The yeast cell wall is a highly dynamic structure that is not only an armour separating cell from its surrounding but that is endowed of surface properties including adherence to inert material leading to biofilm formation, cell-cell adhesion that can yield to flocculation, hydrophobicity which may result in buoyant biofilms. These properties are mediated by a variety of surface proteins called adhesins (Lipke et al., 2018), among which flocculins are adhesins encoded by the FLO genes family (FLO1, FLO5, FLO9, FLO10, and FLO11) in the yeast Saccharomyces cerevisiae (Verstrepen and Klis, 2006; Brückner and Mösch, 2012). In the pathogenic yeast Candida albicans, it has been reported that these adhesins can be organized into clusters of hundreds of proteins at the cell surface leading to adhesion nanodomains. This 3D organization of adhesins into nanodomains has been shown to be triggered by an external physical shear force, such as the extension force of single molecules stretched by an atomic force microscopy (AFM) tip, which then propagates across the cell surface at a speed of about 20 nm/min (Alsteens et al., 2010). This unique phenomenon was explained by the presence of small amyloid-core sequences of five to seven residues (IVIVATT) in adhesin proteins that are also characterized by a high content of β-branched aliphatic Ile, Val, and Thr forming β-aggregates structures as predicted by the β-aggregate predictor TANGO (Fernandez-Escamilla et al., 2004b; Lipke et al., 2018). The formation of these nanodomains is therefore an emergent property of adhesin primary sequence that promotes cell-cell aggregation, resulting in the formation of robust biofilms and eventually promotes fungal infection (Lipke et al., 2018). More recently, Lipke, Dufrêne, and colleagues developed a new pipette technology called fluidic force microscopy enabling to measure forces between single cells (Dehullu et al., 2019b). Using this methodology, these authors showed that amyloid-forming sequences have a dual role, namely in the formation of adhesin clusters (nanodomains), which correspond to cis-interactions between molecule of the same cell, and in homophilic adhesion between adhesins on opposing cells (trans-interaction) (Ho et al., 2019; Dehullu et al., 2019a).

In a previous work aiming to investigate the impact of autolysis process on the nanomechanical properties of the cell wall of different S. cerevisiae strains, we identified by AFM the presence of abundant patches with a mean diameter of 140 nm at the surface of an industrial wine yeast strain (Schiavone et al., 2015). In addition, we found that these patches were formed from highly mannosylated proteins as they interacted strongly with a concanavalin A functionalized AFM tip, showing distances rupture ranging from 50 to 500 nm. Overall, these nanostructures were reminiscent of the adhesion nanodomains formed by the clustering of Als5 adhesins on the surface of C. albicans (Alsteens et al., 2010). In this work, we addressed the question of whether the formation of these patches could result from the aggregation of flocculins, and if this is the case, then which of the flocculins could be responsible, taking into account that a transcriptomic analysis carried out in L69 strain revealed notably a high expression of FLO11, although FLO10, FLO5, and FLO1 were also expressed albeit at lower levels (Schiavone et al., 2015). These flocculins share a common three-domain structure. The C-terminal part (C-domain) contains a glycosylphosphatidylinositol (GPI) anchoring site for covalent attachment of the protein to β (1,6)-glucans of the inner cell wall network (Lu et al., 1995; Bony et al., 1997; Kapteyn et al., 1999). The middle part, stalk-like B-domain, contains serine and threonine enriched tandem repeats (TRs) encoded by conserved DNA sequences that serve as a source of variability triggered by frequent recombination events by which new adhesion/flocculin alleles could be generated (Verstrepen et al., 2005). The N-terminal part (A-domain) which protrudes from the cell surface allows to classify these flocculins into two groups based on their phenotypic properties. The first type includes Flo1, Flo5, Flo9, and Flo10 proteins, which mediates flocs formation. This phenotype is brought about by the A-domain that harbours a unique Ca²⁺ binding motif DcisD for carbohydrate binding (Kraushaar et al., 2015; Veelders et al., 2010). The other group is represented only by Flo11p, which is implicated in several phenotypes including filamentation, flor and biofilm formation (reviewed in Brückner and Mösch, 2012). Like C. albicans adhesins, these S. cerevisiae flocculins contain β-aggregation sequences of five to seven amino acids length that are rich in Ile, Val, and Thr. However, the distribution of these repeats along the flocculins sequence enables to distinguish Flo1p, Flo5p, and Flo9p from Flo10p and Flo11p. For the three former proteins, the β-aggregation-positive sequences predicted by TANGO software (http://tango.crg.es/) revealed a large number of ‘T(V/I)IVI’ motifs that are widely distributed over the B-domain (see details in Supplementary file 1a). Interestingly, a synthetic peptide ‘TDETVIVIRTP’ containing this motif was shown to form amyloid fibers in vitro (Ramsook et al., 2010). On the other hand, β-aggregation-prone sequences are less abundant in Flo10p and Flo11p and they are all located in the C-terminus. Nevertheless, it was reported that a 1331-residue soluble Flo11p can assemble into amyloid fibers in vitro and this feature was ascribed to the presence of the amyloid-β-aggregation-prone sequences ‘VVSTTV’ and ‘VTTAVTT’ at the C-terminus of Flo11p (Ramsook et al., 2010).

The purpose of this work was therefore to investigate the molecular basis and the physiological function of these patches formed on the cell surface of this industrial wine yeast strain. This study was furthermore motivated by the fact that such abundant nanostructures have never been physically observed before in Saccharomyces species, despite the finding that cell-cell aggregation are potentiated under hydrodynamic shear forces and these interactions are antagonized by anti-amyloid dyes (Ramsook et al., 2010; Chan and Lipke, 2014; Chan et al., 2016). Collectively, the reported results lend support to the model that amyloid-forming sequences play a dual role in Flo11p-mediated cell-cell adhesion, namely promoting cis-interaction between Flo11p molecules in the same cell, leading to adhesive nanodomains, and contributing to trans-interaction between opposing cells. They further showed that this dual role is highly dependent on the number of amyloid-core sequences present in the Flo11 protein.

In this work, we showed that the abundant patches observed by AFM on the cell surface of the industrial wine yeast L69 strain (Schiavone et al., 2015) correspond to adhesion nanodomains, with nanomechanical properties very similar to those formed on the cell surface of the pathogen C. albicans (Alsteens et al., 2010; Formosa et al., 2015b). Moreover, we reported that the flocculin encoded by FLO11 is the sole protein involved in the formation of these nanodomains, which can be prevented upon treatment of the cells with anti-amyloid peptides or anti-amyloidophilic dyes. Altogether, these data are in line with the model of Lipke et al., 2018; Lipke et al., 2012 arguing that adhesion nanodomains require cell wall proteins that must harbour in their sequence the following features: (i) short amyloid-β-sheet interaction sequences of five to seven amino acid residues and (ii) serine/threonine-rich ‘T domain’ enriched of β-branched aliphatic amino acids Ile, Val, and Thr. While all flocculins encoded by FLO1, FLO5, FLO9, and FLO11 in S. cerevisiae display these features, it was intriguing to know why such dense and abundant nanodomains had not been observed so far in this yeast species so far, even though physical FLO-dependent modification of cell surface upon hydrodynamic shear force has been recorded (Chan and Lipke, 2014; Chan et al., 2016). The data reported in this work clearly showed that the Flo11 protein of the S. cerevisiae strain L69 differs from that of other Saccharomyces strains sequenced to date by having an additional sequence of ~200 amino acids length (RR2 sequence) near the C-terminus. This sequence probably originates from a twofold repeat of a 90–100 amino acids sequence present near the C-terminus of all Flo11 flocculins that contains the amyloid-forming sequence ‘VVSTTV’ followed by ‘ITTTFV’. This repeat provides therefore two additional amyloid-core sequences to the Flo11p of strain L69, which are clearly determinant in the process of nanodomains formation. Indeed, we showed that the insertion of this RR2 sequence into the Flo11p of BY4741 allowed this strain to form adhesion nanodomains with nanomechanical properties similar to those formed on the cell surface of strain L69.

Our AFM analysis revealed that these nanodomains encompass two components: a first one characterized by a high adhesion force, which corresponds to amyloid nanodomains as they are disrupted under the action of anti-amyloid agents, while the other component has low adhesion force which likely corresponds to isolated Flo11 proteins and possibly other cell wall proteins that may unfold under the action of the AFM tip. This behaviour is fully reminiscent to the nanodomains formed by adhesins on the cell surface of the pathogen C. albicans (Formosa et al., 2015bFormosa et al., 2015b). This analysis also highlighted that, while the amyloid-core sequences are essential in the formation of adhesion nanodomains, other parts of the Flo11p sequence also contribute to their production. In particular, the removal of about half of the TRs in the B-domain almost completely abolished nanodomains formation, while the highly adhesive component of these nanodomains was no longer observed after removal of the N- or C-terminus of Flo11p. These results indicated that the whole Flo11 protein is needed to acquire the right production of adhesion nanodomains and support the model of Lipke et al., 2012 suggesting that upon the traction of an AFM tip or any other physical stretch, the adhesins are unfolded through their TRs, which expose amyloid-core sequences for interaction between neighbouring molecules.

The existence of the unique RR2 sequence in Flo11p of strain L69 led us to investigate the role of amyloid-forming sequences in the physiological function of this protein in relation to the role of other domains of this protein. For this purpose, we used CRISPR-Cas9 technology to create Flo11p variants specifically defective in these domains or sequences and whose expression is strictly dependent on the endogenous FLO11 promoter. This approach should allow us to perform comparative phenotypic analyses under conditions where the expression of these proteins is physiologically comparable. While our qRT-PCR fully confirmed this assertion, proteomic analysis carried out on cell wall/cell membrane preparation from L69 strain failed to detect Flo11p although other relevant cell wall proteins were identified. A likely explanation of this failure is the fact that Flo11p as well as other flocculins are not abundant cell wall proteins. Nevertheless, immunofluorescence experiments enabled to clearly identify Flo11p at the cell periphery and remarkably, same localization was found for all different Flo11p variants, leading to conclude that they are all in the cell wall. This result calls into question the fact that the modified GPI anchor at the C-terminus of Flo protein is essential for these flocculins to be retained at the cell wall (Bony et al., 1997; Kapteyn et al., 1999; Douglas et al., 2007). However, the finding of Flo11 protein in the extracellular medium when the GPI anchor was removed (Douglas et al., 2007) could be explained by the fact the corresponding gene was expressed from a high copy plasmid under the strong PGK1 promoter, which likely resulted in the production of a huge amount of protein that exceeds its ability to be entirely retained in the cell wall. Our immunofluorescence assays are also consistent with AFM data suggesting that Flo11 variants lacking C- or N-terminus are expressed at the cell surface, although firm demonstration of this cell surface localization should require further experiments using ligands, such as antibodies, recognizing specifically these truncated forms of Flo11p.

