasm (as opposed to the nuclear 2-μm plasmid), they have their own transcription machinery, which does not recognise host chromosomal genes. Several important questions remain to be answered for this cytoplasmic system. The rules of plasmid transmission, segregation and recombination are not known, although some factual descriptions are found in early papers. Now that expressible genetic markers can be introduced by plasmid shuffling, there are ways to study these questions. This may provide a model for the genetics of mammalian cytoplasmic DNAs. The mode of action of the killer toxin is different from that of the S. cerevisiae system, and a few laboratories are working on the action mechanism which involves a plasmid-encoded chitinase (H. Kitamoto, Tsukuba, and Schaffrath, Halle). Lactose is a preferred sugar of K. lactis. Although a number of yeasts can aerobically grow on lactose, those that can ferment it are rare. Lactose assimilation in K. lactis is an inducible system triggered by either lactose or galactose. The regulatory circuit has been worked out mostly by two groups (R.C. Dickson group, Lexington, and K.D. Breunig group, Halle). Clearly the lactose regulon is a variation of the galactose regulon of S. cerevisiae. The positive regulator LAC9 (KlGAL4) is an equivalent of the famous GAL4, and interacts with KlGAL80. The LAC4 (encoding β-galactosidase) and LAC12 (lactose permease) genes, absent in S. cerevisiae, are contiguously placed in the K. lactis genome sharing a divergent promoter (co-transferred from external origin?). Several laboratories have constructed S. cerevisiae strains capable of fermenting lactose by introducing both LAC4 and LAC12. The lactose regulon is repressed by glucose in a particular lineage of strains, but not in most laboratory strains. The absence or weakness of glucose repression on many cellular processes (such as synthesis of respiratory enzymes) is a prominent feature of K. lactis physiology, in contrast to S. cerevisiae in which glucose repression is a dominant device of regulation (see below). As of year 2000, one may consider that the main features of the lactose regulon are well understood, even if questions remain concerning its relation to higher order cellular regulation. An important fraction of K. lactis research is focused on the regulation of carbon metabolism (glycolysis, glucose repression, gluconeogenesis, respiration/fermentation switch). There are two main reasons for this choice. Firstly, K. lactis and S. cerevisiae, which diverged more than 100 Myr ago, share similar sets of genes, but show very different patterns of carbon flow regulation. Thus the evolution or variation of regulatory mechanisms may be traceable by comparison of these species. Secondly, K. lactis has turned out to be an organism of choice to obtain glycolysis mutations, because fermentation is dispensable in K. lactis whereas glycolytic mutants of S. cerevisiae are often very sick. The pentose shunt seems to be very active in K. lactis. We may recall that the study of respiratory mutations had been best carried out with S. cerevisiae, precisely because its respiration was dispensable. Another factor that facilitated isolation of glycolysis mutants in K. lactis is the fact that the genes involved in this process are not redundant, whereas the fermentation-oriented S. cerevisiae has most of these genes duplicated. Even for glucose uptake, K. lactis has only one permease for high-affinity transport and one for low-affinity transport. Glycolysis mutants are easy to isolate because they grow poorly on the glucose plates containing a respiratory inhibitor. Some 20 complementation groups are known, including regulatory genes. For the latter category, a new finding, reported at the last meeting, is that a casein kinase I (Rag8p) regulates glycolysis through an Sgc1p-equivalent DNA binding protein (Wésolowski–Louvel group, Lyon). Recent studies show many examples of the differences in regulation between K. lactis and S. cerevisiae. KlCat8p has no apparent role in gluconeogenesis in K. lactis, and regulates acetyl CoA synthetase genes in a manner distinct from its S. cerevisiae counterpart (Breunig group at Halle; I. Ferrero/P. Goffrini group at Parma). Neither disruption of MIG1 gene (Vandenhaute group, Namur) nor disruption of HAP2/3/4/5 complex (Grivell group, Amsterdam; Bolotin–Fukuhara group, Orsay) showed an obvious phenotype in K. lactis, except in a few cases. The lactate-inducible lactic dehydrogenase/permease system is also regulated differently (T. Lodi at Parma, and B. Guiard at Gif-sur-Yvette). The genes orthologous to SNF1, SNF3 and SIP4 have also been identified recently (Halle and Parma groups) whose deletion phenotypes suggest that their target gene spectra are different in K. lactis. The ethanol-inducible, glucose non-repressible mitochondrial alcohol dehydrogenase (KlADH4) is another example of K. lactis specificity (Falcone group, Rome). Such differences may be closely correlated with the modest role of glucose repression in this yeast, but a possible unifying formulation of all these differences is missing. Clearly, the fermentative physiology of S. cerevisiae cannot be readily extrapolated to other yeast species, most of which have a much more aerobic style of life. The aerobic yeasts like Cryptococcus or Yarrowia do not ferment at all on any sugars. K. lactis is conveniently positioned intermediately between them and S. cerevisiae, and appears to be a good instrument for the manipulation of the respiration–fermentation switching mechanisms. For example Crabtree effect-positive/negative