- Research article
- Open Access
Nickel and low CO2-controlled motility in Chlamydomonasthrough complementation of a paralyzed flagella mutant with chemically regulated promoters
BMC Plant Biologyvolume 11, Article number: 22 (2011)
Chlamydomonas reinhardtii is a model system for the biology of unicellular green algae. Chemically regulated promoters, such as the nickel-inducible CYC6 or the low CO2-inducible CAH1 promoter, may prove useful for expressing, at precise times during its cell cycle, proteins with relevant biological functions, or complementing mutants in genes encoding such proteins. To this date, this has not been reported for the above promoters.
We fused the CYC6 and CAH1 promoters to an HA-tagged RSP3 gene, encoding a protein of the flagellar radial spoke complex. The constructs were used for chemically regulated complementation of the pf14 mutant, carrying an ochre mutation in the RSP3 gene. 7 to 8% of the transformants showed cells with restored motility after induction with nickel or transfer to low CO2 conditions, but not in non-inducing conditions. Maximum complementation (5% motile cells) was reached with very different kinetics (5-6 hours for CAH1, 48 hours for CYC6). The two inducible promoters drive much lower levels of RSP3 protein expression than the constitutive PSAD promoter, which shows almost complete rescue of motility.
To our knowledge, this is the first example of the use of the CYC6 or CAH1 promoters to perform a chemically regulated complementation of a Chlamydomonas mutant. Based on our data, the CYC6 and CAH1 promoters should be capable of fully complementing mutants in genes whose products exert their biological activity at low concentrations.
Chlamydomonas reinhardtii is a unicellular green alga, capable of both photosynthetic and fermentative growth. A plethora of mutants in relevant biological processes are available, and nuclear and chloroplast transformation are easy to perform . Its 120-megabase genome has been completely sequenced . Chlamydomonas combines functions typical of higher plants, such as the presence of a chloroplast endowed with two photosystems , of protozoa, such as the presence of motile flagella for swimming , and of archaea, such as the presence of sensory rhodopsins mediating phototaxis .
Flagellar motility in Chlamydomonas is dependent on dynein motors, which drive microtubule sliding, and a multitude of accessory proteins that control dynein activity, including radial spokes and the central pair complex. Immotile mutants missing individual subunits of these components have been identified and, in many cases, rescued by introducing the corresponding wild-type gene driven by its native promoter [6, 7]. The first case of such complementation was achieved in a mutant, pf14, which has paralyzed flagella due to a premature stop codon in the gene encoding radial spoke protein 3 (RSP3) . RSP3 encodes a protein mediating the anchoring to the axoneme of a cAMP-dependent protein kinase that regulates axonemal motility and dynein activity [9, 10]. Flagellar motility can be restored by transformation of the mutant with the wild-type RSP3 gene , thus providing a nice biological assay for activity of the promoter driving RSP3 transcription.
Several chemically regulated promoters have been described in Chlamydomonas: the Nitrate Reductase (NIT1) promoter, induced by ammonium starvation ; the Carbonic Anhydrase (CAH1) promoter, induced by low CO2 ; and the Cytochrome C6 (CYC6) promoter, induced by copper (Cu) depletion or nickel (Ni) addition [13, 14]. In all three cases, inducible expression has been demonstrated using reporter genes such as arylsulfatase or luciferase and, in the case of the NIT1 promoter, through complementation of a paralyzed flagellar mutant, pf14, by expressing the wild type form of the RSP3 gene . No data are available, to our knowledge, on the capacity of the CAH1 and CYC6 inducible promoters to drive complementation of Chlamydomonas mutants.
To assess the capacity of the CYC6 and CAH1 promoters to complement the pf14 mutation in a chemically regulated fashion, we transformed the paralyzed pf14 mutant with the RSP3 gene under the control of the above-mentioned promoters and scored the swimming phenotype. The strong constitutive PSAD promoter  was used as a control.
Constructs used for chemically inducible complementation
The complete RSP3 gene (including introns) was translationally fused to a 9-amino acid HA epitope at its 3' end, to facilitate the immunodetection of the expressed protein . The RSP3-HA hybrid gene was placed under the control of the CYC6 and CAH1 promoters, induced, respectively, by Ni and low CO2 [13, 14, 12] and, as a control, of the strong constitutive PSAD promoter . The constructs are schematically represented in Figure 1.