Cell-cell adhesion is a major cell wall property of Flo11p which is translated into either the sedimentation of cells as flocs at the bottom of the culture medium (flocculation phenotype), the formation of biofilm at the air-liquid interface (flor phenotype), or even biofilms on abiotic surface (adherence to plastic) (Verstrepen and Klis, 2006; Brückner and Mösch, 2012; Alexandre, 2013). Based on their work, Mösch and co-workers (Kraushaar et al., 2015; Brückner et al., 2020) proposed a Flo11p-dependent model of cell-cell adhesion that can furthermore allow for discrimination between cells from the same and from a different kind (kin discrimination). In a first model, it is proposed that Flo11 molecules cluster in cis- via amyloid β-sheet interactions sequences followed by interactions in trans between clusters of Flo11p of opposing cells. An alternative model was to suggest a simple bi-molecular Flo11p-Flo11p trans-interaction. Whatever the model, these authors showed an essential role of the A-domain, which encompasses 185–240 aa length at the N-terminal after the secretion signal of Flo11p, in cell-cell adhesion and indicated that this interaction is promoted at acidic pH and reduced at alkali pH. While our data did not reproduce very well those from Kraushaar et al., 2015 with respect to the importance of A-domain in the formation of cell aggregates and adherence to polystyrene, they however could support the models of cell-cell adhesion proposed by these authors. Indeed, the removal of amyloid-forming sequence in Flo11p of the L69 strain resulted in 70% drop in cellular aggregation, whereas loss of Flo11p abrogated this phenotype, indicating that the remaining aggregation can be due to simple bi-molecular interactions of isolated Flo11p between cells. Conversely, cell-cell interaction was reduced by 50% in a strain expressing Flo11p defective of the A-domain, suggesting that amyloid-forming sequences could in part overcome this defect and thus mediate interaction in trans. The finding that cell-cell aggregation was higher in BY4741 expressing a chimeric Flo11^BY-RR2^L69 protein than in the same strain expressing its own endogenous Flo11p also supports this suggestion, as well as the fact that huge aggregates seen in BY4741 strain overexpressing FLO11^L69 are cancelled by incubation with anti-amyloidophilic agents. Collectively, our results could be put in perspective with the recent work of Lipke and colleagues (Ho et al., 2019; Dehullu et al., 2019a) which used an innovative fluidic force microscopy enabled to manipulate and measured force at the single cells level to show that homotypic interactions between opposing cells can be mediated by amyloid-forming sequences. Nevertheless, the implication of amyloid-forming sequence is likely not an absolute requirement for cell-cell aggregation as witnessed by the finding that strain Σ1278b, which is known to only express Flo11p, forms huge aggregates that are resilient to the treatment with anti-amyloidophilic agents. These data could therefore agree with a model of Flo11p trans-interaction only mediated by A-domain.

Another dominant phenotype brought about by the expression of FLO11 is the ability of haploid yeast cells to undergo invasive growth in agar or pseudohyphal growth for a diploid cell (Gancedo, 2001). In this work, we found that strain L69 exhibited a very intense invasive growth phenotype, which was at first glance surprising since this strain is diploid. This phenotype is commonly not expressed in diploid cells unless FLO11 is highly expressed (Lo and Dranginis, 1998). However, it is unlikely that this pronounced agar invasion of L69 strain was solely due to high expression of FLO11 since our proteomic analysis failed to detect Flo11p. In addition, it should be noticed that the so-called high expression of FLO11 in L69 strain was made in reference with BY4742 strain that does not express this gene (Schiavone et al., 2015). Therefore, other yet unclear mechanisms must exist in this strain to account for this peculiar phenotype. Beside this open question, we showed that the N- and C-terminus of Flo11p are indispensable to elicit agar invasion phenotype, whereas amyloid-forming sequences have no role in this process. Our data are consistent with those of Kraushaar et al., 2015 who previously showed that the deletion of the A (N-terminal) domain of Flo11p abrogates agar adhesion, and also highlight a role for the C-terminus in this phenotype. This latter result could be explained by the inability of the cell to remain trapped in agar due to the fact that Flo11p, in the absence of its C-terminus, is still localized at the cell wall as indicated by our immunofluorescence assay, but loosely retained in this external structure. The invasive growth in agar was also strongly diminished upon deletion of repeat sequences in the B-domain, which can be explained by a reduction of glycosylation of Flo11p as a defect in this process was found to reduce invasive growth in agar (Meem and Cullen, 2012). Alternatively, the removal of TR in B-domain (i.e. RR1 sequence) may prevent Flo11p from reaching the cell surface and therefore reduce the capacity of the cells to remain trapped in the agar. Finally, this study brought to light new data on the process of invasive growth. On the one hand, we found that invasion in agar of strain L69 was equally intense under any kind of growth media, whereas this phenotype in Σ1278b background strain was weak in a galactose agar medium and absent in a glucose agar synthetic medium. In addition, the potent agar invasion phenotype of strain L69 could not be solely attributed to FLO11 gene because the overexpression of this gene in YSTW3a flo11Δ (derivative of Σ1278b) did not provide to this strain a higher capacity to invade agar than that of the isogenic YSTW3α expressing its endogenous gene. More puzzling was to find that the ectopic overexpression of FLO11 in BY4741 did not rescue the inability of this haploid strain to elicit invasive growth phenotype. Altogether, these data indicated that the Flo11p-dependent invasive phenotype involves additional genetic factors that are absent in BY4741 and that are more expressed in L69 strain, accounting for the higher potency of this strain to invade agar.

Based on this work and on several previous works on Flo11p in the yeast S. cerevisiae, one can raise the question why strains that naturally express Flo11p or in which this gene has been ectopically expressed do not exhibit collectively all these phenotypes of flocculation, flor, or growth invasion. Clearly, a large part of the answer lies in the sequence of this protein, and more precisely in differences found in the three domains of this protein, as already suggested by Barua et al., 2016 for the phenotypic comparison between Flo11p from S. cerevisiae var diastaticus and Σ1278b. As an example, the pH dependence and adherence to polystyrene is totally different between L69 and Σ1278b strain. This difference could be explained by the fact that Flo11p in the latter strain has a 15 amino acid insertion in the N-terminal at position 115–130 that was shown to strongly enhance the adhesive property of Flo11p (Brückner and Mösch, 2012). Likewise, the ability of Flo11p to elicit a flor phenotype in yeast has been attributed to high copy number of TRs in the B-domain, which enhances surface hydrophobicity via higher glycosylation (Fidalgo et al., 2006). However, the strain L69 does not exhibit flor phenotype although the expansion of the B-domain in Flo11p of this strain is very similar to that of strain 133d (46 TRs vs. 49 TRs). The lack of flor phenotype of L69 strain could be due either to a lower glycosylation of its Flo11p or to the type of TRs sequences, which in the case of Flo11p of this strain is represented by four different sequences, while only one single sequence of 81 nt is repeated 49 times in the Flo11p of strain 133d. In summary, the Flo11p sequence has been subjected to evolution to adapt yeast to its specific ecological niche and, therefore, the acquisition of additional amyloid-forming sequences in the Flo11p of the L69 strain could be due either to its original niche or to the particular conditions related to the industrial process for its propagation.

The raw dataset has been deposited to Dryad and is accessible at https://doi.org/10/10.5061/dryad.v41ns1rvv. The sequence of the FLO11 gene from the industrial strain used in this study has been deposited at NCBI Gene under the accession number 854836.

1. 1. François JM

   (2021) Dryad Digital Repository

   Data from: Physiological function of Flo11p domains and the particular role of amyloid core sequences of this adhesin in Saccharomyces cerevisiae.

   https://doi.org/10.5061/dryad.v41ns1rvv
2. 1. Bouyx C
   2. Schiavone M
   3. Teste MA
   4. Dague E
   5. Sieczkowski N
   6. Julien A
   7. François JM

   (2021) NCBI Gene

   ID 854836. Saccharomyces cerevisiae FLO11 (FLO11) gene, complete cds.

   https://www.ncbi.nlm.nih.gov/gene/854836

1. 1. Alexandre H

   (2013) Flor yeasts of Saccharomyces cerevisiae--their ecology, genetics and metabolism

   International Journal of Food Microbiology 167:269–275.

   https://doi.org/10.1016/j.ijfoodmicro.2013.08.021

   • Google Scholar
2. 1. Alsteens D
   2. Garcia MC
   3. Lipke PN
   4. Dufrene YF

   (2010) Force-induced formation and propagation of adhesion nanodomains in living fungal cells

   PNAS 107:20744–20749.

   https://doi.org/10.1073/pnas.1013893107

   • Google Scholar
3. 1. Arnoldi M
   2. Frirtz M
   3. Bauerlein E
   4. Radmacher M
   5. Sackmann E
   6. Boulbitch A

   (2000)

   Bacterial turgor pressure can be measured by atomic force microscopy

   The Physical Review E62:1034–1044.

   • Google Scholar
4. 1. Barua S
   2. Li L
   3. Lipke PN
   4. Dranginis AM

   (2016) Molecular basis for strain variation in the Saccharomyces cerevisiae adhesin FLO11P

   MSphere 1:00129-16.

   https://doi.org/10.1128/mSphere.00129-16

   • Google Scholar
5. 1. Bony M
   2. Thines-Sempoux D
   3. Barre P
   4. Blondin B

   (1997) Localization and cell surface anchoring of the Saccharomyces cerevisiae flocculation protein Flo1p

   Journal of Bacteriology 179:4929–4936.

   https://doi.org/10.1128/jb.179.15.4929-4936.1997

   • PubMed
   • Google Scholar
6. 1. Brückner S
   2. Mösch H-U

   (2012) Choosing the right lifestyle: Adhesion and development in Saccharomyces cerevisiae

   FEMS Microbiology Reviews 36:25–58.

   https://doi.org/10.1111/j.1574-6976.2011.00275.x

   • PubMed
   • Google Scholar
7. 1. Brückner S
   2. Schubert R
   3. Kraushaar T
   4. Hartmann R
   5. Hoffmann D
   6. Jelli E
   7. Drescher K
   8. Müller DJ
   9. Oliver Essen L
   10. Mösch H-U

   (2020) Kin discrimination in social yeast is mediated by cell surface receptors of the flo11 adhesin family

   eLife 9:e55587.

   https://doi.org/10.7554/eLife.55587

   • PubMed
   • Google Scholar
8. 1. Chan CXJ
   2. Lipke PN

   (2014) Role of force-sensitive amyloid-like interactions in fungal catch bonding and biofilms

   Eukaryotic Cell 13:1136–1142.

   https://doi.org/10.1128/EC.00068-14

   • PubMed
   • Google Scholar
9. 1. Chan CXJ
   2. El-Kirat-Chatel S
   3. Joseph IG
   4. Jackson DN
   5. Ramsook CB
   6. Dufrêne YF
   7. Lipke PN

   (2016) Force sensitivity in Saccharomyces cerevisiae flocculins

   MSphere 1:e00128-16.

   https://doi.org/10.1128/mSphere.00128-16

   • PubMed
   • Google Scholar
10. 1. Chopinet L
    2. Formosa C
    3. Rols MP
    4. Duval RE
    5. Dague E

    (2013) Imaging living cells surface and quantifying its properties at high resolution using AFM in QI mode

    Micron 48:26–33.

    https://doi.org/10.1016/j.micron.2013.02.003

    • PubMed
    • Google Scholar
11. 1. Dague E
    2. Alsteens D
    3. Latge JP
    4. Verbelen C
    5. Raze D
    6. Baulard AR

    (2007) Chemical force microscopy of single live cells

    Nano Letters 7:3026–3030.

    https://doi.org/10.1021/nl071476k

    • PubMed
    • Google Scholar
12. 1. Dehullu J
    2. Valotteau C
    3. Herman-Bausier P
    4. Garcia-Sherman M
    5. Mittelviefhaus M
    6. Vorholt JA
    7. Lipke PN
    8. Dufrêne YF

    (2019a) Fluidic force microscopy demonstrates that homophilic adhesion by Candida albicans als proteins is mediated by amyloid bonds between cells

