Applied Microbiology and Biotechnology

, Volume 88, Issue 1, pp 31–39

Yeast flocculation and its biotechnological relevance


    • Institute for Wine Biotechnology, Faculty of AgriSciencesUniversity of Stellenbosch
  • Patrick Govender
    • Department of BiochemistryUniversity of KwaZulu-Natal
  • Michael C. Bester
    • Institute for Wine Biotechnology, Faculty of AgriSciencesUniversity of Stellenbosch

DOI: 10.1007/s00253-010-2783-0

Cite this article as:
Bauer, F.F., Govender, P. & Bester, M.C. Appl Microbiol Biotechnol (2010) 88: 31. doi:10.1007/s00253-010-2783-0


Adhesion properties of microorganisms are crucial for many essential biological processes such as sexual reproduction, tissue or substrate invasion, biofilm formation and others. Most, if not all microbial adhesion phenotypes are controlled by factors such as nutrient availability or the presence of pheromones. One particular form of controlled cellular adhesion that occurs in liquid environments is a process of asexual aggregation of cells which is also referred to as flocculation. This process has been the subject of significant scientific and biotechnological interest because of its relevance for many industrial fermentation processes. Specifically adjusted flocculation properties of industrial microorganisms could indeed lead to significant improvements in the processing of biotechnological fermentation products such as foods, biofuels and industrially produced peptides. This review briefly summarises our current scientific knowledge on the regulation of flocculation-related phenotypes, their importance for different biotechnological industries, and possible future applications for microorganisms with improved flocculation properties.


FlocculationMicrobial cell adhesionGenetic engineeringIndustrial bioprocessesYeast


The adhesion properties of microorganisms, which may involve adhering to other cells, tissues or solid substrates, have been the focus of wide-ranging scientific and biotechnological interest (Verran and Whitehead 2005; Verstrepen and Klis 2006; Zhao and Bai 2009). Adhesion properties govern indeed many essential aspects of the life-cycles of microorganisms, including sexual reproduction (Chen et al. 2007), cellular aggregation during processes such as biofilm formation (Reynolds and Fink 2001; Palmer et al. 2007; Ramage et al. 2009), invasive (Gimeno et al. 1992) and/or pathogenic behaviour (Chaffin 2008; Tronchin et al. 2008; ten Cate et al. 2009), and many others.

Adhesion properties of microorganisms are dependent on characteristics of the cellular surface, usually the outer layer of the cell wall. Microorganisms can adjust adhesion properties by changing the structure and/or composition of this cellular surface. Such changes commonly occur as part of adaptive responses to specific environmental parameters and impart environment-specific adhesion phenotypes. Considering the obvious evolutionary importance of cellular properties that allow microorganisms to control their physical interaction in complex extracellular environments, it is perhaps not surprising that current data show that the genes responsible for imparting such phenotypes are subject to highly complex and diverse regulatory networks. In the yeast Saccharomyces cerevisiae, the best studied of all microbial systems, adhesion mechanisms are regulated through numerous signalling pathways, and genes that impart specific adhesion properties are regulated through combinatorial transcription factor networks as well as allele-specific and epigenetic regulation (Verstrepen et al. 2004; Verstrepen et al. 2005; Octavio et al. 2009). These genes furthermore appear to be a target of strong evolutionary pressure, with data showing that yeast genomes contain a number of pseudogenes that are similar to adhesion genes, and that the mutational frequency within the ORFs of these genes appear higher than those for other genomic sequences (Rando and Verstrepen 2007; Ogata et al. 2008). Data show that even single clonal cellular populations contain cells that display different cell surface phenotypes, an adaptation that may ensure that each population contains a certain percentage of cells that is pre-adapted for sudden environmental changes (Octavio et al. 2009).