Constitutive complementation of the pf14 mutant driven by the PSADpromoter
The pf14 mutant strain was transformed with the PSAD:RSP3-HA plasmid and 68 paromomycin-resistant transformants were grown for 4 hours without shaking in the light. Upon microscopic examination, about 40% of the transformants showed swimming cells. The average percentage of swimming cells was about 80% (Table 1). This result shows that the RSP3-HA fusion protein is able to rescue the pf14 mutant. The data of a representative transformant are shown in Figure 2. The majority of the cells (88%) were flagellated and motile (Panel A) and strong signals corresponding to the unphosphorylated (lower band) and phosphorylated (upper band) forms of the RSP3-HA protein were detected in a Western blot using the anti-HA antibody (Panel B).
Chemically inducible complementation of pf14 driven by the CYC6 and CAH1promoters
pf14 cells were transformed with the CYC6:RSP3-HA and the CAH1:RSP3-HA plasmids and 68 transformed colonies were analyzed for each construct. Before analysis, the CYC6:RSP3-HA transformants were inoculated in TAP ENEA2 medium, allowing optimal expression of the CYC6 promoter, and expression was induced in the mid-log phase (6 × 106-8 × 106 cells/ml) by adding 25 μM Ni . We used a rather low Ni concentration, since higher concentrations cause detachment of flagella, preventing the scoring of the swimming phenotype (data not shown). The swimming phenotype was scored 48 hours after induction, when CYC6 promoter expression is maximal . Approximately 8% of the transformants displayed swimming (Table 1) and, in these transformants, an average of 5% of the cells were motile. This difference with respect to the PSAD:RSP3-HA transformants is due to two factors: a much lower percentage of cells are flagellated in the CYC6:RSP3-HA transformants (12% vs 90%) and, of these, a lower percentage are swimming (Table 1). As discussed below, we attribute this difference to a threshold effect. Movies of PSAD:RSP3-HA and CYC6:RSP3-HA transformants are available as Additional files 1 and 2.
In order to prevent loss of flagella at high cell densities (see below), cells were also induced with 25 μM Ni in early log phase (1 × 106-2 × 106 cells/ml), but no rescue was observed (data not shown). This is consistent with the observation of Quinn et al.  that activation of the CYC6 promoter is stronger when the cells are induced at mid-late log phase, probably because Ni uptake is higher.
The CAH1:RSP3-HA transformants were grown in air containing 5% CO2, in minimal medium supplemented with extra phosphate buffer to keep the pH stable. Expression of the CAH1 promoter was induced in early-log phase by transferring the plate to air and cells were scored for swimming 6 hours after induction, when the CAH1 promoter shows high expression . Approximately 7% of the transformants showed swimming and, as for the CYC6:RSP3-HA transformants, approximately 5% of the cells were motile in the rescued transformants (Table 1).
The percentage of swimming and flagellated-immotile cells was determined for two representative CYC6:RSP3-HA transformants showing restored motility, 48 hours after Ni addition (Figure 3), when the CYC6 promoter shows high expression  and cell density is high (1 × 107-2 × 107 cells/ml). The percentage of swimming cells was about 5% in both cases, whereas the flagellated/immotile cells ranged between 5% and 8%. Loss of flagella is independent of addition of Ni at 25 μM, since it is observed also in the non-induced transformants at high cell densities (Figure 3, gray bars). Cell density-dependent loss of flagella is not observed in wild type or PSAD:RSP3-HA transformants, suggesting that continuous, or high level, expression of RSP3 prevents this phenomenon.
The percentage of swimming cells in two representative CAH1:RSP3-HA transformants was determined 6 hours after transfer to low CO2 (Figure 4), when the CAH1 promoter shows high expression . In this case, cell density was low (2 × 106-4 × 106 cells/ml) and the percentage of flagellated cells was high (approx. 90%). However, as for the CYC6:RSP3-HA transformants, the percentage of motile cells was low (5%-6% of total cells).
