Display of Recombinant Proteins on Bacillus subtilis Spores, Using a Coat-Associated Enzyme as the Carrier

Bacterial strains, media, and general techniques.The B. subtilis strains used in the present study are listed in Table 1 . They are derived from strain PY79 (1A747; Bacillus Genetic Stock Center). E. coli K-12 was used to PCR amplify the uidA gene, encoding for β-glucuronidase (GusA). The high-fidelity Herculase polymerase (Stratagene) was used to generate PCR products. Tryptone blood agar broth (BD Difco) was used as the standard solid medium for routine propagation of all B. subtilis strains. Difco sporulation medium (DSM) was used to induce sporulation by nutrient exhaustion (32).

Construction of B. subtilis strains expressing CotG-Phy OxdD-Phy and OxdD-GusA fusion proteins.A schematic view of the cloning strategy used for the translational fusions is presented in Fig. 1B. The original promoter of the spore-anchoring motif (300 bp of DNA upstream from oxdD start codon, or 465 bp of DNA upstream cotG start codon) was used to synchronize expression with the passenger protein during sporulation. Passenger proteins were fused in frame with the carboxyl terminus of the anchoring motif (either OxdD or CotG), deleted of its stop codon. B. subtilis native phytase Phy (GenBank accession number BG11198), was fused to either OxdD (BG13484) or CotG (BG11017) (numbering according to Genolist, Institut Pasteur, 1999-2001; Subtilist [http://genolist.pasteur.fr/SubtiList ], and Colibri [http://genolist.pasteur.fr/Colibri]) . The region encoding the signal peptide of Phy (encompassing the 26 first residues; Swiss-Prot P42094) was excluded. A linker made of 10 alanine residues was introduced between the carrier and passenger proteins. Gene fusions cotG-phy (2,173 bp long) and oxdD-phy (2,594 bp long) were synthesized as synthetic genes and sequenced by DNA2.0, Inc. (California). They were then cloned between the BamHI and HindIII sites of pDG364 suicide vector (7), resulting in plasmids pSD21 (cotG-phy) and pSD22 (oxdD-phy). To construct a fusion of the Escherichia coli uidA gene, coding for β-glucuronidase (GusA; GenBank accession number EG11055) to the 3′ end of the oxdD gene, we first introduced a NheI restriction site into the 10-residue linker of pSD22, leading to the replacement of the sixth alanine residue by a serine. Then, the uidA gene of E. coli was cloned between NheI and HindIII sites of plasmid pSD22, to yield pSD27, carrying a 3,339-bp open reading frame coding for a OxdD-GusA fusion protein under the control of the native oxdD promoter. After linearization with XhoI restriction endonuclease, pSD21, pSD22, and pSD27 were used to independently transform B. subtilis PY79 with selection for chloramphenicol (5 μg/ml). Transformants that resulted from a marker replacement recombination event (double crossover) at the nonessential amyE locus were identified and designated SD48 (expressing the CotG-Phy fusion), SD50 (OxdD-Phy), and SD60 (OxdD-GusA).

Construction of SD58 (E. coli His₆-Phy, N terminus).A 1.1-kb PCR fragment containing the phy gene without its first 84 codons (corresponding to the signaling peptide) was amplified by using primers phy-for 5′-GGAATTCCATATGGTGAATGAGGAACATCATTTCAAAG-3′ and phy-rev 5′-ATCGCTCGAGGCCGTCAGAACGGTCTTTCAGCTTCCTC-3′, which generated NdeI and XhoI restriction site ends, respectively. Chromosomal DNA of B. subtilis 168 was used for amplification. The PCR fragment was digested and inserted into pET16b (Novagen) to create pSD26. Selection of transformants was done on LB plates containing 100 μg of ampicillin/ml. Plasmid pSD26 carries an N-terminal fusion of B. subtilis phytase (without its signaling peptide) fused to a His tag. After purification, pSD26 was introduced into E. coli BL21 competent cells (Invitrogen) to allow efficient overexpression of the heterologous protein. Selection of transformants was done in LB plates containing 100 μg of ampicillin/ml and 50 μg of chloramphenicol/ml. The resulting strain was named SD58.