Besides the obvious importance of adhesion properties in natural environments, adhesion properties also control phenotypes that have become relevant targets for the improvement of microorganisms in industrial processes. In particular, the natural ability of some microorganisms to display controlled asexual aggregation, also referred to as bio-flocculation, which can lead to the formation of extensive but highly compact groups of cells or flocs, has been exploited by fermentation-based industries to improve various processes, including yeast-based processes such as winemaking (Caridi 2006), brewing (Verstrepen et al. 2003), bioethanol production (Zhao and Bai 2009) and wastewater treatment from both municipal sewage and industrial sources (Liu et al. 2009; Narihiro and Sekiguchi 2007; Vijayaraghavan and Yun 2008).

As a consequence, and initially largely driven by the observed cost-saving nature of ale yeast that generate self-clearing beers at the end of fermentations, research has provided a significant depth of understanding of the genetic, biochemical and physiological aspects of yeast flocculation. Most of these data focus on the most commonly used industrial yeast species, S. cerevisiae. This yeast species is capable of changing the property of its cellular surface leading to a number of specific phenotypes depending on environmental conditions (Fig. 1). Compared to S. cerevisiae, our understanding pertaining to the flocculation of other microorganisms is rather limited. For this reason, much of the background information on the molecular nature of flocculation phenotypes presented here will be based on data generated with this organism.
Fig. 1

S. cerevisiae displays a variety of cell–cell and/or cell-surface-related phenotypes. Shown are phenotypes as observed under defined laboratory conditions

Flocculating microorganisms of industrial relevance

Yeast flocculation

Yeast flocculation is defined as the asexual, reversible and calcium-dependent aggregation of yeast cells to form flocs containing large numbers of cells that rapidly sediment to the bottom of the liquid growth substrate (Bony et al. 1997; Stratford 1989). Unlike brewing, other industrial fermentation processes primarily employ non-flocculent strains of S. cerevisiae to produce a wide range of commodities that include wine, bioethanol and small-molecule metabolites of such as insulin (Kjeldsen 2000), l-lactic acid (Saitoh et al. 2005) and polyketides (Maury et al. 2005; Kealey et al. 1998). These industries resort to the use of expensive separation procedures such as centrifugation and/or filtration to remove cells. Consequently, there exists a major research impetus to modify the flocculation profile of industrial S. cerevisiae strains so that they could be of similar benefit to related industries. Similar strategies could also be of financial benefit to other biotechnological industries that employ bacteria to generate biodiesel, biopolymers, enantiopure pharmaceuticals, agrochemicals and flavour compounds.

Bacterial flocculation

In contrast flocculation involving bacteria is rather ill defined. The extra cellular interaction systems, as well as the mechanisms that control them, have not been worked out in such detail as for S. cerevisiae. Thus far, bacterial flocculation has been shown to be of importance in the process of sludge treatment during sewage processing. Affecting the physico-chemical properties of sludge flocs it improves denser and stronger aggregate structure settling and ensured solid-effluent separation (for a recent review, see Liu et al. 2009). Many of these bacteria from sludge have been identified (for a recent review, see Nielsen et al. 2009). Evidence suggests that bacterial flocculation involves extra cellular polymeric substances (Subramanian et al. 2008) with properties not unlike the lectins and adhesins identified for yeast (Park and Novak 2009). The study from Park and Novak (2009) notably highlights the role of extra cellular glycoproteins together with the suggested involvement of divalent ions (ethylenediaminetetraacetic acid inhibition of flocculation), and both these factors are required for lectin-based yeast flocculation.