Figure 5 (Panels A and B) shows a Western blot of several CYC6:RSP3-HA and CAH1:RSP3-HA transformants, grown in the same conditions of Figures 3 and 4, and probed with an anti-HA antibody. Only transformants showing motility in the swimming assay (Figures 3 and 4) showed the two bands corresponding to the RSP3 protein. The signal of the two bands is very weak compared to the PSAD:RSP3-HA transformants, suggesting that the low percentage of swimming cells is probably due to low expression of the RSP3 protein. The swimming transformants were re-grown in the same conditions used in Figure 6, and probed 0 h and 48 h (CYC6 transformants) or 0 h and 6 h (CAH1 transformants) after induction. The results (Panel C) show that the RSP3 protein is completely absent in non-induced, and readily detectable in induced cells.
Kinetics of induction of the swimming phenotype
We then determined the kinetics of appearance of swimming cells in one representative CYC6:RSP3-HA and one CAH1:RSP3-HA transformant (Figure 6). In the case of the CYC6 promoter, swimming cells were observed as early as 24 hours after Ni addition. At 48 hours the number of swimming cells reached a maximum and then decreased at 72 hours. This is in agreement with the kinetics of activation of the CYC6 promoter, measured with the luciferase reporter gene, which reaches maximum activity after two days of induction and then decreases at three days . In the CAH1:RSP3-HA transformant, swimming cells were observed as early as 2 hours after transfer to low CO2. The maximum number of swimming cells was reached 5 hours after transfer, and then declined at 8 hours. However, considering the standard deviations at 5 and 8 hours, this decline is not significant.
Through the use of the chemically regulated CYC6 and CAH1 promoters and of a genomic RSP3 clone fused to an HA epitope, we have achieved the chemically regulated motility of Chlamydomonas cells. While the vast majority of cells showed motility when RSP3 expression was driven by the constitutive PSAD promoter, only a minority (5%) of cells showed motility after induction of the CYC6 and CAH1 promoters. This is probably due to a threshold effect: the levels of RSP3-HA protein driven by PSAD are much higher than those driven by CYC6 and CAH1. The low levels of RSP3-HA protein expressed from the CAH1 promoter after 6 hours of induction contrast markedly with the high levels of CAH1 protein expressed from the endogenous gene (data not shown). Low expression of exogenously introduced constructs in Chlamydomonas is a well-known phenomenon, which has been attributed to gene silencing .
The low levels of RSP3-HA protein expressed from the CYC6 promoter after 48 hours of induction are also puzzling, since, in TAP ENEA2 medium, the CYC6 promoter is able to drive levels of luciferase expression comparable to those driven by PSAD . We attribute this difference in RSP3 vs luciferase expression to the fact that RSP3 accumulates over time when it is expressed from PSAD, while expression for 48 hours (from CYC6) or 6 hours (from CAH1) allows accumulation of low RSP3 levels (Figure 5). This implies that the RSP3 protein is more stable than luciferase (whose estimated half-life in Chlamydomonas is <2 hours ). Whatever the case, the low levels of expressed RSP3-HA protein are sufficient for achieving motility in 5% of the transformed cells. To our knowledge, this is the first example of the use of the CYC6 and CAH1 promoters for achieving chemically regulated complementation of a Chlamydomonas mutant, as well as the first example of metal- or CO2-regulated motility engineered in a living organism. The partial complementation observed is probably due to the fact that the RSP3 protein, to exert its function, is required in high concentrations. Although the number of radial spokes required to restore motility to flagella is not known, each wild type flagellum contains approximately 2,000 radial spokes .
Zhang and Lefebvre  have used the RSP3 gene under the control of the ammonium-repressible NIT1 promoter to complement the pf14 mutant in a nitrogen source-dependent fashion. In that study, 81 out of 2,000 cotransformants showed motility in permissive conditions, i.e. a fraction of about 4%, comparable to the 7-8% reported here for the CYC6 and CAH1 promoters. At least one of the transformants, containing multiple copies of the NIT1:RSP3 plasmid, showed full complementation, i.e. a large number of swimming cells, a fact we did not encounter in the case of the CYC6 and CAH1 promoters, probably due to the smaller number of colonies screened in our study and to the fact that the vast majority of the insertions, in our case, are single-copy (Additional file 3). Whatever the case, the frequency of swimming transformants obtained with the strong PSAD promoter is 40% (Table 1), i.e. much higher than what can be obtained using either the NIT1, CYC6, or CAH1 promoters in permissive conditions. A chemically regulated promoter system allowing such high complementation efficiencies in permissive conditions has yet to be worked out.
