Intimin-Mediated Export of Passenger Proteins Requires Maintenance of a Translocation-Competent Conformation

E. coli strains.The E. coli strains used in this study are listed in Table 1. Bacteria were grown at 37°C, except where otherwise noted. The medium used for the growth and maintenance of E. coli strains was dYT (1% [wt/vol] yeast extract, 1.6% [wt/vol] Bacto Tryptone, 0.5% [wt/vol] NaCl). Anhydrotetracycline (Acros, Morris Plains, N.J.) was added to the liquid medium at a final concentration of 0.2 μg/ml. Other drugs were used at the following final concentrations: chloramphenicol, 25 μg/ml; kanamycin, 75 μg/ml; ampicillin, 100 μg/ml; and tetracycline, 10 μg/ml.

Reagents.Restriction enzymes and DNA-modifying enzymes were purchased from MBI Fermentas (Vilnius, Lithuania) or New England BioLabs (Beverly, Mass.). Tfl polymerase was obtained from Promega (Madison, Wis.), and biotinylated goat anti-mouse antibody was obtained from Sigma (St. Louis, Mo.). A monoclonal antibody (VII-E-7) against a 13-residue C-terminal epitope (Send epitope) of Sendai virus L protein (DGSLGDIEPYDSS) (12) was a gift from H. Einberger and H. P. Hofschneider (Max-Planck-Institut für Biochemie, Martinsried bei München, Federal Republic of Germany). An anti-E epitope monoclonal antibody was obtained from Amersham Pharmacia Biotech (Uppsala, Sweden). Streptavidin R-phycoerythrin (PE) conjugate was purchased from Molecular Probes (Eugene, Oreg.). All other chemicals were of analytical grade and were obtained from Sigma.

Oligonucleotides.Synthetic oligonucleotides were purchased from SigmaARK (Darmstadt, Federal Republic of Germany) or MWG Biotech (Ebersberg, Federal Republic of Germany). The sequences were as follows: BlaSmaUp, 5′-GCGCCCCGGGCACCCAGAAACGCTGGTG-3′; Bla-Bgl12dwn, 5′-GCCCAGATCTCCAATGCTTAATCAGTGAG-3′; BLIP-up, 5′-GCGCGCCCGGGGCGGGGGTGATGACCGGGG-3′; BLIP-low, 5′-GCGCGGATCCTTATGATGAATCGTATGGTTCGATATCACCTAATGATCCATCTACAAGGTCCCACTGCCG-3′; Calmodulin-SmaI-up, 5′-GCGCCCCGGGATGGCTGATCAGCTGACCGAAGAAC-3′; Calmodulin-BglII-low, 5′-GCGCAGATCTCTTTGCAGTCATCATCTGTACAAAC-3′; UbiSmaUp, 5′-GCGCCCCGGGATGCAGATTTTCGTGAAAACCCTT-3′; UbiBglIILow, 5′-GCGCAGATCTACCACCACGAAGTCTCAACACAAGA-3′; and BlaSigNdeUp, 5′-GAACATATGCACCCAGAAACGCTGG-3′.

DNA procedures.Standard DNA procedures, such as plasmid isolation, ligation, and restriction analysis and isolation of DNA fragments, were carried out as described previously (39). PCR with Tfl polymerase was carried out as follows: 30 s of denaturation at 94°C, 30 s of annealing at 53°C, and 30 s of elongation at 72°C for 30 cycles.