Overproduction and purification of B. subtilis phytase.A starter culture of strain SD58, grown overnight at 37°C with shaking, was used to inoculate 100 ml of fresh LB broth supplemented with ampicillin and chloramphenicol, which was incubated at 37°C and 150 rpm until reaching an optical density at 600 nm of ∼0.6. At this point, IPTG (isopropyl-β-d-thiogalactopyranoside) was added to a final concentration of 1 mM, and the culture was incubated for another 4 h. The culture was then centrifuged 20 min at 8,000 × g and 4°C, and the cell pellet was resuspended in 4 ml of ice-cold lysis buffer (20 mM phosphate, 300 mM NaCl, 1 mMCaCl₂, 10 mM imidazole, 1 mM phenylmethylsulfonyl fluoride, 1× anti-protease EDTA free solution). The cell lysate was obtained by passage through a French press cell at 19,000 lb/in², and the supernatant fraction was used to purify the His₆-tagged phytase on a Ni²⁺-nitrilotriacetic acid (NTA) column (Qiagen). The column was operated according to the manufacturer's instructions. Bound phytase was eluted at an imidazole concentration of 300 mM. All purification steps were carried out at 4°C.

Isolation of B. subtilis extracellular proteins.The extracellular proteome was prepared essentially as described by Antelmann et al. (1). Briefly, cultures (20 ml) and its derivative bearing a Δphy mutation (AH7666), were grown in a synthetic medium containing 15 mM (NH₄)₂SO₄, 8 mM MgSO₄·7H₂O, 27 mM KCl, 7 mM sodium citrate·2H₂O, 50 mM Tris-HCl (pH 8) supplemented with 0.16 mM KH₂PO₄, 2 mM CaCl₂·2H₂O, 1 μM FeSO₄·7H₂O, 10 μM MnSO₄·4H₂O, 4.5 mM glutamic acid, 780 μM tryptophan, 860 μM lysine, and 0.2% (wt/vol) glucose. Samples were collected 1 h after the end of the exponential phase of growth by centrifugation for 30 min at 21,000 × g and 4°C. Proteins present on the culture supernatant were then precipitated during 1 h at 4°C with trichloroacetic acid (TCA) at a final concentration of 5% and recovered by centrifugation during 30 min at 21,000 × g and 4°C. The protein pellet was once washed with 70% ice-cold ethanol and finally resuspended in 200 μl of a buffer containing 10 mM Tris (pH 8), 10 mM MgCl₂, 500 mM EDTA, 200 mM NaCl, 10% glycerol, 1 mM PMSF, and 1× anti-protease EDTA free solution.

Construction of AH7666 (B. subtilis wild type with phy deleted).Long flanking homology PCR was used to delete the endogenous copy of phy in PY79 background (38). The primers P1-phy (5′-ATATCTGCCTAAAAAAAGTGC-3′) and P2-phy-spec (5′-ACATGTATTCACGAACGAAAATCGAAATAGAAAGCAGCTTGTGCAGC-3′) were used to amplify a fragment upstream of the phy gene, while the primers P3-phy-spec (5′-ATTTTAGAAAACAATAAACCCTTGCACCTTCATTTGGTACCCTCC-3′) and P4-phy (5′-CACCTGTTTAGGTGTAAGCAG-3′) were used to amplify a fragment downstream of phy. Both PCR products were bridged by a fragment containing the spectinomycin cassette (11). The final PCR product was then used to replace the endogenous phy gene by the spectinomycin cassette via double crossover.

Spore purification and analysis.Spores were harvested by centrifugation of DSM cultures 24 h after the onset of sporulation. After washing, spores were purified with a 20 to 50% Gastrografin (Schering) step gradient, as previously described for Renocal-76 gradients (12). Purified spores were suspended in cold buffers (as specified below) supplemented with protease inhibitors (Complete EDTA-free; Roche). Samples of the purified spore suspension were used to analyze the coat polypeptide composition as described before (12). The total cell count, as well as the titer of heat or lysozyme-resistant spores, was determined in samples of DSM cultures 24 h after the onset of sporulation (12).