We have demonstrated low level, chemically regulated complementation of the paralyzed flagella pf14 mutant by the RSP3 gene, encoding a component of the flagellar radial spoke complex, cloned under the control of the CYC6 and CAH1 promoters. Maximum complementation is reached with very different kinetics (6 hours for CAH1, 48 hours for CYC6). In principle, these promoters should be capable of fully complementing mutants in genes whose products exert their biological activity at low concentrations (e.g. receptor/signalling protein kinases). Test of this hypothesis is under way, as well as the optimization of the CYC6 and CAH1 promoters, for full complementation of mutants in genes encoding abundant intracellular proteins.
Strains and culture conditions
The paralyzed flagella mutant pf14  was used for all experiments. Nuclear transformation was performed as described . Plasmids were digested with Sca I and 300 ng of DNA were used for each transformation. Transformants were selected on TAP agar plates containing paromomycin (10 μg/ml).
Unless indicated differently, cells were grown photomixotrophically in TAP medium at 25°C under irradiation (16 L: 8 D) with fluorescent white light (200 μE m-2 s-1). For the initial screening, 68 transformants for each construct were grown in 200 μL in 96-well microtiter plates with shaking (900 rpm). For quantitative measurements of motility (Figures 2, 3, 4, 5A and 5B), transformants were grown in 2 mL in 24-well microtiter plates with shaking (500 rpm). For the experiments described in Figure 5C and in Figure 6 cells were grown in 6 ml in 6-well microtiter plates with shaking (120 rpm) (CYC6 transformants) or in 150 ml in 250-ml Erlenmeyer flasks (CAH1 transformants) with bubbling. The plates were covered with Breathe-Easy membrane (Diversified Biotech, cat. BEM-1), to prevent evaporation without limiting gas and light exchange. For Ni induction, cells were grown in TAP ENEA2 medium  and induced at mid-log phase (6 × 106-8 × 106 cells/ml) by adding 25 μM Ni. For low-CO2 induction, cells were grown in minimal medium with doubled phosphate buffer concentration, to keep the pH stable in high CO2 conditions , in air containing 5% CO2 and induced by shifting to air in early log phase (1 × 106-2 × 106 cells/ml).
The complete RSP3 gene (including introns) was amplified using the following oligonucleotides:
Reverse: GAAGATCT TTA GGCGTAGTCGGGCACGTCGTAGGGGTACGCGCCCTCCGCCTCGGCGAAC
The forward oligonucleotide inserts an Xba I restriction site, the reverse oligonucleotide inserts a 9- amino acid HA-tag (the corresponding nucleotide sequence is in italics) followed by a TAA stop codon and a Bgl II restriction site (both restriction sites are in bold). The two oligonucleotides were used to amplify the RSP3 gene and the RSP3-HA insert was used to replace the cRLuc sequence in the PSAD:cRLuc and CYC6:cRLuc plasmids . To construct the CAH1: RSP3-HA plasmid, the -651 +41 region of the CAH1 promoter was amplified with the following oligonucleotides: CAH1 forward: CCGCTCGAGTCAGCTTCTCTCCCGCCAGC; CAH1 reverse: GCTCTAGAGGTGTTCAAGTGGGTTGCAG. The CAH1 forward oligonucleotide inserts an Xho I restriction site, the CAH1 reverse primers inserts an Xba I restriction site (both restriction sites are in bold). The two oligonucleotides were used to amplify the -651 +41 region of the CAH1 promoter and the insert obtained was used to replace the CYC6 sequence in the CYC6:RSP3-HA plasmid.
Transgene copy number was estimated via quantitative Real-Time PCR. Total DNA was extracted from the PSAD:RSP3-HA, CYC6-RSP3-HA and CAH1-RSP3-HA transformants and from the pf14 strain using the DNeasy Plant Mini Kit (Qiagen, cat. N. 69104). A calibration curve (Additional file 3) was made by mixing 10 ng of total DNA from the pf14 strain with different amounts of linearized PSAD:RSP3-HA plasmid, corresponding to 0.5, 1 and 2 copies per genome. The CYC6 gene was used as an internal standard for normalization. The oligonucleotides used to amplify the RSP3-HA transgene are: RSP3HA forward: TACGCCTAAAGATCTGAATTCGG; RSP3HA reverse: TCAGCGAAATCGGCCATC. These oligonucleotides amplify the PSAD:RSP3-HA, CYC6:RSP3-HA, CAH1:RSP3-HA constructs and the corresponding transformants, but not the endogenous RSP3 gene. The oligonucleotides used to amplify the CYC6 gene are: CYC6 forward; TGGATTTCCTCCTCCGACAG; CYC6 reverse: CTGGATGGCGGCTTCAAG. The Real-Time PCR amplification conditions were: 95°C, 10 min; 50 cycles of 95°C, 15 sec and 60°C, 1 min. The SYBR Green PCR Master Mix (Applied, Cat. N. 4309155) was used in a final volume of 20 μl.