The molecular regulation of adhesion in yeast

The mechanism of flocculation in S. cerevisiae

The revised lectin hypothesis as presented by Teunissen and Steensma (1995) proposed that specific surface cell-wall-anchored glycoproteins or flocculins specifically recognise and bind to α-mannan carbohydrates of adherent yeast cells. Furthermore, Miki et al. (1982) suggests that Ca2+ ions act as cofactors in maintaining the active conformation of surface proteins, thereby enhancing the capacity of lectins to interact with α-mannan carbohydrates. Flocculation (Flo) protein members, all carrying lectin domains, play a central role in this process as is illustrated by the specific expression of FLO genes that can transform non-flocculent S. cerevisiae strains into flocculent ones (Watari et al. 1994; Cunha et al. 2006; Wang et al. 2008; Watari et al. 1991; Chambers et al. 2004; Guo et al. 2000; Verstrepen et al. 2001; Govender et al. 2008). Furthermore, Flo proteins are involved in all of the phenotypes depicted in Fig. 1 (Verstrepen and Klis 2006), thereby suggesting a link between “pathogen-like” behaviour and both cell–cell and cell–substrate interactions. Although lectins have clearly been shown to be involved in flocculation, it is still unclear how lectins and ligands specifically interact.

Flocculation phenotypes of S. cerevisiae

The interactions mediated by Flo glycoproteins can be divided into two categories namely lectin-like (cell to cell adhesion) and sugar-insensitive (adhesion to abiotic surfaces) adhesion phenotypes (Verstrepen and Klis 2006). This review will focus on the cell–cell adhesion (flocculation) phenotypes which can be divided into three sub-categories depending on the degree this phenotype is inhibited by the presence of certain sugars (sugar sensitivity) (Masy et al. 1992; Stratford and Assinder 1991). The Flo1-phenotype is constitutively flocculent, exclusively mannose-sensitive and is generally associated with laboratory S. cerevisiae strains that express FLO1, FLO5 and FLO9-encoded flocculins (Govender et al. 2008; Guo et al. 2000; Teunissen and Steensma 1995; Van Mulders et al. 2009). The NewFlo-phenotype is typically associated with brewing strains that display stationary phase-specific flocculation that is both glucose- and mannose-sensitive. The NewFlo-phenotype is aligned with yeast strains expressing FLO1 homologues corresponding to either Lg-FLO- or FLONS-encoded flocculins (Kobayashi et al. 1998; Liu et al. 2007). A third group of strains in which flocculation is mannose-insensitive (MI) and independent of Ca2+ ions is also reported (Masy et al. 1992). Masy et al. speculated that flocculation in these strains could be the result of hydrophobic interactions or other specific interactions not involving mannans. Alternatively, Stratford and Rose (1992) suggested that mannose-insensitivity probably results from very low specificity to monosaccharides since lectins may have much greater affinity for tri- or polysaccharides than for simple sugars. It is nonetheless highly probable that the flocculation mechanism in these strains would differ from the modified-lectin mechanism of Flo1 and NewFlo strains. Recent studies from our lab seem to indicate the involvement of Flo11p in the generation of an MI flocculent phenotype (data not shown).

FLO gene encoded adhesins or flocculins

Analysis of the genomic sequence of the haploid S. cerevisiae laboratory strain S288C reveals five unlinked dominant FLO genes (Verstrepen et al. 2004), with four of these genes; FLO1, FLO5, FLO9 and FLO10 located adjacent to telomeres (Teunissen and Steensma 1995). In addition, a fifth gene, MUC1 (Lambrechts et al. 1996) also known as FLO11 (Lo and Dranginis 1996) is located in a more central chromosomal position.

The FLO genes encode cell wall proteins that are collectively referred to as adhesins and they are characterised by a common modular organisation that consists of three domains. Firstly, an amino-terminal domain (A) that is proposed to harbour the binding site to carbohydrate receptors (mannan) which confers adhesion (Kobayashi et al. 1998). Initially attached to this domain is a secretory sequence that is removed as the protein migrates through the secretory pathway to the cell wall. This is followed by a central domain (B) that is extremely rich in serine and threonine residues (Caro et al. 1997), and thirdly, a carboxyl-terminal region (C) that contains a site for covalent attachment of a glycosyl phosphatidylinositol anchors (Caro et al. 1997; De Groot et al. 2003; Hamada et al. 1998). A comparison of putative Flo proteins reveals that Flo5p, Flo9p and Flo10p share 96%, 94% and 58% similarity respectively to Flo1p (Teunissen and Steensma 1995). In contrast, the Flo11p adhesin is the most divergent member of the family and displays 37% similarity to Flo1p (Lo and Dranginis 1996).