    Nano Letters 19:3846–3853.

    https://doi.org/10.1021/acs.nanolett.9b01010

    • PubMed
    • Google Scholar
13. 1. Dehullu J
    2. Vorholt JA
    3. Lipke PN
    4. Dufrêne YF

    (2019b) Fluidic force microscopy captures amyloid bonds between microbial cells

    Trends in Microbiology 27:728–730.

    https://doi.org/10.1016/j.tim.2019.06.001

    • PubMed
    • Google Scholar
14. 1. Douglas LM
    2. Li L
    3. Yang Y
    4. Dranginis AM

    (2007) Expression and characterization of the flocculin Flo11/Muc1, a Saccharomyces cerevisiae mannoprotein with homotypic properties of adhesion

    Eukaryot Cell 6:2214–2221.

    https://doi.org/10.1128/EC.00284-06

    • PubMed
    • Google Scholar
15. 1. Dowell RD
    2. Ryan O
    3. Jansen A
    4. Cheung D
    5. Agarwala S
    6. Danford T

    (2010) Genotype to phenotype: a complex problem

    Science 328:469.

    https://doi.org/10.1126/science.1189015

    • PubMed
    • Google Scholar
16. 1. Dranginis AM
    2. Rauceo JM
    3. Coronado JE
    4. Lipke PN

    (2007) A biochemical guide to yeast adhesins: glycoproteins for social and antisocial occasions

    Microbiology and Molecular Biology Reviews 71:282–294.

    https://doi.org/10.1128/MMBR.00037-06

    • PubMed
    • Google Scholar
17. Software

    1. Durbin R
    2. Rice P
    3. Longden I
    4. Bleasby A

    (2000) EMBOSS ETANDEM

    European Bioinformatics Institute.

    https://emboss.bioinformatics.nl/cgi-bin/emboss/etandem
18. 1. Fernandez-Escamilla AM
    2. Rousseau F
    3. Schymkowitz J
    4. Serrano L

    (2004a) Prediction of sequence-dependent and mutational effects on the aggregation of peptides and proteins

    Nature Biotechnology 22:1302–1306.

    https://doi.org/10.1038/nbt1012

    • PubMed
    • Google Scholar
19. Software

    1. Fernandez-Escamilla AM
    2. Rousseau F
    3. Schymkowitz J
    4. Serrano L

    (2004b) TANGO

    Centre for Genomic Regulation.

    http://tango.crg.es/
20. 1. Fidalgo M
    2. Barrales RR
    3. Ibeas JI
    4. Jimenez J

    (2006) Adaptive evolution by mutations in the flo11 gene

    PNAS 103:11228–11233.

    https://doi.org/10.1073/pnas.0601713103

    • Google Scholar
21. 1. Formosa C
    2. Pillet F
    3. Schiavone M
    4. Duval RE
    5. Ressier L
    6. Dague E

    (2015a) Generation of living cell arrays for atomic force microscopy studies

    Nature Protocols 10:199–204.

    https://doi.org/10.1038/nprot.2015.004

    • Google Scholar
22. 1. Formosa C
    2. Schiavone M
    3. Boisrame A
    4. Richard ML
    5. Duval RE
    6. Dague E

    (2015b) Multiparametric imaging of adhesive nanodomains at the surface of Candida albicans by atomic force microscopy

    Nanomedicine 11:57–65.

    https://doi.org/10.1016/j.nano.2014.07.008

    • Google Scholar
23. 1. Free SJ

    (2013) Fungal cell wall organization and biosynthesis

    Advances in Genetics 81:33–82.

    https://doi.org/10.1016/B978-0-12-407677-8.00002-6

    • PubMed
    • Google Scholar
24. 1. Frieman MB
    2. Cormack BP

    (2003) The omega-site sequence of glycosylphosphatidylinositol-anchored proteins in Saccharomyces cerevisiae can determine distribution between the membrane and the cell wall

    Molecular Microbiology 50:883–896.

    https://doi.org/10.1046/j.1365-2958.2003.03722.x

    • PubMed
    • Google Scholar
25. 1. Gancedo JM

    (2001) Control of pseudohyphae formation in Saccharomyces cerevisiae

    FEMS Microbiology Reviews 25:107–123.

    https://doi.org/10.1111/j.1574-6976.2001.tb00573.x

    • PubMed
    • Google Scholar
26. 1. Gietz RD
    2. Schiestl RH

    (2007) High-efficiency yeast transformation using the LiAc/SS carrier DNA/PEG method

    Nature Protocols 2:31–34.

    https://doi.org/10.1038/nprot.2007.13

    • PubMed
    • Google Scholar
27. 1. Goffeau A

    (1998) The yeast genome

    Pathologie-Biologie 46:96–97.

    https://doi.org/10.1016/0168-9525(96)10027-5

    • PubMed
    • Google Scholar
28. 1. Guo B
    2. Styles CA
    3. Feng Q
    4. Fink GR

    (2000) A Saccharomyces gene family involved in invasive growth, cell-cell adhesion, and mating

    PNAS 97:12158–12163.

    https://doi.org/10.1073/pnas.220420397

    • PubMed
    • Google Scholar
29. 1. Halme A
    2. Bumgarner S
    3. Styles C
    4. Fink GR

    (2004) Genetic and epigenetic regulation of the FLO gene family generates cell-surface variation in yeast

    Cell 116:405–415.

    https://doi.org/10.1016/s0092-8674(04)00118-7

    • PubMed
    • Google Scholar
30. 1. Ho V
    2. Herman-Bausier P
    3. Shaw C
    4. Conrad KA
    5. Garcia-Sherman MC
    6. Draghi J
    7. Dufrene YF
    8. Lipke PN
    9. Rauceo JM

    (2019) An amyloid core sequence in the major Candida albicans adhesin ALS1P mediates cell-cell adhesion

    MBio 10:e01766-19.

    https://doi.org/10.1128/mBio.01766-19

    • PubMed
    • Google Scholar
31. 1. Jl H
    2. Bechhoefer J

    (1993) Calibration of atomic force microscope tips

    Rev Scientific Instruments 64:1868–1873.

    https://doi.org/10.1063/1.1143970

    • Google Scholar
32. 1. Kapteyn JC
    2. Van den Ende H
    3. Klis FM

    (1999) The contribution of cell wall proteins to the organization of the yeast cell wall

    Biochem Biophys Acta 1426:373–383.

    https://doi.org/10.1016/s0304-4165(98)00137-8

    • PubMed
    • Google Scholar
33. 1. Klis FM
    2. Boorsma A
    3. de Groot PW

    (2006) Cell wall construction in Saccharomyces cerevisiae

    Yeast 23:185–202.

    https://doi.org/10.1002/yea.1349

    • PubMed
    • Google Scholar
34. 1. Kobayashi O
    2. Yoshimoto H
    3. Sone H

    (1999) Analysis of the genes activated by the FLO8 gene in Saccharomyces cerevisiae

    Current Genetics 36:256–261.

    https://doi.org/10.1007/s002940050498

    • PubMed
    • Google Scholar
35. 1. Kraushaar T
    2. Brückner S
    3. Veelders M
    4. Rhinow D
    5. Schreiner F
    6. Birke R
    7. Pagenstecher A
    8. Mösch HU
    9. Essen LO

    (2015) Interactions by the fungal flo11 adhesin depend on a fibronectin type iii-like adhesin domain girdled by aromatic bands

    Structure 23:1005–1017.

    https://doi.org/10.1016/j.str.2015.03.021

    • PubMed
    • Google Scholar
36. 1. Lemesle-Varloot L
    2. Henrissat B
    3. Gaboriaud C
    4. Bissery V
    5. Morgat A
    6. Mornon JP

    (1990) Hydrophobic cluster analysis: procedures to derive structural and functional information from 2-D-representation of protein sequences

    Biochimie 72:555–574.

    https://doi.org/10.1016/0300-9084(90)90120-6

    • PubMed
    • Google Scholar
37. 1. Levin DE

    (2005) Cell wall integrity signaling in Saccharomyces cerevisiae

    Microbiology and Molecular Biology Reviews 69:262–291.

    https://doi.org/10.1128/MMBR.69.2.262-291.2005

    • PubMed
    • Google Scholar
38. 1. Lipke PN
    2. Garcia MC
    3. Alsteens D
    4. Ramsook CB
    5. Klotz SA
    6. Dufrene YF

    (2012) Strengthening relationships: amyloids create adhesion nanodomains in yeasts

    Trends in Microbiology 20:59–65.

    https://doi.org/10.1016/j.tim.2011.10.002

    • PubMed
    • Google Scholar
39. 1. Lipke PN
    2. Klotz SA
    3. Dufrene YF
    4. Jackson DN
    5. Garcia-Sherman MC

    (2018) Amyloid-Like beta-aggregates as force-sensitive switches in fungal biofilms and infections

    Microbiology and Molecular Biology Reviews 82:00035-17.

    https://doi.org/10.1128/MMBR.00035-17

    • PubMed
    • Google Scholar
40. 1. Liu H
    2. Styles CA
    3. Fink GR

    (1996)

    Saccharomyces cerevisiae S288C has a mutation in FLO8, a gene required for filamentous growth

    Genetics 144:967–978.

    • PubMed
    • Google Scholar
41. 1. Livak KJ
    2. Schmittgen TD

    (2001) Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) Method

    Methods 25:402–408.

    https://doi.org/10.1006/meth.2001.1262

    • PubMed
    • Google Scholar
42. 1. Lo WS
    2. Dranginis AM

    (1996) Flo11, a yeast gene related to the STA genes, encodes a novel cell surface flocculin

    Journal of Bacteriology 178:7144–7151.

    https://doi.org/10.1128/jb.178.24.7144-7151.1996

    • PubMed
    • Google Scholar
43. 1. Lo WS
    2. Dranginis AM

    (1998) The cell surface flocculin Flo11 is required for pseudohyphae formation and invasion by Saccharomyces cerevisiae

    Molecular Biology of the Cell 9:161–171.

    https://doi.org/10.1091/mbc.9.1.161

    • PubMed
    • Google Scholar
44. 1. Lu CF
    2. Kurjan J
    3. Lipke PN

    (1994) A pathway for cell wall anchorage of Saccharomyces cerevisiae alpha-agglutinin

    Molecular and Cellular Biology 14:4825–4833.

    https://doi.org/10.1128/mcb.14.7.4825-4833.1994

    • PubMed
    • Google Scholar
45. 1. Lu CF
    2. Montijn RC
    3. Brown JL
    4. Klis F
    5. Kurjan J
    6. Bussey H
    7. Lipke PN

    (1995) Glycosyl phosphatidylinositol-dependent cross-linking of alpha-agglutinin and beta 1,6-glucan in the Saccharomyces cerevisiae cell wall

    The Journal of Cell Biology 128:333–340.

    https://doi.org/10.1083/jcb.128.3.333

    • PubMed
    • Google Scholar
46. 1. Mans R
    2. van Rossum HM
    3. Wijsman M
    4. Backx A
    5. Kuijpers NG
    6. van den Broek M

    (2015) CRISPR/Cas9: a molecular Swiss army knife for simultaneous introduction of multiple genetic modifications in Saccharomyces cerevisiae

    FEMS Yeast Research 15:fov004.

    https://doi.org/10.1093/femsyr/fov004

    • Google Scholar
47. 1. Meem MH
    2. Cullen PJ

    (2012) The impact of protein glycosylation on Flo11-dependent adherence in Saccharomyces cerevisiae