Proteins electrophoresis and Western blotting
Two ml of Chlamydomonas culture were centrifuged at 10,000 × g at room temperature for 1 minute and the cell pellets were resuspended in 300 μl of 60 mM DTT, 60 mM Na2CO3, 2% SDS, 12% sucrose and shaken for 20 minutes at room temperature to extract the proteins. The protein extracts were centrifuged at 10,000 × g for 1 minute and the supernatant collected. To measure protein concentration, 10 μl of protein extracts were mixed with 800 μl of 0.5% Amido Black in 90% methanol and 10% glacial acetic acid. The samples were vortexed and centrifuged at 10,000 × g at 4°C for 10 minutes. The pellets were washed two times with 90% methanol and 10% glacial acetic acid at 4°C. Finally the pellets were resuspended in 800 μl of 0.2 M NaOH and the absorbance measured at 615 nm. To calculate protein concentration a standard curve with BSA was used.
Western blot analyses were performed on total protein extracts obtained as described above. About 30 μg protein/sample were separated on an 8% SDS-PAGE gel . Proteins were blotted on nitrocellulose membrane in a buffer containing 25 mM Tris, 192 mM glycine, 20% ethanol for 2 hours at 250 mA using a Hoefer TE22 apparatus. A commercial anti-HA antibody (Ascites Fluid Mono HA 11, 16B12, Covance) was used in a 1:250 dilution. The secondary antibody (anti-mouse, phosphatase conjugated, Thermo Scientific 31325) was used in a 1:2500 dilution. Detection was performed placing the nitrocellulose membrane in 100 mM NaCl, 5 mM MgCl2, 100 mM Tris (pH 9.5) containing Nitrotetrazolium Blue chloride (NBT) 0.33 mg/ml and 5-Bromo-4-chloro-3-indolyl phosphate disodium salt (BCIP) 0.165 mg/ml. To stop the reaction, the membrane was rinsed with Phosphate-Buffered saline (PBS) containing 20 mM EDTA.
Cell motility and the presence of flagella were assessed using an Olympus BX41 microscope with 16 X and 40 X objective lenses, respectively. Movies were recorded with a Cool Snap HQ camera (Photometrics) on a Nikon Eclipse TE2000 inverted microscope using a 10 X objective lens.
Stern D, Harris EH, Witman GB, (Eds.): The Chlamydomonas Sourcebook. New York: Academic Press; 2008,
Merchant SS, Prochnik SE, Vallon O, Harris EH, Karpowicz SJ, Witman GB, Terry A, Salamov A, Fritz-Laylin LK, Marechal-Drouard L, et al: The Chlamydomonas genome reveals the evolution of key animal and plant functions. Science. 2007, 318: 245-250. 10.1126/science.1143609.
Rochaix JD: Assembly, function, and dynamics of the photosynthetic machinery in Chlamydomonas reinhardtii. Plant Physiol. 2001, 127: 1394-1398. 10.1104/pp.010628.
Wilson NF, Iyer JK, Buchheim JA, Meek W: Regulation of flagellar length in Chlamydomonas. Semin Cell Dev Biol. 2008, 19: 494-501. 10.1016/j.semcdb.2008.07.005.
Sineshchekov OA, Jung KH, Spudich JL: Two rhodopsins mediate phototaxis to low- and high-intensity light in Chlamydomonas reinhardtii. Proc Natl Acad Sci USA. 2002, 99: 8689-8694.
Diener DR, Curry AM, Johnson KA, Williams BD, Lefebvre PA, Kindle KL, Rosenbaum JL: Rescue of a paralyzed-flagella mutant of Chlamydomonas by transformation. Proc Natl Acad Sci USA. 1990, 87: 5739-5743. 10.1073/pnas.87.15.5739.