Genetic regulation of flocculation

Intensive studies on the transcriptional regulation of the FLO11 gene and to a far lesser extent that of FLO1 reveals that S. cerevisiae has evolved sophisticated mechanisms to sense and respond to environmental signals by activating developmental switches that result in coordinated changes in cell physiology, morphology and cell adherence (Gagiano et al. 2002; Verstrepen and Klis 2006; Rubio-Texeira et al. 2010). The data suggest that these genes are subject to a large diversity of regulatory mechanisms, and are evolving at a much faster rate than other genome sequences (Rando and Verstrepen 2007).

The expression of FLO11, the gene for which most information has been generated, is regulated in response to nutrient availability and many other stress factors (Bauer and Pretorius 2000; Verstrepen and Klis 2006). The three best characterised signalling pathways that are implicated in transmitting the nutritional status of the environment to the promoter of FLO11 are the cAMP-PKA pathway (Tamaki 2007), the invasive growth MAP-kinase (Chen and Thorner 2007) and the main glucose repression pathway (Gancedo 1998; Kuchin et al. 2002). These pathways regulate the gene via a set of transcriptional activators and repressors that include Flo8p (Kobayashi et al. 1999), Ste12p and Tec1p (Madhani and Fink 1997; Bardwell et al. 1998) as well as Sfl1p (Conlan and Tzamarias 2001) and Mss11p (van Dyk et al. 2005). For an extensive review on FLO11 regulation by these pathways, see reviews by Palecek et al. (2002) and Verstrepen and Klis (2006). The data suggest that the above mentioned signalling pathways and regulatory proteins converge on this promoter to regulate the primary FLO11 phenotypes of invasive growth and pseudohyphal differentiation (Rupp et al. 1999; Gagiano et al. 1999). Moreover, these investigations have shown that FLO11 transcriptional regulation is particularly dependent on the nutritional status and the specific composition of the growth environment (Gagiano et al. 2002).

Far less information is available regarding the regulation of other FLO genes. However, Verstrepen and Klis (2006) suggested that other S. cerevisiae FLO genes may be controlled by similar (but not identical) pathways to those that control FLO11. This could well be the case as it has been reported that both Flo8p (Kobayashi et al. 1999) and Mss11p (Bester et al. 2006) act as transcriptional activators of both FLO1 and FLO11. Data by Fleming and Pennings (2001) show that the Swi-Snf co-activator and Tup1-Ssn6 co-repressor control an extensive domain (>5 kb) in which regulation of the FLO1 gene takes place. The promoter region of FLO1 was also observed to contain a putative GCN4-box at position 268 and numerous stress responsive heat-shock elements (Teunissen et al. 1993). Flocculation is controlled by nutritional status signals such as carbon and/or nitrogen starvation (Sampermans et al. 2005) and other environmental indicators such as pH (Soares and Seynaeve 2000) and ionic strength (Jin and Speers 2000). It is likely that some of these factors, especially those related to nutrient status, also affect FLO1 expression, but no molecular data are available.

Besides FLO gene activity being regulated at the transcriptional level, it has also been shown to be modulated by other regulatory systems. In particular, data suggest that these genes are often under promoter-specific epigenetic control allowing S. cerevisiae cells in a homogenous population to reversibly switch between active FLO gene expression and silent modes (Halme et al. 2004; Octavio et al. 2009).

Industrial applications

Table 1 presents a summary of the various industrial applications of modified flocculent yeasts. Some of these applications are briefly discussed in the following section.
Table 1

Optimisation of flocculation phenotypes for industrial processes involving Saccharomyces and non-Saccharomyces yeast strains



Genetic modification



Kluyveromyces marxianus

Ethanol production from molasses

TDH3 promoter-based constitutive expression of S. cerevisiae FLO1, FLO5, FLO9 or FLO10 genes

Ferments arabinose, cellobiose and xylose at 40 °C.