    FEMS Yeast Research 12:809–818.

    https://doi.org/10.1111/j.1567-1364.2012.00832.x

    • Google Scholar
48. 1. Nielsen H

    (2017a) Predicting Secretory Proteins with SignalP

    Methods in Molecular Biology 1611:59–73.

    https://doi.org/10.1007/978-1-4939-7015-5_6

    • PubMed
    • Google Scholar
49. Software

    1. Nielsen H

    (2017b) Signalp 4.1 server

    DTU Health Tech.

    http://www.cbs.dtu.dk/services/SignalP-4.1/
50. 1. Pierleoni A
    2. Martelli PL
    3. Casadio R

    (2008a) PredGPI: a GPI-anchor predictor

    BMC Bioinformatics 9:392.

    https://doi.org/10.1186/1471-2105-9-392

    • PubMed
    • Google Scholar
51. Software

    1. Pierleoni A
    2. Martelli PL
    3. Casadio R

    (2008b) PREDGPI

    Bologna Biocomputing Group.

    http://gpcr.biocomp.unibo.it/predgpi/
52. 1. Purevdorj-Gage B
    2. Sheehan KB
    3. Hyman LE

    (2006) Effects of low-shear modeled microgravity on cell function, gene expression, and phenotype in Saccharomyces cerevisiae

    Applied and Environmental Microbiology 72:4569–4575.

    https://doi.org/10.1128/AEM.03050-05

    • Google Scholar
53. 1. Purevdorj-Gage B
    2. Orr ME
    3. Stoodley P
    4. Sheehan KB
    5. Hyman LE

    (2007) The role of FLO11 in Saccharomyces cerevisiae biofilm development in a laboratory based flow-cell system

    FEMS Yeast Research 7:372–379.

    https://doi.org/10.1111/j.1567-1364.2006.00189.x

    • Google Scholar
54. 1. Ramsook CB
    2. Tan C
    3. Garcia MC
    4. Fung R
    5. Soybelman G
    6. Henry R

    (2010) Yeast cell adhesion molecules have functional amyloid-forming sequences

    Eukaryot Cell 9:393–404.

    https://doi.org/10.1128/EC.00068-09

    • Google Scholar
55. 1. Reynolds TB
    2. Fink GR

    (2001) Bakers’ yeast, a model for fungal biofilm formation

    Science 291:878–881.

    https://doi.org/10.1126/science.291.5505.878

    • PubMed
    • Google Scholar
56. 1. Rice P
    2. Longden I
    3. Bleasby A

    (2000) EMBOSS: the European Molecular Biology Open Software Suite

    Trends in Genetics 16:276–277.

    https://doi.org/10.1016/s0168-9525(00)02024-2

    • PubMed
    • Google Scholar
57. 1. Roberts RL
    2. Fink GR

    (1994) Elements of a single MAP kinase cascade in Saccharomyces cerevisiae mediate two developmental programs in the same cell type: mating and invasive growth

    Genes & Development 8:2974–2985.

    https://doi.org/10.1101/gad.8.24.2974

    • PubMed
    • Google Scholar
58. 1. Ryan OW
    2. Poddar S
    3. Cate JH

    (2016) CRISPR-Cas9 genome engineering in Saccharomyces cerevisiae cells

    Cold Spring Harbor Protocols 2016:86827.

    https://doi.org/10.1101/pdb.prot086827

    • PubMed
    • Google Scholar
59. 1. Sarode N
    2. Miracle B
    3. Peng X
    4. Ryan O
    5. Reynolds TB

    (2011) Vacuolar protein sorting genes regulate mat formation in Saccharomyces cerevisiae by Flo11p-dependent and -independent mechanisms

    Eukaryot Cell 10:1516–1526.

    https://doi.org/10.1128/EC.05078-11

    • PubMed
    • Google Scholar
60. 1. Schiavone M
    2. Sieczkowski N
    3. Castex M
    4. Dague E
    5. Marie Francois J

    (2015) Effects of the strain background and autolysis process on the composition and biophysical properties of the cell wall from two different industrial yeasts

    FEMS Yeast Research 15:012.

    https://doi.org/10.1093/femsyr/fou012

    • PubMed
    • Google Scholar
61. 1. Sievers F
    2. Higgins DG

    (2014) Clustal omega

    Current Protocols in Bioinformatics 48:13.

    https://doi.org/10.1002/0471250953.bi0313s48

    • PubMed
    • Google Scholar
62. 1. Solis-Escalante D
    2. Kuijpers NG
    3. Bongaerts N
    4. Bolat I
    5. Bosman L
    6. Pronk JT

    (2013) amdSYM, a new dominant recyclable marker cassette for Saccharomyces cerevisiae

    FEMS Yeast Research 13:126–139.

    https://doi.org/10.1111/1567-1364.12024

    • PubMed
    • Google Scholar
63. 1. Strahl-Bolsinger S
    2. Gentzsch M
    3. Tanner W

    (1999) Protein O-mannosylation

    Biochimica et Biophysica Acta 1426:297–307.

    https://doi.org/10.1016/s0304-4165(98)00131-7

    • PubMed
    • Google Scholar
64. 1. Teste MA
    2. Duquenne M
    3. Francois JM
    4. Parrou JL

    (2009) Validation of reference genes for quantitative expression analysis by real-time RT-PCR in Saccharomyces cerevisiae

    BMC Molecular Biology 10:99.

    https://doi.org/10.1186/1471-2199-10-99

    • PubMed
    • Google Scholar
65. 1. Van Mulders SE
    2. Christianen E
    3. Saerens SM
    4. Daenen L
    5. Verbelen PJ
    6. Willaert R

    (2009) Phenotypic diversity of Flo protein family-mediated adhesion in Saccharomyces cerevisiae

    FEMS Yeast Research 9:178–190.

    https://doi.org/10.1111/j.1567-1364.2008.00462.x

    • PubMed
    • Google Scholar
66. 1. Veelders M
    2. Bruckner S
    3. Ott D
    4. Unverzagt C
    5. Mosch HU
    6. Essen LO

    (2010) Structural basis of flocculin-mediated social behavior in yeast

    PNAS 107:22511–22516.

    https://doi.org/10.1073/pnas.1013210108

    • PubMed
    • Google Scholar
67. 1. Verstrepen KJ
    2. Jansen A
    3. Lewitter F
    4. Fink GR

    (2005) Intragenic tandem repeats generate functional variability

    Nature Genetics 37:986–990.

    https://doi.org/10.1038/ng1618

    • PubMed
    • Google Scholar
68. 1. Verstrepen KJ
    2. Klis FM

    (2006) Flocculation, adhesion and biofilm formation in yeasts

    Molecular Microbiology 60:5–15.

    https://doi.org/10.1111/j.1365-2958.2006.05072.x

    • PubMed
    • Google Scholar
69. 1. Voordeckers K
    2. Kominek J
    3. Das A
    4. Espinosa-Cantú A
    5. De Maeyer D
    6. Arslan A
    7. Van Pee M
    8. van der Zande E
    9. Meert W
    10. Yang Y
    11. Zhu B
    12. Marchal K
    13. DeLuna A
    14. Van Noort V
    15. Jelier R
    16. Verstrepen KJ

    (2015) Adaptation to high ethanol reveals complex evolutionary pathways

    PLOS Genetics 11:e1005635.

    https://doi.org/10.1371/journal.pgen.1005635

    • PubMed
    • Google Scholar
70. 1. Zara S
    2. Bakalinsky AT
    3. Zara G
    4. Pirino G
    5. Demontis MA
    6. Budroni M

    (2005) Flo11-based model for air-liquid interfacial biofilm formation by Saccharomyces cerevisiae

    Applied and Environmental Microbiology 71:2934–2939.

    https://doi.org/10.1128/AEM.71.6.2934-2939.2005

    • PubMed
    • Google Scholar

Acceptance summary:

Your study confirms and expands our understanding of fungal adhesins in general, and the role of amyloid-forming regions in particular. Specifically, the data show that amyloid sequences are needed for cis-interactions leading to clusters of Flo11p at the cell surface.

Decision letter after peer review:

Thank you for submitting your article "Physiological function of Flo11p domains and the particular role of amyloid core sequences of this adhesin in Saccharomyces cerevisiae" for consideration by eLife. Your article has been reviewed by 3 peer reviewers, and the evaluation has been overseen by a Reviewing Editor and Vivek Malhotra as the Senior Editor. The following individuals involved in review of your submission have agreed to reveal their identity: Peter Lipke (Reviewer #1).

The reviewers have discussed their reviews with one another, and the Reviewing Editor has drafted this to help you prepare a revised submission.

While all reviewers agree that your work is an interesting confirmation of previous findings, they also agreed that the study suffers from several loose ends that need to be resolved before the paper can be considered for publication in eLife.

Firstly, there is no direct evidence that the different adhesin constructs are present in more or less equal amounts at the cell surface. Expression levels may not be a good proxy for the levels of active protein at the cell surface, especially for constructs lacking the GPI domain medium and decreased surface anchorage. Hence, further experiments to address this issue, for example using previously established methods to observe adhesins, such as atomic force microscopy and immunofluorescence microscopy, seem essential.

Secondly, the formation of amyloid-like structures was not demonstrated directly. Demonstrating that the relevant peptides can form amyloids that can be broken by ThioS or the breaker peptide, would take care of that issue.

Third, it does not seem appropriate to only show data from one cell instead of reporting the aggregated data from all cells.

A last important issue is that the structure of the abstract and in fact the whole manuscript is perhaps not optimal. The authors convey and mix too many different messages, and the summary and main text would benefit from simplification and a better separation of the different questions that are addressed. In addition, the authors are also encouraged to better discuss their findings against the background of previous literature, explicitly pointing out where their findings confirm previous observations.

Reviewer #1 (Recommendations for the authors):

The manuscript would benefit from discussion of relevant data and models in 3 recent papers:

Ho et al., 2019. An Amyloid Core Sequence in the Major Candida albicans Adhesin Als1p Mediates Cell-Cell Adhesion. mBio. 2019 Oct 8;10(5):e01766-19. doi: 10.1128/mBio.01766-19.

Dehullu et al., 2019. Fluidic Force Microscopy Captures Amyloid Bonds between Microbial Cells. Trends Microbiol. 2019 Sep;27(9):728-730. doi: 10.1016/j.tim.2019.06.001.

Dehullu et al., 2019. Fluidic Force Microscopy Demonstrates That Homophilic Adhesion by Candida albicans Als Proteins Is Mediated by Amyloid Bonds between Cells. Nano Lett. 2019 Jun 12;19(6):3846-3853. doi: 10.1021/acs.nanolett.9b01010.