Construction of plasmids.pASKInt100-ΔP was constructed by cleavage of pASKInt100-IL-4 (47) with SmaI and BglII, followed by religation of the vector after filling in of DNA ends with T4 DNA polymerase. To obtain vector pASKInt100-Bla, the gene for RTEM-1 β-lactamase from E. coli was amplified by PCR with oligonucleotides BlaSmaUp and Bla-Bgl12dwn and with pHK5-20 (24) as the template DNA. The PCR product was digested with SmaI and BglII and ligated to similarly digested pASKInt100-TmDegP, a derivative of pASKInt100 (47), to yield vector pASKInt100-Bla. The β-lactamase inhibitor protein (BLIP)-encoding gene sequence was amplified by PCR with a BLIP-containing vector (kindly provided by N. C. J. Strynadka) as a template and with primers BLIP-up and BLIP-low. The resulting PCR product was digested with SmaI and BamHI and ligated to similarly digested pASKInt100-EETI-CK^Send (47). For the construction of a C-terminally truncated intimin-calmodulin gene fusion vector (pASKInt100-Cal), pASKInt100-TmDegP was digested with SmaI and BglII and ligated to a similarly digested PCR product encompassing the calmodulin gene, which was obtained from PCR amplification with a human heart cDNA library as a template and oligonucleotide primers Calmodulin-SmaI-up and Calmodulin-BglII-low. Finally, to obtain plasmid pASKInt100-Ubi, the ubiquitin-encoding gene sequence was amplified by PCR with oligonucleotides UbiSmaUp and UbiBglIILow and with human heart cDNA as a template. The resulting product was digested with SmaI and BglII and ligated to similarly digested pASKInt100-EETI-CK^Send (47).

Preparation of an E. coli membrane fraction.Cultures of E. coli strains containing the respective expression plasmids were grown overnight and subcultured 1:50 until they reached an optical density at 600 nm (OD₆₀₀) of 0.8. After induction with anhydrotetracycline (0.2 μg/ml) for 180 min, cells (50 ml) were pelleted by centrifugation and resuspended in 100 mM Tris-Cl (pH 8.0). The membrane fraction was prepared as described previously (16), with minor modifications. Cells were lysed by sonication with a Branson (Danbury, Conn.) Sonifier. Remaining large bacterial fragments were sedimented by centrifugation at 5,000 × g for 10 min. After incubation of the lysate on ice for 30 min in 100 mM Tris-HCl (pH 8.0)-10 mM EDTA-1% (wt/vol) Triton X-100, the membrane fraction was isolated by centrifugation of the cleared solution at 100,000 × g for 120 min at 15°C. Membranes were solubilized in sample buffer (50 mM Tris-HCl [pH 6.8], 100 mM dithiothreitol, 2% [wt/vol] sodium dodecyl sulfate [SDS], 0.1% [wt/vol] bromophenol blue, 10% [vol/vol] glycerol) and subjected to SDS-polyacrylamide gel electrophoresis (PAGE) (12.5% acrylamide-bisacrylamide [30:0.8]) followed by immunoblotting.

Trypsin treatment of intact cells and preparation of periplasmic proteins.Cells were grown in dYT and treated with trypsin as previously described (47). After being washed, cells were resuspended in 200 mM Tris-Cl (pH 9.0)-100 mM EDTA-20% (wt/vol) sucrose and incubated on ice for 1 h. Cells were pelleted by centrifugation and subjected to osmotic shock by resuspension in 10 mM Tris-Cl (pH 9.0). Bacterial fragments were removed by centrifugation at 13,000 × g for 30 min. The cleared osmotic shock fluid was subjected to SDS-PAGE (12.5% acrylamide-bisacrylamide [30:0.8]) followed by immunoblotting.

Flow cytometric analysis.For flow cytometric analysis, cultures of E. coli strains containing the respective expression plasmids were grown overnight and subcultured 1:50 until they reached an OD₆₀₀ of 0.2. After induction with anhydrotetracycline (0.2 μg/ml) for 60 min, cells (200 to 500 μl) were pelleted by centrifugation in a tabletop centrifuge for 1 min and resuspended in 10 μl of phosphate-buffered saline (PBS). After the addition of 1 μl of the respective antibodies (1 mg/ml), cells were incubated for 5 min at room temperature. After the addition of 500 μl of PBS, cells were centrifuged for 1 min and resuspended in 10 μl of PBS containing biotinylated goat anti-mouse immunoglobulin (1:10 dilution). After 5 min of incubation at room temperature and the addition of 500 μl of PBS, cells were pelleted as described above and resuspended in 10 μl of PBS containing Streptavidin R-PE conjugate (1:10 dilution). Finally, after 5 min of incubation at room temperature and the addition of 500 μl of PBS, cells were pelleted as described above and resuspended in 100 μl of PBS for flow cytometric analysis. A total of 300,000 events were collected with a Cytomation MoFlo cell sorter. Parameters were set as follows: forward scatter and side scatter—730 (LIN mode, amplification factor 6); FL1 (fluorescein isothiocyanate)—600 (LOG mode); FL2 (PE)—600 (LOG mode); and trigger parameter—side scatter. The sample flow rate was adjusted to an event rate of approximately 30,000 s⁻¹.