(Nonklang et al. 2009)

Hybrid Saccharomyces pombe and S. cerevisiae strain

Ethanol production from saccharified corn powder

Inter-species hybridization strategy

Biomass recovery in a continuous system

(Xu et al. 2005; Zhao and Bai 2009)

S. cerevisiae

Beer and wine production

ADH2 or HSP30 promoter-mediated controlled expression of the native FLO1, FLO5, or FLO11 open reading frames

Induction of flocculation upon the completion of fermentation for the production of clearer beer and wine

(Verstrepen et al. 2001; Govender et al. 2008; Govender et al. 2010)

S. cerevisiae

Beer production

ADH2 promoter-mediated controlled expression of the FLONS gene

Production of beer with lower diacetyl content

(Wang et al. 2008)

S. cerevisiae

Ethanol production

ADH2 promoter-mediated controlled expression of the FLO1 open reading frame

Decreased ethanol production

(Cunha et al. 2006)

S. cerevisiae

Biodiesel production from soybean oil and methanol

Expression of chimeric protein consisting of Rhizopus oryzae lipase fused to native Flo1p

Whole-cell factory in a solvent-free system for the methanolysis reaction.

(Matsumoto et al. 2002)

S. cerevisiae

Ethanol production from starch

Expression of chimeric protein where Flo1p N-terminal domain is replaced by Rhizopus oryzae glucoamylase

Increased amylase substrate interaction achieved by this presentation method

(Sato et al. 2002)

S. cerevisiae (S. cerevisiae and S. diastaticus parents)

Ethanol production from starch

Expression of chimeric protein consisting of the Rhizopus oryzae glucoamylase fused to the sexual agglutinin Aga1p cell wall anchor

“high level” ethanol production by flocculent strain

(Kondo et al. 2002; Murai et al. 1997)

S. cerevisiae

Ethanol production from molasses

ADH1 promoter-based constitutive expression of the FLO1 gene

Increased ethanol production

(Ishida-Fujii et al. 1998)

S. cerevisiae

Production of ethanol and fructose from hydrolyzed Jerusalem artichoke extracts

ADH1 promoter-mediated constitutive expression of the FLO1 gene

99% enrichment of fructose

(Remize et al. 1998)

S. cerevisiae

Wine production

Rare-mating hybridisation strategy

Increased fermentation efficiency and less turbid white wines

(Lahtchev and Pesheva 1991)

S. cerevisiae

Beer, whisky, wine, sake, and shochu production

Constitutive multicopy plasmid-based FLO1 expression

Increased flocculation of all industrial strains tested

(Watari et al. 1991)

Bioremediation with yeast

As already mentioned the dominant flocculation phenotype requires the presence of Ca2+ ions. Thus yeast that flocculates effectively removes Ca2+ ions from the medium as the flocs sediment out of the liquid. In the same manner, it was shown that flocculent yeast is able to remove other divalent metal ions, namely Cu2+, Ni2+ and Zn2+, from solution (Machado et al. 2008; Machado et al. 2009). Furthermore, these authors show that dead, but still intact yeast cells more effectively remove these same ions. This could be through the further interaction of intracellular components with metal ions. Machado et al. (2008) also refers to five other studies that observed metal ion fixation by either live or dead yeast cells. These data highlight the clear potential of native or modified yeast to be employed as biological “scrubbers” at heavy metal contaminated sites or of contaminated industrial effluent.