Detection of β-glucuronidase (GusA) activity by fluorescence microscopy. In situ detection of β-glucuronidase activity was performed using the fluorogenic substrate C₁₂FDGlcU, a lipophilic analog of fluorescein di-β-d-glucuronic acid containing a 12-carbon aliphatic chain supplied with the ImaGene Green C₁₂FDGlcU GUS gene expression kit from Molecular Probes. Cleavage by β-glucuronidase releases the yellow-colored, green-fluorescent compound 5-dodecanoylaminofluorescein (maximal excitation wavelength = 495 nm, maximal emission wavelength = 518 nm). This substrate was used on purified spores of PY79 and SD60, according to the indications of the manufacturer. Before adding the substrate, part of the spore suspension was washed with water and incubated for 1 h at 37°C in a 0.1% trypsin solution (Amimed), as a control. Spores were washed three times with phosphate-buffered saline (PBS) at pH 7.4 (Roche). The spores were then examined by fluorescence microscopy (Eclipse E600; Nikon) using phase-contrast optics and a standard filter for the visualization of green fluorescence emission (Ex = 490 nm, Em = 520 nm), reflecting GusA activity at the spore surface. Images were acquired with a digital camera (CoolSnap EZ; Visitron Systems) and processed with ImageJ (W. S. Rasband, ImageJ, 1997-2008 [http://rsb.info.nih.gov/ij/ ]; National Institutes of Health, Bethesda, MD) for quantification of the fluorescence signal. Background was measured for each picture and substracted from the fluorescence intensity of the samples. The quantification was made on 150 spores per strain.

Assay of B. subtilis phytase activity.Whole-cell extracts were prepared from 24-h DSM cultures of strains SD48, SD50, and PY79. Cells and mature spores were harvested by centrifugation (20 min, 8,000 × g) and suspended in a buffer containing 100 mM Tris-HCl (pH 7.4; Fluka), a cocktail of protease inhibitors (Complete EDTA-free), and 2 mM CaCl₂ (Sigma), required to maintain active conformation of B. subtilis phytase (24, 36). The mixture was passed twice in French cell press (Sim-Aminco) at 900 lb/in² to release both forespores and the cytoplasm content of the mother cells. Spores were purified from whole-cell extracts as described previously (40). Phytase activity was measured by incubating either 10 μl of whole-cell extracts or 10 μl of pure spores suspensions with 40 μl of 2 mM sodium phytate (Sigma) in the above-mentioned buffer. After 30 min of incubation at 55°C, the reaction was stopped by adding 50 μl of 5% TCA (Merck) (36). After centrifugation (4 min, 8,000 × g), 50 μl of the supernatant was diluted in 500 μl of water. The released orthophosphate (P_i) was photometrically measured at 700 nm by monitoring the production of phosphomolybdate for 15 min at 50°C after the addition of 500 μl of the color reagent. The color reagent was prepared freshly by mixing 1 volume of 2.5% ammonium molybdate (Fluka), 1 volume of 10% sodium ascorbate (Fluka), and 3 volumes of 1 M H₂SO₄ (Fluka) (27). One unit of phytase was defined as the amount of enzyme required to release 1 μmol of P_i from sodium phytate in 1 min. The total amount of protein was assayed with the bicinchoninic acid method (Pierce), in order to calculate specific activity. The dry weight was measured after drying spores suspensions for 24 h at 80°C.

Immunodetection of spore displayed phytase.An anti-phytase rabbit polyclonal antibody was produced by immunizing rabbits with a mix of two synthetic phytase-specific peptides (NH₂-CAEPGGGSKGQVVDRA-COOH and NH₂-CHKQVNPRKLKDRSDG-COOH; Eurogentec). Before immunolabeling, part of the spores was incubated at 37°C for 1 h in a 0.1% trypsin solution (Amimed), as a control. They were washed three times with PBS (pH 7.4; Roche) supplemented with 2% bovine serum albumin (Sigma) and then probed with primary and secondary antibodies. Immunostained spores were mounted on agarose-coated microscope slides and observed by fluorescence, as described above. The fluorescence signal was quantified on 150 spores per strain.