Specific comments.

l.40: The wall integrity statement is not supported by the data: the data shows that expression of Flo11 contributes to resistance to wall-perturbing drugs. The wording would be modified accordingly

l. 48: "reported that an adhesin…"

l. 48, 58: also shown for Als1p: Ho et al., 2019

l 61 and elsewhere: The authors may also find that Dehullu et al., 2019a, b are useful in data analysis and discussion. These papers describe and summarize formation of amyloid-like bonds between cells to mediate cellular aggregation.

l. 76: It would be helpful to mention that Flo11 is not homologous to Flo1-family adhesins. There is no apparent homology of Flo11 in any region of the Flo1-family proteins.

l. 106-112: the bases for these statements are not convincing

l. 138-141: sentence not clear

l. 159 delete "a" in two places, because the subject is plural

l 157-161: Figure 3A shows ~75% inhibition of the low-adhesion domains as well. This inhibition contradicts the statement.

l. 180: suggest "facilitated" rather than "warranted"

l. 184-187: the sentence contains a double negative "…the lack of nanodomains formation in BY4741 is not solely due…" and would be clearer if rephrased.

l. 206: reference should be Dranginis 2007.

l. 223-225: If the VASTTVT mutant was not made and tested, I suggest alternate wording "…Finally, it is interesting to notice that mutating the VVSTTV to VASTTVT (the sequence used in Figure 2a as anti-amyloid peptide) dropped the TANGO-predicted β-aggregation potential of the in this region of the Flo11 sequence to 26% (Table S3)." Residue position numbers would help.

l. 253: give quantitative estimates or % reduction in stiffness.

l. 254, 321, etc.: All discussion of effects of the δ C-Ter form are suspect because we do not know whether any Flo11 is present on the cell surface, nor how it is attached if present, nor its cell-surface orientation. Labeling and detection of surface Flo11 levels would be required.

l. 262. The adhesion peaks in Figure S10 look like 300-500 pN, rather than the 100 pN stated.

l. 322. The loss of aggregation might also be caused by differences in cell surface concentration.

l. 331: typo: should be "not"

l. 348: differences might also be caused by differences in cell surface concentration due to different culture methods.

l. 418 see comment on l. 40

l. 460-470: This discussion is not clear. The sentence on l. 462-3 appears to be self-contradictory: how can L69 not exhibit a flocculation phenotype? The concluding sentence of the paragraph has a double negative, and its meaning is not clear.

l. 848 has typos: "Figure 6: The Flo11p cell-cell aggregation is potentiated by increased number of amyloid-forming sequences."

Figure 8 overlapping label on 2nd column

Figure S10 correct to '…a number obtained from a single cell of…"

Reviewer #2 (Recommendations for the authors):

I find the abstract a bit unfocussed and trying to convey too many messages. Please simplify it.

Reviewer #3 (Recommendations for the authors):

In the introduction, include in the discussion the state-of-the art on Flo11p, such the structure Flo11 proteins that have been recently solved as well as functional studies on Flo11p-based cell-cell interaction (including AFM-based single-cell interaction studies),

see

– Brückner S, Schubert R, Kraushaar T, Hartmann R, Hoffmann D, Jelli E, Drescher K, Müller DJ, Oliver Essen L, Mösch HU. Kin discrimination in social yeast is mediated by cell surface receptors of the Flo11 adhesin family. eLife. 2020 Apr 14;9:e55587;

– Kraushaar T, Brückner S, Veelders M, Rhinow D, Schreiner F, Birke R, Pagenstecher A, Mösch HU, Essen LO. Interactions by the Fungal Flo11 Adhesin Depend on a Fibronectin Type III-like Adhesin Domain Girdled by Aromatic Bands. Structure. 2015 Jun 2;23(6):1005-17.

Also include recent reviews on fungal adhesins such as

– Willaert RG. Adhesins of Yeasts: Protein Structure and Interactions. J Fungi (Basel). 2018 Oct 27;4(4):119.

– Essen LO, Vogt MS, Mösch HU. Diversity of GPI-anchored fungal adhesins. Biol Chem. 2020 Nov 26;401(12):1389-1405.

Other relevant (landmark) previous research on Flo11p should be included in the introduction and/or the discussion of the results:

– Lambrechts MG, Bauer FF, Marmur J, Pretorius IS. Muc1, a mucin-like protein that is regulated by Mss10, is critical for pseudohyphal differentiation in yeast. Proc Natl Acad Sci U S A. 1996 Aug 6;93(16):8419-24..

– Bayly JC, Douglas LM, Pretorius IS, Bauer FF, Dranginis AM. Characteristics of Flo11-dependent flocculation in Saccharomyces cerevisiae. FEMS Yeast Res. 2005 Dec;5(12):1151-6. doi: 10.1016/j.femsyr.2005.05.004.

– Douglas LM, Li L, Yang Y, Dranginis AM. Expression and characterization of the flocculin Flo11/Muc1, a Saccharomyces cerevisiae mannoprotein with homotypic properties of adhesion. Eukaryot Cell. 2007 Dec;6(12):2214-21. doi: 10.1128/EC.00284-06.

– Purevdorj-Gage B, Orr ME, Stoodley P, Sheehan KB, Hyman LE. The role of FLO11 in Saccharomyces cerevisiae biofilm development in a laboratory-based flow-cell system. FEMS Yeast Res. 2007 May;7(3):372-9.

– Van Mulders SE, Christianen E, Saerens SM, Daenen L, Verbelen PJ, Willaert R, Verstrepen KJ, Delvaux FR. Phenotypic diversity of Flo protein family-mediated adhesion in Saccharomyces cerevisiae. FEMS Yeast Res. 2009 Mar;9(2):178-90.

– Goossens KV, Willaert RG. The N-terminal domain of the Flo11 protein from Saccharomyces cerevisiae is an adhesin without mannose-binding activity. FEMS Yeast Res. 2012 Feb;12(1):78-87.

p. 4: In the description of the protein architecture the function of the A-domain as an adhesin is now missing and should be described in detail, since it is the A-domain that contains the Flo11 domain (see Pfam database: PF10182 or the interpro database). The Flo11p domain is now generally accepted as the domain of the Flo11 protein that is responsible for the cell-cell interaction.

Nanodomains were observed using AFM and a bare tip. Some details on how the measurements were performed are missing from the Materials and Methods section, such as which cantilever(s) of the MLCT were used, in which buffer (or growth medium) (what was the pH?) were the experiments performed. Were the cantilevers plasma or O3 cleaned?

A question that arises when looking at the adhesion maps and adhesion force determination is which type of interaction is present here? Is it an electrostatic or hydrophobic/hydrophilic interaction? Plasma/O3 cleaning can change the hydrophilicity of the cantilever and the adhesion strength. It would be interesting to better characterize the type of interaction since now there is no control. As an additional measurement, the cantilever could be coated with compounds that make the surface hydrophobic or positive or negative charged.

Additionally, since the type of interaction is not well characterized (as well as the conditions), it is not clear now with which part of the Flo11p (and/or the glycans (Meem 7 Cullen, 2012); since yeast glycans contain (charged) phosphates that can interact (Goossens KV, Ielasi FS, Nookaew I, Stals I, Alonso-Sarduy L, Daenen L, Van Mulders SE, Stassen C, van Eijsden RG, Siewers V, Delvaux FR, Kasas S, Nielsen J, Devreese B, Willaert RG. Molecular mechanism of flocculation self-recognition in yeast and its role in mating and survival. mBio. 2015 Apr 14;6(2):e00427-15). For example, Kraushaar et al., (Structure. 2015 Jun 2;23(6):1005-17) showed that the Flo11 domain trans-interact via a hydrophobic interaction due to the presence of 2 aromatic bands in the structure. They also discovered that this interaction only occurs at low pH (5.5) and not at higher pH (6.8)).

These aspects will influence the results on the measured adhesion forces. A bimodal distribution of the adhesion forces were observed. The weaker adhesion forces were attributed to surface proteins that unfold upon retraction of the AFM tip (p. 6, line 137; p. 7, line 160). Are these surface proteins the Flo11 proteins? There is no direct evidence from the results that these nanodomains are composed of Flo11p. This could be obtained by adhesion maps using an anti-Flo11p antibody or with the purified Flo11p, or Flo11p on other cells by performing single cell force-spectroscopy (Brückner et al., eLife. 2020 Apr 14;9:e55587). What was the effect of lowering the maximal applied force on the adhesion forces?

If the weaker adhesion forces were due to unfolding of proteins, this should be visible in the force-distance curves. If it was not visible, this hypothesis should be rejected. Please provide these graphs.

p. 11, line 254: "On the other hand, the cell surface of L69flo11-∆Cter cell was characterized by needle-shaped nanostructures whose height was 3 times greater than nanodomains formed on the surface of the L69 strain. These nanodomains displayed adhesion forces that were scattered from a few pN to max 200 pN and stiffness that was roughly 30% higher than that of nanodomains from L69 strain (see Figure S9, in supplementary data)."

– How could these results be explained? When the C-terminal part of the protein has been removed, is it still anchored in the cell wall? Is there proof of this? Suppose that the truncated Flo11p is present: Does the bare AFM tip interact specifically with the remainder of Flo11p and does cis-interaction results again in these nanodomains? Why are these nanodomains then larger?

The force-distance curves shown in Figure S.9 G, indicate 2 slopes indicating that first a softer layer is pushed through. Not clear what this could be?

p. 12, line 293: …"This residual clumping is likely a specific feature from Σ1278b background strain, which does not hamper the notion that Flo11p is critical in cell-cell interaction. This is further supported by the lack of aggregates in BY4741 as this strain does not express FLO11 due to non-sense mutation in FLO8 encoding its major transcriptional activator.

– Not clear what is meant with a "specific" feature? How could this remaining interaction be explained? Remark that Brückner et al., (2020) demonstrated quantitatively using single-cell force spectroscopy that the Flo11p A-domain is needed to have a significant cell-cell interaction (although some variation is present in these single-cell experiments).

– Were these aggregation experiments also performed at a lower pH and could the results of Kraushaar et al., (2015) be confirmed?

– It has been shown recently that the BY4741 strain can flocculate depending on the condition and this flocculation was independent of Flo5p (see Degreif D, de Rond T, Bertl A, Keasling JD, Budin I. Lipid engineering reveals regulatory roles for membrane fluidity in yeast flocculation and oxygen-limited growth. Metab Eng. 2017 May;41:46-56). This has been shown for Flo1p. In many conditions, Flo1p and Flo11p are be coexpressed. Remark that the expression of FLO11 is very complex and control has been attributed to many transcription factors (Brückner S, Mösch HU. Choosing the right lifestyle: adhesion and development in Saccharomyces cerevisiae. FEMS Microbiol Rev. 2012 Jan;36(1):25-58.).

Has it been assessed that the flocs that were observed could be deflocculated by adding EDTA? Was the expression of other Flo proteins also measured?

– How many aggregates were counted to obtain the results? It is not clear if this method can detect significant small changes in interaction forces.

– An additional quantitative method of measuring the interaction between cells is needed to confirm the obtained results based on observing aggregates.

p. 14, line 330: … "Adherence was assayed on polystyrene surfaces."

– Note that this was also investigated before: see Van Mulders et al., 2009, and should be included in the discussion (since here also the adhesion of cells that express other Flo proteins was investigated).

p. 14, line 335: "Moreover, the moderate adherence of BY4741 cell to plastic at pH 5.0 must be due to cell wall proteins other than Flo11p, since this gene is not expressed in this strain due to a non-sense mutation in FLO8 encoding its transcriptional activator (Liu et al., 1996).

– Same comment as before: Under the selected conditions, was it confirmed that no other Flo proteins were expressed?

p. 16 and discussion: Invasive growth experiments => discuss the results also in relation to previous published results on this (such as Van Mulders et al., 2009).

p. 19, line 454: "The finding that this phenotype was not abrogated when domain A was deleted could be at first glance in contradiction with the data of Kraushaar et al., (Kraushaar et al., 2015) who showed that this domain is essential for conferring cell-cell adhesion by facilitating homotypic interactions".

– Also in contrast to others: Bayly et al., (2005), Goossens and Willaert (2012).