Bio-processing with flocculating yeast

The brewing industry has been utilising the natural flocculating ability of some S. cerevisiae brewing strains to cost-effectively separate biomass from the fermented product upon completion of fermentation. Various immobilisation systems have also been used for yeast (Verbelen et al. 2006), but flocculation has proven to be a cheap and relatively easy way of separating yeast from the fermented product. In addition cells within flocs have been shown to display increased survival rates due to the protective nature of the floc structure (Smukalla et al. 2008), important when the aim is to re-use yeast in subsequent fermentations (re-pitching of yeast). Primarily due to preference of S. cerevisiae for simple hexoses (glucose, fructose and maltose) sugars from raw materials such as barley, corn, grape must, wheat and rice, its use is mainly restricted to alcoholic-beverage producing and baking industries. Through the use of engineered flocculating strains great improvement in flocculation behaviour has been demonstrated with optimised industrial processes as the goal. Although of no industrial significance as yet, the ability to naturally flocculate has also been demonstrated by other yeast genera which include Hansenula anomala (Saito et al. 1990), Kluveromyces bulgaricus (Almahmood et al. 1991), Kluveromyces lactis, and Schizosaccharomyces pombe (El-Behhari et al. 1998; El-Behhari et al. 2000). The growing global demand for ethanol as an alternative fuel and the need for yeast that are capable of fermenting other sugars than glucose, fructose and maltose (xylose, arabinose and cellobiose) has driven and will continue to drive the development of these strains.

Future improvements of flocculating phenotypes

Our current knowledge on the molecular nature and regulation of adhesion systems is mainly based on studies of S. cerevisiae and some human pathogenic yeast and fungi (Verstrepen et al. 2004; Wheeler and Fink 2006; Chaffin 2008; Tronchin et al. 2008; ten Cate et al. 2009). This knowledge has already led to the engineering of strains with improved attributes to help optimise industrial processes. The current increase in high throughput DNA sequencing and -omics technologies will lead to the discovery and characterisation of novel industrially relevant adhesion interaction systems of non-S. cerevisiae organisms within the near future, allowing us to expand such engineering approaches to other species.

Up to now, the improvement strategies targeted the expression levels of individual flocculation genes. There is strong evidence however that adhesion phenotypes are dependent on more than single adhesins (Guo et al. 2000). Thus a better understanding of the interaction between various factors is required to allow strategies that would coordinate regulation of more than one of these factors to further refine flocculation phenotypes. We have mentioned the complex epigenetic and conventional control exerted by multiple transcriptional regulators over FLO11 expression (Octavio et al. 2009). Not only do these factors control the level of heterogenicity of flocculation within a clonal population, but evidence exist for the combinational control of more than one FLO gene (Halme et al. 2004; Fichtner et al. 2007; Bester et al. 2006). Biotechnological “tweaking” of the levels of these regulators thus could be a very attractive strategy to further refine the flocculation phenotype and improve the timing of expression and/or the possible optimal level of heterogeneous expression within a population during industrial processing.

A significant proportion of adhesin genes contain internal tandem repeat regions that display length instability upon an increase in generation number. Currently, it is not clear what molecular mechanism controls these mutational frequencies that lead to great phenotypic variability over relatively short evolutionary periods (Verstrepen et al. 2005; Rando and Verstrepen 2007). Such variation could prove to be problematic if a consistent phenotype is required over a prolonged time period, for instance with the constant re-pitching of brewing yeast in the production of beer. Thus engineering strains with stable (1) internal FLO repeat lengths as well as (2) epigenetic control over expression will likely result in strains that retain specific and desired flocculation phenotypes over multiple generations.

Concluding remarks

Our knowledge regarding the molecular nature of the flocculation process in S. cerevisiae has enabled the design of specific modification strategies that can lead to desirable flocculation phenotypes. However, since this process is of a highly complex and species-specific nature, the insights gained in this yeast may not result in applicable strategies for other organisms. Future studies will have to focus on improving our understanding of the combinatorial impact of cell wall proteins, other cell-wall-modifying factors, and the impact of matrix components on specific flocculation properties. The complexity of this challenge will require the use of systems-based approaches to improve our knowledge of the molecular nature and regulation of adhesion phenotypes in microorganisms of industrial interest. Such knowledge will enable the future development of organisms with desired as well as stable flocculation profiles.


This work was financially supported by the National Research Foundation (NRF) and the South African Wine Industry (Winetech).

Copyright information

© Springer-Verlag 2010