conversion of K. lactis has been obtained by Y. Steensma group (Leiden), in which acetyl CoA synthetases and pyruvate dehydrogenase were shown to be key elements. Many yeasts are Kluyver effect-positive, in other words they cannot grow on certain sugars in a fermentative mode (galactose, maltose, raffinose, etc). To assimilate these sugars, respiration seemed to be required for unknown reasons. This long-standing question, which has important issues in the industrial utilisation of non-glucose sugars, has perhaps found its explanation according to a recent work on K. lactis (C. Donnini group, Parma). The proposed answer is that yeast does not always have a high enough level of permease activities for uptake of certain sugars. Since fermentative growth on sugars demands a much higher flow of substrate than does the respiratory growth, the permease activities become limiting. By introduction of additional doses of permease genes, K. lactis became capable of growing fermentatively on galactose and other ‘Kluyver effect sugars’. S. cerevisiae is Kluyver effect-negative on most sugars it assimilates, but after disruption of the main galactose permease gene GAL2, it became Kluyver effect-positive, growing on galactose solely in a respiratory mode. S. cerevisiae does not usually secrete much protein into media, whereas many other species do (K.-K. Huo and Y.-Y. Li, Shanghai). K. lactis also secretes little protein, but its potential to secrete large proteins was predictable by the fact that the pGKL1 killer toxin protein contained a very large subunit. The exoprotease level is negligible. Demonstration of g l−1 level production of recombinant calf prochymosin in 1990, followed similarly by high secretion of human serum albumin in 1991, has prompted many laboratories to try to make other recombinant proteins. A great challenge now is the production of ‘difficult’ proteins of mammalian origin. For example a few groups (M. Dion group at Nantes; A. Fournier group at Paris) are attempting to obtain single-chain antibody fragments (ScFv) in a secreted form, seemingly with modest success so far. Attempts to target the periplasmic enzymes into media, such as β-galactosidase, is another ambitious challenge (E. Cerdan group, La Coruña). K. lactis, Hansenula polymorpha and Pichia pastoris are the three favourite species for recombinant protein production, often compared for their respective performance. Regretfully, the detail of protein secretion mechanisms is little studied in these yeasts. However, a few secretory mutants began to emerge (C. Palleschi group, Rome; W.-G. Bao, Orsay; Galeotti, Siena). If we suppose that these yeasts have a secretory system analogous to that of S. cerevisiae, the species Y. lipolytica, in contrast, might represent a quite different host as its main secretory pathway functions in an SRP-dependent, co-translational mode similar to that of the mammalian systems (C. Gaillardin, Grignon). We may need a rational basis in the choice of yeast species as a production host for each recombinant protein. The molecular manipulation of genes in K. lactis started in the 1980s by the isolation of K. lactis ARSs and a 2-μm type circular plasmid pKD1. From this plasmid, many replicative vectors have been developed. These tools are now almost as complete as those for S. cerevisiae, including K. lactis–S. cerevisiae shuttle vectors and shuttle libraries. For K. lactis, as for many other yeasts, large-scale genome mapping and sequencing are in progress with random sequence tag collections and BAC libraries. The K. lactis genome, about the size of the S. cerevisiae genome, contains six chromosomes. On issue of the French ‘Génolevures’ project (FEBS Lett. 487(1), December 2000 special issue, in which the genomes of 13 hemiascomycetous yeasts were compared including K. lactis, Kluyveromyces marxianus and Kluyveromyces thermotolerans), more than 2200 genes of the K. lactis genome have been identified (http://cbi.labri.u-bordeaux.fr/Genolevures/Genolevures.php3). Most of them were orthologues of S. cerevisiae genes, but some seemed unique to K. lactis (Bolotin–Fukuhara group, Orsay; M. Wésolowski–Louvel group, Lyon; O. Ozier–Kalogéropoulos, Paris). It is hoped that a partial or full set of gene microarrays will soon become available. K. Wolfe (Dublin) had proposed the attractive idea that the S. cerevisiae genome might have arisen through a total genome duplication of a K. lactis-like prototype genome. However, the Génolevures people seem to think differently after a comparative analysis of partial genome sequences of the above 13 species. This still limited set of sequence data may not facilitate the debate. The Kluyveromyces workshop, as a European network, does not have a regular participation of the researchers from other continents where many more topics are being studied with K. lactis. To quote a few, we may mention telomere-maintaining system (M.J. McEachern/E. Blackburn group, Athens, GA, USA), glycosylation (C. Abeijon group, Worcester, MA, USA), mitochondrial ATPase (X.-J. Chen–G.D. Clark-Walker group, Canberra), G-proteins (R. Coria group, Mexico), multi-drug resistance genes (J. Subik group, Bratislava and X.-J. Chen, Canberra, or variant insulin production (Y.-M. Feng, Shanghai). A regularly updated address list of the K. lactis research community is available (hiroshi.fukuhara@curie.u-psud.fr) and waiting for new members. The 2002 workshop will take place on September 6–8 (a weekend as usual) at the Smolenice castle in Slovakia. The organiser will be Julius Subik, Comenius University, Bratislava (subik@fns.uniba.sk).