Spore display of B. subtilis phytase.Although green fluorescence protein (GFP) fused to OxdD has been successfully exposed at the surface of B. subtilis spores (6), spore display of a functional enzyme has not been reported. The B. subtilis Phy passenger enzyme, truncated for its 26-residue signal peptide, was translationally fused to the C terminus of the anchoring motif OxdD. A linker, made of 10 aliphatic amino acids, was introduced between the carrier and passenger proteins to minimize any potential steric effect that could disturb the correct folding of either protein (Fig. 1B). Since CotG had been previously described as the carrier for the spore display of recombinant proteins (16, 25, 28), a control fusion CotG-Phy was also designed (Fig. 1B). The translational oxdD-phy and cotG-phy fusions are carried by strains SD50 and SD48, respectively (Table 1).

Since in spores of cotG insertional mutants the normal structural organization of the outer coat is severely disturbed (14), we have chosen to allow expression of the wild-type cotG allele in strains expressing passenger-carrier fusions (strain SD48). A similar approach was used with OxdD as a carrier (strain SD50), since the function of OxdD in spore assembly is not yet fully elucidated (6). Both integrations at the amyE locus of the wild-type strain PY79 locus had no detectable effect on spore viability and resistance to heat and to lysozyme (Table 2). Moreover, examination of the collection of coat polypeptides extracted from spores of strains SD48 or SD50 resulted in a pattern that did not differ from that obtained for the wild-type PY79 spores (data not shown). We concluded that expression of the wild-type and engineered carrier proteins did not interfere in any detectable way on the assembly and function of the spore coat lattice.

To test whether the fusion proteins were presented at the spore surface, we conducted immunofluorescence microscopy experiments. For these studies, we used a rabbit polyclonal anti-Phy antibody (see Materials and Methods). This antibody specifically recognized the purified His₆-Phy protein (Fig. 2A). In addition, it recognized the native Phy protein in extracts prepared from culture supernatants of the wild-type strain PY79, grown under conditions where Phy is known to accumulate (1), and did not react with any protein in extracts prepared in parallel from a phy deletion mutant (AH7666) (Fig. 2B). Spores were collected 24 h after the onset of sporulation, purified on density gradients (see Materials and Methods) and either directly analyzed or treated with trypsin as a control. Significant fluorescence was observed on purified spores harboring the CotG-Phy or OxdD-Phy fusions (Fig. 3A). As expected, no significant fluorescence was observed at the surfaces of spores of the wild-type strain PY79 strain. Importantly, treatment with trypsin resulted in loss of the fluorescence signal. Treatment with trypsin also resulted in the complete loss of fluorescence from spores of a strain (AH2873; Table 1) expressing an oxdD-gfp fusion (data not shown). As an additional control, the primary antibody was omitted to test for specific binding of the secondary antibody coupled with the fluorophore. No fluorescence was observed under these conditions. We then quantified the fluorescence signal for spores of strains SD48 and SD50 in comparison to wild-type PY79 spores. The results show that the average signals on SD48 (85 average units [AU] and SD50 (91 AU) were ∼3-fold higher than the baseline signal observed for PY79 spores (29 AU) (Fig. 3B and 4A). In spite of a similar average intensity for SD48 and SD50, 75% of SD48 spores displayed at least 2-fold more fluorescence than the baseline (>60 AU), whereas 100% did so for SD50. Together, these results suggest that the OxdD-Phy fusion was more efficiently displayed at the spore surface.

• Open in new tab
• Download powerpoint

FIG. 2.

Purification of B. subtilis phytase (Phy) and specificity of an anti-Phy antibody. (A) SDS-PAGE analysis (lanes 1 to 3) and immunoblot analysis (lanes 4 to 6) of extracts of E. coli strain SD58 noninduced (lanes 1 and 4), induced with 1 mM IPTG (lanes 2 and 5), and of the His₆-Phy protein (lanes 3 and 6), partially purified by Ni²⁺ affinity chromatography. The asterisk indicates a cross-reactive species or a degradation product of the His₆-phytase. (B) SDS-PAGE (lanes 1 and 2) and immunoblot analysis (lanes 3 and 4) of the cell culture supernatants of the PY79 wild-type strain of B. subtilis (lanes 1 and 3) and of strain AH7666, bearing a Δphy mutation (lanes 2 and 4).