However, the lack of A domain can be in part overcome by amyloid-forming sequences as these later could promote clustering of Flo11 molecules in cis, which enables cell-cell adhesion by trans-interaction as proposed by these same authors (Kraushaar et al., 2015, Bruckner et al., 2020).

– In the model hypothesis of Kraushaar et al., 2015; Bruckner et al., 2020 (based and their results as well on previous results from e.g. Ramsook et al.,), they propose that there is trans interaction between interacting cells, which is based on the interaction of the Flo11 domains, and that cis-interaction between various Flo11 proteins via the β-aggregation-prone amyloid-forming sequences could precede the trans-interaction. This model needs further experimental confirmation to be generally accepted.

It could be that the observed nanodomains correspond to trans-interacting Flo11 proteins. However, the cell-cell interaction would be significantly reduced if the Flo11p domains are not present in the truncated protein. This is however in contrast to what has been observed by the authors in this manuscript. Therefore, it is not clear how this can be explained.

While all reviewers agree that your work is an interesting confirmation of previous findings, they also agreed that the study suffers from several loose ends that need to be resolved before the paper can be considered for publication in eLife.

Firstly, there is no direct evidence that the different adhesin constructs are present in more or less equal amounts at the cell surface. Expression levels may not be a good proxy for the levels of active protein at the cell surface, especially for constructs lacking the GPI domain medium and decreased surface anchorage. Hence, further experiments to address this issue, for example using previously established methods to observe adhesins, such as atomic force microscopy and immunofluorescence microscopy, seem essential.

It is indeed clear that there is no direct correlation between expression level of a given gene and level of the encoded protein and so we agreed that qRT-PCR, although showing similar expression levels of the genes encoding the wild type and the variants of Flo11 should be completed by a more direct quantification of the protein. We solve this issue in two ways:

A – As suggested by reviewer #1 and Editor, we carried out immunofluorescence assays. To do this, Flo11 and its variants were flanked at their C-ter of a 6-His tag which was inserted into the gene by CRISPR-cas9. Then, the immunofluorescence assay was carried out using Alexa Fluor 488-conjugated Anti-6x-His Tag Monoclonal Antibody. The result of this experiment clearly shows strong and distinct fluorescence spots at the cell periphery, regardless of the Flo11 variant used. This result, which we present as Figure 5 in the revised paper, leads to two relevant pieces of information which are (i) loss of amyloid β-aggregation sequence does not abrogate this localisation and (ii) more importantly, loss of C-ter and thus of the GPI anchor does not apparently prevent Flo11 protein to be retained at the cell surface. We then wanted to quantify the level of Flo11 protein produced in strains using flow cytometer, but the signal level was close to the autofluorescence, likely because our constructs are expressed at a single copy from the endogenous FLO11 promoter.

B – The second approach was to quantify Flo11p by a shot gun proteomic analysis on samples enriched with cell wall and cell membranes according to a protocol taken from Sarode et al., (EC, 10, 1516-1526, 2011). We did this experiment twice and identified around 700 proteins, from cell wall, cell membrane and mitochondria (likely because of the process of centrifugation that was suggested in the protocol). In cell wall, we identified about 20 cell wall among which Gas1, Fks1, Fks2, cwp1, several cell mannoproteins such as Pir1, Pir4, Pst1, Cwp1, but we failed to detect Flo11p as well as other flocculins. When looking carefully to the median abundance of the cell wall proteins that were identified in this shot-gun proteomic analysis, using SGD database , it turns out that the detection limit for statistical confidence in protein identification roughly corresponded to 3000 molecules per cell (ie the case for EXG2 encoding an endo1,3-glucanase). Thus, it may be possible that level of Flo11p was below this level, which could also explain that we were unable to detect Flo11p by Western blot using the same strain that was used for immunofluoresecence and thus using the anti-HIS antibody to cross react with the 6xHis tag that is at the C-ter of the Flo11p. It should be noticed that detection of Flo11p reported in previous work was done using strains that were transformed with a 2-µ plasmid that carries FLO11 under a strong promoter such as PGK1 (see work of Douglas et al., 2007) or under GAL1. Thus, the data we have reported here actually suggest that Flo11p is not a very abundant protein and that if this protein is highly overproduced, this may result in the expression of phenotypic behaviours that no longer reflect physiological reality. For the interest of reviewers and readers, we have provided the proteomic data for strain L69 in the supplementary data (data of one typical proteomic analysis is reported Supplementary File 3, as the second provided very similar results).

Secondly, the formation of amyloid-like structures was not demonstrated directly. Demonstrating that the relevant peptides can form amyloids that can be broken by ThioS or the breaker peptide, would take care of that issue.

It was already shown that Flo11p assembled into amyloid fibers in vitro (see reference Ramsook et al., Eukaryotic Cell, 9: 393-404, and this reference is given in our paper). This was not the purpose of this work to rediscover that, but to show that the nanodomains at the surface of the L69 strain was due to the presence of amyloid-core sequences in the Flo11p. Our data support the Lipke’ model that these β-aggregation prone sequences are required for the production of these nanostructures in response to an external force such as the retraction of AFM tip from the surface, and that these nanostructures are abolished when these sequences are removed in the protein (Flo11-ΔRR2) or upon incubation with an anti-amyloid peptide or amyloid perturbing agents such as thioflavin S.

Third, it does not seem appropriate to only show data from one cell instead of reporting the aggregated data from all cells.

As indicated in MandM, all AFM experiments were repeated 3 times on each strain expressing a Flo11 variant (ie 3 biological replicates), with 4 to 8 cells per replicate. For each cell, 1024 force-distance curves have been recorded given rise to histograms of the distribution of adhesion forces or spring constants. In figure 1E, we reported such a representation for one representative cell. By aggregating all data, we found obviously a similar bimodal distribution of adhesion values, characterized however by a high variability. Thus, we suggested to present the aggregated values in the manuscript as box plots (see figure 1 Gand H) to illustrate both this bimodality of the distribution and the variability of the adhesion force and stiffness.

A last important issue is that the structure of the abstract and in fact the whole manuscript is perhaps not optimal. The authors convey and mix too many different messages, and the summary and main text would benefit from simplification and a better separation of the different questions that are addressed. In addition, the authors are also encouraged to better discuss their findings against the background of previous literature, explicitly pointing out where their findings confirm previous observations.

We have extensively revised our paper to focus on the importance and role of amyloid sequences in the function of Flo11 in yeast and in particular, our data show that amyloid sequences are needed for cis-interaction leading to clusters of Flo11p making nanodomain. Also they may contribute to trans-interaction since ablation of these sequences strongly reduce cell-cell interaction, or alternatively, insertion of the RR2 domain from Flo11 of L69 strain that is enriched in amyloid core sequences in the Flo11 of BY strain leads to strong cell-cell aggregation, what barely occurred in the BY strain expressing its normal Flo11p. We thus suggested this new title to "The dual role of the amyloid core sequences in the physiological function of FLO11-encoded flocculin in the yeast Saccharomyces cerevisiae".

Reviewer #1 (Recommendations for the authors):

The manuscript would benefit from discussion of relevant data and models in 3 recent papers:

Ho et al., 2019. An Amyloid Core Sequence in the Major Candida albicans Adhesin Als1p Mediates Cell-Cell Adhesion. mBio. 2019 Oct 8;10(5):e01766-19. doi: 10.1128/mBio.01766-19.

Dehullu et al., 2019. Fluidic Force Microscopy Captures Amyloid Bonds between Microbial Cells. Trends Microbiol. 2019 Sep;27(9):728-730. doi: 10.1016/j.tim.2019.06.001.

Dehullu et al., 2019. Fluidic Force Microscopy Demonstrates That Homophilic Adhesion by Candida albicans Als Proteins Is Mediated by Amyloid Bonds between Cells. Nano Lett. 2019 Jun 12;19(6):3846-3853. doi: 10.1021/acs.nanolett.9b01010.

We are very grateful to this reviewer for alerting us to these 3 papers. They are in fact extremely relevant for our paper because they consolidate our experimental data suggesting that the amyloid core sequences are needed for nanoclustering of Flo11 molecules by cis-interaction, but also can contribute to cell-cell adhesion by trans-interaction or, alternatively, that this homophilic adhesion which is dependent on the A-domain of Flo11p according to the work of Kraushaar et al., (Structure, 2015) may require amyloid-β sheet interaction sequence to be the most effective. This suggestion is based on our data that cell-cell aggregation are strongly diminished in a strain expressing a Flo11 variant lacking amyloid core sequences, whereas this phenotype is remarkably exacerbated in BY cells either upon expression of the FLO11 of L69 strain or insertion of additional amyloid-core sequences in its endogenous FLO11.

Specific comments.

l.40: The wall integrity statement is not supported by the data: the data shows that expression of Flo11 contributes to resistance to wall-perturbing drugs. The wording would be modified accordingly

We showed that loss of Flo11 results in a higher sensitivity to cell wall drugs, which are currently used to assess the cell wall integrity. We do not mean that Flo11/flocculin is needed for wall integrity but only suggested that it may contribute to it. We reformulated therefore this meaning in the revised version (see line 413-415).

l. 48: "reported that an adhesin…"

corrected

l. 48, 58: also shown for Als1p: Ho et al., 2019

corrected

l 61 and elsewhere: The authors may also find that Dehullu et al., 2019a, b are useful in data analysis and discussion. These papers describe and summarize formation of amyloid-like bonds between cells to mediate cellular aggregation.

Indeed, this is a good suggestion and we have included these excellent papers in our manuscript (see notably Introduction).

l. 76: It would be helpful to mention that Flo11 is not homologous to Flo1-family adhesins. There is no apparent homology of Flo11 in any region of the Flo1-family proteins.

Indeed, there are not homologous but we stated about the fact that these flocculins harbor a similar sequence architecture divided in three major region A, B and C.

l. 106-112: the bases for these statements are not convincing

Is it a question of formulation? In any case, we have revised this part of the introduction and written that:

This has been revised (-see line 95-105)

l. 138-141: sentence not clear

This has been revised (see now line 131 -135)

l. 159 delete "a" in two places, because the subject is plural

Corrected

l 157-161: Figure 3A shows ~75% inhibition of the low-adhesion domains as well. This inhibition contradicts the statement.

This sentence and this part of the results have been revised; See line (140 to 151 in the revised version).

l. 180: suggest "facilitated" rather than "warranted"

Corrected

l. 184-187: the sentence contains a double negative "…the lack of nanodomains formation in BY4741 is not solely due…" and would be clearer if rephrased.

Sentence has been rephrased (line 159 -170).

l. 206: reference should be Dranginis 2007.

Indeed, this is corrected.

l. 223-225: If the VASTTVT mutant was not made and tested, I suggest alternate wording "…Finally, it is interesting to notice that mutating the VVSTTV to VASTTVT (the sequence used in Figure 2a as anti-amyloid peptide) dropped the TANGO-predicted β-aggregation potential of the in this region of the Flo11 sequence to 26% (Table S3)." Residue position numbers would help.

Indeed, the VASTTVT variant was not made and tested experimentally. This was purely a value obtained by simulation using TANGO. Upon revision of this paper, it came to us that this sentence should be removed as it did not provide any useful information.

l. 253: give quantitative estimates or % reduction in stiffness.

We explicitly wrote that the stiffness of nanodomains on the cell surface of the Flo11-Nter expressing strain was 60% lower than that of the L69 strain.

l. 254, 321, etc.: All discussion of effects of the δ C-Ter form are suspect because we do not know whether any Flo11 is present on the cell surface, nor how it is attached if present, nor its cell-surface orientation. Labeling and detection of surface Flo11 levels would be required.