Mitchell DR, Kang Y: Identification of oda6 as a Chlamydomonas dynein mutant by rescue with the wild-type gene. J Cell Biol. 1991, 113: 835-842. 10.1083/jcb.113.4.835.
Williams BD, Velleca MA, Curry AM, Rosenbaum JL: Molecular cloning and sequence analysis of the Chlamydomonas gene coding for radial spoke protein 3: flagellar mutation pf-14 is an ochre allele. J Cell Biol. 1989, 109: 235-245. 10.1083/jcb.109.1.235.
Howard DR, Habermacher G, Glass DB, Smith EF, Sale WS: Regulation of Chlamydomonas flagellar dynein by an axonemal protein kinase. J Cell Biol. 1994, 127: 1683-1692. 10.1083/jcb.127.6.1683.
Gaillard AR, Diener DR, Rosenbaum JL, Sale WS: Flagellar radial spoke protein 3 is an A-kinase anchoring protein (AKAP). J Cell Biol. 2001, 153: 443-448. 10.1083/jcb.153.2.443.
Ohresser M, Matagne RF, Loppes R: Expression of the arylsulphatase reporter gene under the control of the nit1 promoter in Chlamydomonas reinhardtii. Curr Genet. 1997, 31: 264-271. 10.1007/s002940050204.
Kucho K, Ohyama K, Fukuzawa H: CO(2)-responsive transcriptional regulation of CAH1 encoding carbonic anhydrase is mediated by enhancer and silencer regions in Chlamydomonas reinhardtii. Plant Physiol. 1999, 121: 1329-1338. 10.1104/pp.121.4.1329.
Quinn JM, Kropat J, Merchant S: Copper response element and Crr1-dependent Ni(2+)-responsive promoter for induced, reversible gene expression in Chlamydomonas reinhardtii. Eukaryot Cell. 2003, 2: 995-1002. 10.1128/EC.2.5.995-1002.2003.
Ferrante P, Catalanotti C, Bonente G, Giuliano G: An Optimized, Chemically Regulated Gene Expression System for Chlamydomonas. PLoS ONE. 2008, 3: e3200-10.1371/journal.pone.0003200.
Zhang D, Lefebvre PA: FAR1, a negative regulatory locus required for the repression of the nitrate reductase gene in Chlamydomonas reinhardtii. Genetics. 1997, 146: 121-133.
Fischer N, Rochaix JD: The flanking regions of PSAD drive efficient gene expression in the nucleus of the green alga Chlamydomonas reinhardtii. Mol Genet Genomics. 2001, 265: 888-894. 10.1007/s004380100485.
Kelman Z, Yao N, O'Donnell M: Escherichia coli expression vectors containing a protein kinase recognition motif, His6-tag and hemagglutinin epitope. Gene. 1995, 166: 177-178. 10.1016/0378-1119(95)00556-7.
Cerutti H, Johnson AM, Gillham NW, Boynton JE: Epigenetic silencing of a foreign gene in nuclear transformants of Chlamydomonas. Plant Cell. 1997, 9: 925-945. 10.1105/tpc.9.6.925.
Fuhrmann M, Hausherr A, Ferbitz L, Schodl T, Heitzer M, Hegemann P: Monitoring dynamic expression of nuclear genes in Chlamydomonas reinhardtii by using a synthetic luciferase reporter gene. Plant Mol Biol. 2004, 55: 869-881.
Hopkins JM: Subsidiary components of the flagella of Chlamydomonas reinhardii. J Cell Sci. 1970, 7: 823-839.
Kindle KL: High-frequency nuclear transformation of Chlamydomonas reinhardtii. Proc Natl Acad Sci USA. 1990, 87: 1228-1232. 10.1073/pnas.87.3.1228.
Schnell RA, Lefebvre PA: Isolation of the Chlamydomonas regulatory gene NIT2 by transposon tagging. Genetics. 1993, 134: 737-747.
Laemmli UK: Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature. 1970, 227: 680-685. 10.1038/227680a0.
This work was supported by grants of the Italian Ministry of Agriculture (project Hydrobio) to GG and National Institutes of Health (GM014642) to JLR.
PF performed experiments. DRD and GG supervised experiments. All authors designed experiments, interpreted data and wrote the manuscript.