• Open in new tab
• Download powerpoint

FIG. 3.

Immunofluorescence detection of CotG-Phy and OxdD-Phy protein fusions at the surface of purified spores. FITC, fluorescein isothiocyanate. (A) Phase-contrast microscopy (PC) and fluorescence microscopy (FITC). Scale bars, 2 μm. (B) Fluorescence intensity was quantified on 150 spores. AU, arbitrary units. Purified spores were submitted or not to trypsin proteolysis. noAb, control without primary antibody. Spores were purified from cultures of the following strains: PY79, wild type; SD48, CotG-Phy; and SD50, OxdD-Phy.

• Open in new tab
• Download powerpoint

FIG. 4.

Abundance of the displayed phytase at the spore surface and specific activity of B. subtilis phytase. (A) Immunofluorescence detection of CotG-Phy and OxdD-Phy protein fusions at the surfaces of purified spores (average signal). (B) Assay of whole-cell extracts prepared from B. subtilis cultures in sporulating medium, 24 h after the onset of sporulation. (C) Assay of purified spores from the same cultures. DW, dry weight. Spores were purified from cultures of the following strains: PY79, wild type; SD48, CotG-Phy; SD50, OxdD-Phy. One unit (U) of phytase was defined as the amount of enzyme required to release 1 μmol of inorganic phosphate from sodium phytate in 1 min.

Phytase activity at the surface of B. subtilis spores.Having determined that both CotG and OxdD served as carriers for the display of Phy at the surface of B. subtilis spores, we then wanted to determine whether the passenger protein was presented in an active form. For this purpose, we conducted measurements of Phy specific activity. We first measured Phy activity in whole-cell extracts prepared from samples directly collected from DSM cultures of strains SD48, SD50, and PY79, 24 h after the onset of sporulation (Fig. 4B). Although a specific activity of 3.1 U/mg of protein was obtained for PY79, reflecting the endogenous B. subtilis phytase activity, values of 11.4 U/mg and 6.9 U/mg were obtained for SD48 and SD50, respectively (Fig. 4B). For reference, a partially purified His₆-tagged version of the B. subtilis phytase missing its signal peptide (Fig. 2A) gave a specific activity of 23 U/mg of protein when produced in E. coli (data not shown). The phytase activity detected for SD48 and SD50 extracts over the activity obtained for PY79 cultures clearly demonstrated expression of a functional phytase during sporulation, from the sporulation specific promoters that drive expression of either cotG or oxdD. Next, we sought to determine whether phytase activity could be detected on density gradient-purified spores—in other words, whether the CotG-Phy or OxdD-Phy fusion detected at the spore surface (see above) represented active enzyme (Fig. 4C). Since not all proteins are extractable from the spore coat, units of phytase activity in purified spores were normalized with respect to spore dry weight. Activities of 5.7 × 10³ U/g (dry weight) and of 2.7 × 10³ U/g (dry weight) were obtained for SD48 and SD50, respectively. No activity could be detected for PY79 spores or for spores of strains SD48 and SD50 that were treated with trypsin prior to the assay (data not shown). These results indicate that both the CotG or OxdD carriers resulted in the display of active phytase at the spore surface.

Although the OxdD-Phy fusion protein appeared more abundant than CotG-Phy at the spore surface, as assessed by immunofluorescence (see above), spores of SD50 (OxdD-Phy) showed a 2-fold-decreased specific activity compared to those of SD48 (CotG as the carrier). One possibility is that the OxdD-Phy fusion protein, presumed to be mostly associated with the inner-coat layers, is less accessible to the substrate than is CotG-Phy, thought to be mainly located in the outer coat.