See comments above in reply to the Editor who also considered that this comment was very crucial to the paper. Overall, our immunofluorescence assay showed that all Flo11 variants include the one lacking the C-ter which bears the GPI anchor are clearly localized at the cell periphery like the normal protein.

l. 262. The adhesion peaks in Figure S10 look like 300-500 pN, rather than the 100 pN stated.

You are right indeed, but it must be insisted that the presence of these patches were extremely difficult to track on the cell surface, as they appeared only on few cells other did not show that peak. So, we did not want therefore to expand more on these peak but just mention that if they exist, they are really different from nanodomains formed by normal Flo11p.

l. 322. The loss of aggregation might also be caused by differences in cell surface concentration.

This is also a possibility although our qRT-PCR combined with the immunofluorescence may suggest that the cell surface concentration of this Flo11 construct is comparable to that in the wild type strain.

l. 331: typo: should be "not"

Corrected

l. 348: differences might also be caused by differences in cell surface concentration due to different culture methods.

This is indeed a possible reason that cannot be excluded. In addition, we found that the pH effect was not comparable between strains, which could be due to difference in expression of Flo11p with respect to pH, and not the vulture media in our condition because all strains were cultivated in the same condition. We however did not investigate further this question.

l. 418 see comment on l. 40

We do not grasp this point since the sentence only says that the strain expressing Flo11 defective in C-ter is more sensitive to cell wall drugs. We reformulate this sentence (see line 412-415).

l. 460-470: This discussion is not clear. The sentence on l. 462-3 appears to be self-contradictory: how can L69 not exhibit a flocculation phenotype? The concluding sentence of the paragraph has a double negative, and its meaning is not clear.

We completely revised this part of the discussion (now see lines 557 – 566), in particular to take into account this very justified comment.

l. 848 has typos: "Figure 6: The Flo11p cell-cell aggregation is potentiated by increased number of amyloid-forming sequences."

Corrected

Figure 8 overlapping label on 2nd column

Corrected

Figure S10 correct to '…a number obtained from a single cell of…"

The legend has been corrected.

Reviewer #2 (Recommendations for the authors):

I find the abstract a bit unfocussed and trying to convey too many messages. Please simplify it.

We have fully revised the abstract taking into account the various comments raised by the reviewers and editors and we now hope it is more in focus with the content of the paper.

Reviewer #3 (Recommendations for the authors):

In the introduction, include in the discussion the state-of-the art on Flo11p, such the structure Flo11 proteins that have been recently solved as well as functional studies on Flo11p-based cell-cell interaction (including AFM-based single-cell interaction studies),

see

– Brückner S, Schubert R, Kraushaar T, Hartmann R, Hoffmann D, Jelli E, Drescher K, Müller DJ, Oliver Essen L, Mösch HU. Kin discrimination in social yeast is mediated by cell surface receptors of the Flo11 adhesin family. eLife. 2020 Apr 14;9:e55587;

– Kraushaar T, Brückner S, Veelders M, Rhinow D, Schreiner F, Birke R, Pagenstecher A, Mösch HU, Essen LO. Interactions by the Fungal Flo11 Adhesin Depend on a Fibronectin Type III-like Adhesin Domain Girdled by Aromatic Bands. Structure. 2015 Jun 2;23(6):1005-17.

Also include recent reviews on fungal adhesins such as

– Willaert RG. Adhesins of Yeasts: Protein Structure and Interactions. J Fungi (Basel). 2018 Oct 27;4(4):119.

– Essen LO, Vogt MS, Mösch HU. Diversity of GPI-anchored fungal adhesins. Biol Chem. 2020 Nov 26;401(12):1389-1405.

Other relevant (landmark) previous research on Flo11p should be included in the introduction and/or the discussion of the results:

– Lambrechts MG, Bauer FF, Marmur J, Pretorius IS. Muc1, a mucin-like protein that is regulated by Mss10, is critical for pseudohyphal differentiation in yeast. Proc Natl Acad Sci U S A. 1996 Aug 6;93(16):8419-24..

– Bayly JC, Douglas LM, Pretorius IS, Bauer FF, Dranginis AM. Characteristics of Flo11-dependent flocculation in Saccharomyces cerevisiae. FEMS Yeast Res. 2005 Dec;5(12):1151-6. doi: 10.1016/j.femsyr.2005.05.004.

– Douglas LM, Li L, Yang Y, Dranginis AM. Expression and characterization of the flocculin Flo11/Muc1, a Saccharomyces cerevisiae mannoprotein with homotypic properties of adhesion. Eukaryot Cell. 2007 Dec;6(12):2214-21. doi: 10.1128/EC.00284-06.

– Purevdorj-Gage B, Orr ME, Stoodley P, Sheehan KB, Hyman LE. The role of FLO11 in Saccharomyces cerevisiae biofilm development in a laboratory-based flow-cell system. FEMS Yeast Res. 2007 May;7(3):372-9.

– Van Mulders SE, Christianen E, Saerens SM, Daenen L, Verbelen PJ, Willaert R, Verstrepen KJ, Delvaux FR. Phenotypic diversity of Flo protein family-mediated adhesion in Saccharomyces cerevisiae. FEMS Yeast Res. 2009 Mar;9(2):178-90.

– Goossens KV, Willaert RG. The N-terminal domain of the Flo11 protein from Saccharomyces cerevisiae is an adhesin without mannose-binding activity. FEMS Yeast Res. 2012 Feb;12(1):78-87.

All these papers are indeed relevant for the paper. We have revised the Introduction (see ie lines 78 to 90 in the revised version), taking into account some of the data related to Flo11 and data in these paper are more extensively considered in the Discussion section (and most of them have been mentioned in the initial submission) with regards to our data.

p. 4: In the description of the protein architecture the function of the A-domain as an adhesin is now missing and should be described in detail, since it is the A-domain that contains the Flo11 domain (see Pfam database: PF10182 or the interpro database). The Flo11p domain is now generally accepted as the domain of the Flo11 protein that is responsible for the cell-cell interaction.

We have provided some additional description related to A-domain in the revised manuscript (see line 68 to 84), but as written above, discussed these important data in regards to our data (see line 484 -514).

Nanodomains were observed using AFM and a bare tip. Some details on how the measurements were performed are missing from the Materials and Methods section, such as which cantilever(s) of the MLCT were used, in which buffer (or growth medium) (what was the pH?) were the experiments performed. Were the cantilevers plasma or O3 cleaned?

All these technical details are given in the Mand M, except that we did not specify that plasma with air was used to clean the cantilever before the experiment which has now been added.

A question that arises when looking at the adhesion maps and adhesion force determination is which type of interaction is present here? Is it an electrostatic or hydrophobic/hydrophilic interaction? Plasma/O3 cleaning can change the hydrophilicity of the cantilever and the adhesion strength. It would be interesting to better characterize the type of interaction since now there is no control. As an additional measurement, the cantilever could be coated with compounds that make the surface hydrophobic or positive or negative charged.

Additionally, since the type of interaction is not well characterized (as well as the conditions), it is not clear now with which part of the Flo11p (and/or the glycans (Meem 7 Cullen, 2012); since yeast glycans contain (charged) phosphates that can interact (Goossens KV, Ielasi FS, Nookaew I, Stals I, Alonso-Sarduy L, Daenen L, Van Mulders SE, Stassen C, van Eijsden RG, Siewers V, Delvaux FR, Kasas S, Nielsen J, Devreese B, Willaert RG. Molecular mechanism of flocculation self-recognition in yeast and its role in mating and survival. mBio. 2015 Apr 14;6(2):e00427-15). For example, Kraushaar et al., (Structure. 2015 Jun 2;23(6):1005-17) showed that the Flo11 domain trans-interact via a hydrophobic interaction due to the presence of 2 aromatic bands in the structure. They also discovered that this interaction only occurs at low pH (5.5) and not at higher pH (6.8)).

These aspects will influence the results on the measured adhesion forces. A bimodal distribution of the adhesion forces were observed. The weaker adhesion forces were attributed to surface proteins that unfold upon retraction of the AFM tip (p. 6, line 137; p. 7, line 160). Are these surface proteins the Flo11 proteins?

We do not fully understand your comment and do not see if there is a question behind. What can be commented on our side are:

A – AFM techniques used is all what is the more common.

B – The adhesion forces (what you mention as interaction) have been analysed in detail for each strain expressing the different variant. Only nanodomains from L69 strain expressing the wild type Flo11 exhibits a bimodal distribution of adhesion forces.

C – The weaker adhesion force is likely due to protein being unfolded, which can be Flo11 but not exclusively indeed as we used a bare tip that interacts with cell surface component in general, not specifically. This can be seen in Figure 3 in which this low-adhesion forces was still observed but a lower frequency upon treatment of the cells with the anti-amyloid peptide.

There is no direct evidence from the results that these nanodomains are composed of Flo11p. This could be obtained by adhesion maps using an anti-Flo11p antibody or with the purified Flo11p, or Flo11p on other cells by performing single cell force-spectroscopy (Brückner et al., eLife. 2020 Apr 14;9:e55587).

The evidence that the nanodomains are composed of Flo11p are:

A – deletion of FLO11 in strain L69 completely abrogates the formation of adhesive nanodomains upon AFM tip.

B – the laboratory BY strain that does not form nanodomains with its own FLO11 gene even after overexpression, but only upon ectopic expression of this peculiar Flo11p from strain L69.

C – nanodomains are not produced in cells that express a Flo11 lacking its amyloid-core –sequence.

D – the finding that purified Flo11p forms amyloids in vitro has been already reported by Lipke group (Ramsook, EC, 2010). This is the most direct evidence that nanodomains are made from clusters of Flo11p!

What was the effect of lowering the maximal applied force on the adhesion forces?

We did not carry out this experiment and do not believe it will bring anything new.

If the weaker adhesion forces were due to unfolding of proteins, this should be visible in the force-distance curves. If it was not visible, this hypothesis should be rejected. Please provide these graphs.

This is indeed visible, as indicated in Figure 2E or Figure 3A, Figure 5 —figure supplement 1 and 2 an example. All raw data have been deposited at https://datadryad.org/

p. 11, line 254: "On the other hand, the cell surface of L69flo11-∆Cter cell was characterized by needle-shaped nanostructures whose height was 3 times greater than nanodomains formed on the surface of the L69 strain. These nanodomains displayed adhesion forces that were scattered from a few pN to max 200 pN and stiffness that was roughly 30% higher than that of nanodomains from L69 strain (see Figure S9, in supplementary data)."

– How could these results be explained? When the C-terminal part of the protein has been removed, is it still anchored in the cell wall? Is there proof of this? Suppose that the truncated Flo11p is present: Does the bare AFM tip interact specifically with the remainder of Flo11p and does cis-interaction results again in these nanodomains? Why are these nanodomains then larger?

The force-distance curves shown in Figure S.9 G, indicate 2 slopes indicating that first a softer layer is pushed through. Not clear what this could be?

As reported, we carried out immunofluorescence assay with all the Flo11 variants and reported in a new figure (Figure 4) that they have the same localization as the wild type (so clearly marked at the cell periphery), indicating that the loss of C-ter does not prevent anchorage of Flo11 at the cell wall. How this protein could be retained would need additional work that is out of the scope of this study.