Spore display of active E. coli β-glucuronidase.The results presented above show that the inner-coat protein, OxdD, can be used as a carrier for the display of active, monomeric phytase at the spore surface. We then wanted to test whether OxdD could be used for spore display of a larger, oligomeric enzyme. We tested whether E. coli β-glucuronidase, encoded by the uidA gene, could be presented in active form at the spore surface as a fusion to OxdD. The crystal structure of human GusA has been determined and suggests that the enzyme is likely to function as a 273-kDa homotetramer (3) with the catalytic site formed from a large cleft at the interface of two monomers (20). Presumably, the E. coli enzyme, which shares 50% amino acid identity with the human version, is also a homotetramer (10). To attempt the functional display of E. coli GusA at the surface of B. subtilis spores, the uidA gene was fused to the 3′-end of the oxdD gene with, as for the oxdD-phy fusion (see above), an interspacing 10 amino acid-linker (Fig. 1B). Ectopic integration of the gene fusion at amyE resulted in strain SD60 (Table 1). Expression of the OxdD-GusA fusion at the amyE locus of PY79 did not measurably interfere with spore function or assembly (Table 2 and data not shown).

Spores of strain SD60 were collected 24 h after the onset of sporulation in DSM, purified, and assayed for in situ detection of β-glucuronidase activity using the fluorogenic substrate C₁₂FDGlcU, (see Materials and Methods). In parallel, activity assays were conducted in trypsin-treated SD60 spores, and in spores of the wild-type strain PY79. Cleavage of the colorless C₁₂FDGlcU substrate into-5-dodecanoylaminofluorescein results in the production of a stable green-fluorescent product, allowing β-glucuronidase activity to be monitored by fluorescence microscopy. Fluorescence was emitted by SD60 spores (Fig. 5 A). In contrast, no fluorescence was detected for PY79 spores, as well as for SD60 or PY79 spores treated with trypsin prior to the assay (Fig. 5A). Quantification of the fluorescence signal (Fig. 5B and C) showed that 40% of SD60 spores emitted at least 2-fold more fluorescence than the baseline (>60 AU). Possible explanations for the occurrence of nonfluorescent spores might be that the fluorogenic product of cleavage diffuses away from spores or that the high molecular weight (882 g/mol) and lipophilic nature (12-carbon aliphatic chain) of the fluorogenic substrate limits its penetration into the inner coat, where OxdD-GusA presumably resides. It can also be speculated that OxdD-GusA fusion has not assembled into an active form in nonfluorescent spores. In any event, our results show that active GusA can be displayed at the spore surface as a fusion to the inner-coat protein OxdD.

• Open in new tab
• Download powerpoint

FIG. 5.

Assay for the activity of GusA at the spore surface, through conversion of the colorless substrate C₁₂FDGlcU (see Materials and Methods) into a yellow fluorescent product. (A) Detection of a functional OxdD-GusA fusion by phase-contrast microscopy (PC) and fluorescence (FITC) microscopy. Bars, 2 μm. (B) The fluorescence intensity was quantified on 150 spores. Purified spores submitted or not to trypsin proteolysis. Spores were purified from cultures of the following strains: PY79, wild type; and SD60, OxdD-GusA. (C) Average fluorescence intensity quantified on 150 spores. AU, arbitrary units.

Concluding remarks.Further work will address the optimization of strain engineering to increase activity of the passenger enzyme fused to OxdD. For instance, improvement might result from increased multicopy-driven expression, from expression in protease-deficient B. subtilis strains or through the selection of passenger enzymes with higher specific activity (e.g., phytase from Citrobacter braakii). Nevertheless, based on the current activity measured at the surface of spores displaying B. subtilis OxdD-Phy (2.7 × 10³ U/g [dry weight]), ∼0.4 g of spores (dry weight) would represent 1,000 U of phytase. This amount of enzyme, added to 1 kg of feed in a daily swine diet, would replace 1 g of inorganic phosphorus supplementation and therefore could reduce total phosphorus excretion by 30 to 50% (29).

Based on the results presented here, spore-associated enzymes, such as OxdD, seem to be valuable carriers for spore display of passenger proteins. One question raised by our study is whether exposure of the spores to acid or proteases in the upper parts of the gastrointestinal tract would result in the inactivation of the displayed carrier-enzyme fusion protein. This may indeed happen to some extent. However, B. subtilis spores were recently shown to have the ability to germinate and resporulate upon exiting the stomach, leading to persistence of the B. subtilis strain in the gut (2, 31) and to constant surface display of freshly synthesized and fully active enzymes. Spore surface display thus provides the basis for novel means of formulation for proteins of interest such as feed enzymes (phytases, lipases, hemicellulases, and others).