We then performed AFM analysis on cells expressing these Flo11 variants and with the exception of Flo11-RR2, they all showed nanodomains formation with however a very different typology and a loss of bimodal behaviour. There is no reason that the tip does not interact with the reminder of the Flo11 protein, but we cannot exclude that the expression of this truncated form of Flo11p has somehow altered the cell surface, leading to this adhesion histogram. However, these data indicated that the high adhesion forces that result from amyloid-dependent clustering of Flo11p are no longer present. We have carefully revised our paper to take into account this remark (see now lines 266 -282 in the revised version).

Finally, the force-distance curves shown in Figure 9G is an example for the specific cell of this figure. It shows 2 slopes, but the stiffness has been calculated on the first one.

p. 12, line 293: …"This residual clumping is likely a specific feature from Σ1278b background strain, which does not hamper the notion that Flo11p is critical in cell-cell interaction. This is further supported by the lack of aggregates in BY4741 as this strain does not express FLO11 due to non-sense mutation in FLO8 encoding its major transcriptional activator.

– Not clear what is meant with a "specific" feature? How could this remaining interaction be explained? Remark that Brückner et al., (2020) demonstrated quantitatively using single-cell force spectroscopy that the Flo11p A-domain is needed to have a significant cell-cell interaction (although some variation is present in these single-cell experiments).

– Were these aggregation experiments also performed at a lower pH and could the results of Kraushaar et al., (2015) be confirmed?

The ∑1278 haploid strain exhibits a very different aggregation phenotype with respect to L29 strain, as deletion of FLO11 did not completely abolish this phenotype as reported in Figure 6 (now 7 in the revised version). Whether the remaining aggregation is due to the expression of another FLO gene is unlikely since Flo11p is the only flocculin expressed in this strain. As FLO gene is under silencing, it cannot be excluded that deletion of FLO11 has allowed desilencing of some FLO through epigenetic mechanism as (reported by Halme et al., Cell, 116, 405-415, 2004). This is only a speculation because we did not verify this suggestion. We revised this paragraph accordingly (see lines 302-312).

Our experiments of cell-cell interaction in YNGal agar was made at pH 5.0 which is a condition reported by Kraushaar et al., to favour hemophilic Flo11-Flo11 interaction. Our results are in part in support the finding of Kraushaar et al., although we did not show it directly. Although our data show that amyloid-core sequence contribute to cell-cell interaction (see figure 6), they are not absolutely essential for this phenotype as shown with the lack of effect of the anti-amyloid agents thioflavin S to be unable to disrupt aggregates formed with cells from ∑1278b strain. This data indirectly indicates that other components are implied in this phenotype, which is likely due to the A-domain (revised in lines 312 -315).

Finally, we tested the adherence at different pH and confirmed that this phenotype was sensitive to pH (hence higher at acidic than alkali pH, see figure 7). But surprisingly, we found that this pH effect was not similar between yeast strains. While strains ∑1278C showed better adherence at pH 8.0 than pH 5.0, this was the opposite for L69 strain and BY4741!. This difference would need additional experiments that is not in the scope of this work.

In our hands, the contribution of the N-terminal domain was more important at pH 8.0 than pH 5.0 for the Flo11p of L29 strain, whereas the amyloid core sequence has no role. However, what was striking was the finding that the adherence to plastic in quantitative term was dramatically increases upon overexpression of FLO11 gene whatsoever the origin of the gene (either FLO11 from BY or from L69). This result somehow alerted us because in the experiments carried out by Kraushaar et al., (Structure, 2016), FLO11 is expressed in CEN plasmid under the dependence of the strong PGK1 promoter. Thus, it may be possible that under these conditions, the effect of removal A-domain was quantitatively more effective.

– It has been shown recently that the BY4741 strain can flocculate depending on the condition and this flocculation was independent of Flo5p (see Degreif D, de Rond T, Bertl A, Keasling JD, Budin I. Lipid engineering reveals regulatory roles for membrane fluidity in yeast flocculation and oxygen-limited growth. Metab Eng. 2017 May;41:46-56). This has been shown for Flo1p. In many conditions, Flo1p and Flo11p are be coexpressed. Remark that the expression of FLO11 is very complex and control has been attributed to many transcription factors (Brückner S, Mösch HU. Choosing the right lifestyle: adhesion and development in Saccharomyces cerevisiae. FEMS Microbiol Rev. 2012 Jan;36(1):25-58.).

Has it been assessed that the flocs that were observed could be deflocculated by adding EDTA? Was the expression of other Flo proteins also measured?

– How many aggregates were counted to obtain the results? It is not clear if this method can detect significant small changes in interaction forces.

– An additional quantitative method of measuring the interaction between cells is needed to confirm the obtained results based on observing aggregates.

BY strain bears a non-sense mutation in FLO8 which prevents expression of FLO1 and FLO11 (Liu et al., Genetics, 144, 967-978, 1996; Fichtner et al., Mol Microbiol., 66: 1276-1289, 2007). Van Mulders et al., (FEMS Yeast Res, 9: 178 -190, 2009) did an extensive functional analysis of expressing each of the FLO gene in this strain background, and get data related the phenotype of each FLO genes, which are in line with our data.

In a previous paper (Schiavone et al., FEMS Yeast Res 2015), we reported transcriptomic data on L69 strain showed that FLO11 was the most expressed, but we also identified very weak expression of FLO1, FLO5 and FLO10, as well as FIGURE 2.

The procedure for counting cell aggregates are described more extensively now in M and M SCFS has been used by Bruckner et al., (eLife 2020) and Fluid Force microscopy as developed by Dufrene and coll (see Nanoletters, 2019) could be a good mean to measure the force between cells. We believe that this is not needed for the present work but these experiments are considered in a future work.

p. 14, line 330: … "Adherence was assayed on polystyrene surfaces."

– Note that this was also investigated before: see Van Mulders et al., 2009, and should be included in the discussion (since here also the adhesion of cells that express other Flo proteins was investigated).

This paper has been referenced but the methodology was not established by these authors but by Reynolds and Fink (Science 2011). Also, we only restricted the discussion to Flo11 protein and its variant an not to the other Flo protein which was not relevant in this paper.

p. 14, line 335: "Moreover, the moderate adherence of BY4741 cell to plastic at pH 5.0 must be due to cell wall proteins other than Flo11p, since this gene is not expressed in this strain due to a non-sense mutation in FLO8 encoding its transcriptional activator (Liu et al., 1996).

– Same comment as before: Under the selected conditions, was it confirmed that no other Flo proteins were expressed?

We did not check for expression of other FLO gene in BY strain. But we may refer to the paper of Halme et al., (Cell, 116, 405-415, 2004) that FLO gene is under epigenetic regulation, and that their expression is metastably silenced. Although this regulation was examined in ∑1272b strain, it might be also valuable in BY strain. We have rephrased our discussion taking into account these data, although this is quite speculative and this suggestion shall merit to be verified, which is out of the scope of this work (see lines 290 -294).

p. 16 and discussion: Invasive growth experiments => discuss the results also in relation to previous published results on this (such as Van Mulders et al., 2009).

Here, we compared the invasion phenotype between Flo11 variant not between different Flo proteins, thus except indicating that our invasive phenotype of FLO11 from BY strain agrees with data of Van Muders et al., there is no need to expend much on their results.

p. 19, line 454: "The finding that this phenotype was not abrogated when domain A was deleted could be at first glance in contradiction with the data of Kraushaar et al., (Kraushaar et al., 2015) who showed that this domain is essential for conferring cell-cell adhesion by facilitating homotypic interactions".

– Also in contrast to others: Bayly et al., (2005), Goossens and Willaert (2012).

However, the lack of A domain can be in part overcome by amyloid-forming sequences as these later could promote clustering of Flo11 molecules in cis, which enables cell-cell adhesion by trans-interaction as proposed by these same authors (Kraushaar et al., 2015, Bruckner et al., 2020).

– In the model hypothesis of Kraushaar et al., 2015; Bruckner et al., 2020 (based and their results as well on previous results from e.g. Ramsook et al.,), they propose that there is trans interaction between interacting cells, which is based on the interaction of the Flo11 domains, and that cis-interaction between various Flo11 proteins via the β-aggregation-prone amyloid-forming sequences could precede the trans-interaction. This model needs further experimental confirmation to be generally accepted.

Recent data from Lipke et coll (see Ho et al., mBIO. mBio. 2019 Oct 8;10(5):e01766-19. doi: 10.1128/mBio.01766-19; Dehullu et al., 2019, Nano Lett. 2019 Jun 12;19(6):3846-3853. doi: 10.1021/acs.nanolett.9b01010) have now clearly demonstrated that amyloid-forming sequences are required for cis-interaction (interaction between adhesion molecule) leading to cluster of adhesion molecule and to trans-interaction, ie interaction cell to cell. Although this demonstration was done with ALS1 and ALS5 encoding Als adhesion of C. albicans, our data are in line with this statement. Indeed, we show that the nanodomains at the cell surface of L69 strain are due to the presence of several amyloid forming sequences present in this particular Flo11p, and this leads to cis-interaction which can be abolished using either the anti-amyloid peptide or an amyloid perturbant, and likely to trans-interaction because loss of RR2 domain that contains these amyloid sequences results in a strong reduction of cell-cell interaction (or cell-cell aggregation).

It could be that the observed nanodomains correspond to trans-interacting Flo11 proteins. However, the cell-cell interaction would be significantly reduced if the Flo11p domains are not present in the truncated protein. This is however in contrast to what has been observed by the authors in this manuscript. Therefore, it is not clear how this can be explained.

This is not what we show in Figure 6 and 7 since the loss of RR2 that contains the amyloid forming sequences results in > 70% reduction of cell-cell interaction, whereas insertion of this fragment in the Flo11 from BY strain allows cells of this strain to form aggregate at a rate that higher than by ectopic overexpression of its endogenous FLO11 (see figure 6).

Author details

1. Clara Bouyx

   Toulouse Biotechnology Institute, INSA, Toulouse, France

   Contribution

   Investigation, Methodology, Writing - original draft

   Competing interests

   none

2. Marion Schiavone

   1. Toulouse Biotechnology Institute, INSA, Toulouse, France
   2. Lallemand, Lallemand SAS, Blagnac, France

   Contribution

   Conceptualization, Investigation, Methodology, Writing - original draft

   Competing interests

   none

3. Marie-Ange Teste

   Toulouse Biotechnology Institute, INSA, Toulouse, France

   Contribution

   Investigation, Methodology

   Competing interests

   none

   "This ORCID iD identifies the author of this article:" 0000-0001-9173-9190

4. Etienne Dague

   LAAS, CNRS, Toulouse, France

   Contribution

   Investigation, Methodology, Supervision

   Competing interests

   none

5. Nathalie Sieczkowski

   Lallemand, Lallemand SAS, Blagnac, France

   Contribution

   Conceptualization, Methodology

   Competing interests

   none

6. Anne Julien

   Lallemand, Lallemand SAS, Blagnac, France

   Contribution

   Conceptualization, Methodology, Supervision

   Competing interests

   none

7. Jean Marie François

   Toulouse Biotechnology Institute, INSA, Toulouse, France

   Contribution

   Conceptualization, Funding acquisition, Supervision, Writing - original draft, Writing - review and editing

   For correspondence

   fran_jm@insa-toulouse.fr

   Competing interests

   none

   "This ORCID iD identifies the author of this article:" 0000-0001-9884-5535

• 672

  Page views

• 101

  Downloads

• 7

  Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.