Binding of MykBla^Send to cell surface-exposed calmodulin.Cells harboring pASK100-Cal were grown overnight and subcultured 1:50 in dYT containing 20 mM EGTA until they reached an OD₆₀₀ of 0.2. After the induction of gene expression with anhydrotetracycline (0.2 μg/ml) for 60 min, cells (200 to 500 μl) were pelleted by centrifugation in a tabletop centrifuge for 1 min, resuspended in 500 μl of Tris-Cl (pH 8.0)-50 mM CaCl₂, and incubated on ice for 30 min. Cells were pelleted and resuspended in 10 μl of 100 mM borate buffer (pH 7.0) containing MykBla^Send fusion protein (350 μg/ml). MykBla^Send is a tripartite fusion protein consisting of a calmodulin-binding segment (MKRRWKKNFIAVSAANRFKKISSSGAL) of human light-chain myosin kinase, RTEM β-lactamase, and a Send epitope. After 30 min of incubation on ice, cells were successively incubated with anti-Send epitope antibody, biotinylated goat anti-mouse antibody, and Streptavidin R-PE conjugate as described above. Finally, cells were resuspended in 20 μl of PBS for fluorescence microscopy by using a Zeiss (Göttingen, Federal Republic of Germany) Axioscope with Zeiss filter set 487715.

Generation of Bla mutants. E. coli strain 71-18mutS (26) was successively transformed with plasmids pASKInt100-Bla and pZA22-mutD5 (H. Kolmar, unpublished results), a derivative of low-copy-number plasmid pZA22-MCS1 (31) containing the mutD5 gene under P_lac promoter control. mutD5 encodes a dominant-negative variant of the DNA polymerase III ε subunit lacking proofreading activity (49). Cells were grown overnight in 50 ml of dYT supplemented with chloramphenicol, kanamycin, and 1 mM isopropyl-β-d-thiogalactopyranoside (IPTG) at 37°C in quadruplicate. A 50-μl portion of each culture was transferred to 50 ml of fresh medium containing antibiotics and the inducer IPTG, and cells were grown overnight. This procedure was repeated two more times. We serendipitously found that the overproduction of wild-type Bla fused to C-terminally truncated intimin in the presence of 0.6% (wt/vol) SDS in the growth medium results in cell lysis.

To eliminate clones expressing a C-terminally truncated intimin-wild-type Bla fusion protein, 250-μl samples from the cultures were pooled and used to inoculate 50 ml of dYT supplemented with the appropriate antibiotics, 0.6% (wt/vol) SDS, 1 mM EDTA, and 15 mM 2-mercaptoethanol (2-ME). At an OD₆₀₀ of 0.2, the expression of the C-terminally truncated intimin-Bla fusion was induced by the addition of tetracyline to a 10-ml portion of the culture. Cells were grown for 2 h, and then 5 ml of this culture was used to inoculate 50 ml of fresh dYT supplemented with the appropriate antibiotics, 0.6% (wt/vol) SDS, 1 mM EDTA, and 15 mM 2-ME. Cells were grown at 30°C overnight and subcultured 1:50 ml in fresh dYT supplemented with the appropriate antibiotics, 0.6% (wt/vol) SDS, 1 mM EDTA, and 15 mM 2-ME until they reached an OD₆₀₀ of 0.2. The expression of the C-terminally truncated intimin-Bla fusion was induced by the addition of anhydrotetracyline. Cells were grown for 1 h, successively labeled with anti-Send epitope antibody, biotinylated anti-mouse antibody, and Streptavidin R-PE conjugate, and subjected to fluorescence-activated cell sorting (FACS).

A total of 150,000 cells that fell within a window of 15 to 1,333 relative fluorescence units out of a total of 10⁵ units were sorted out, plated on dYT plates containing 10 μg of tetracycline/ml and 20 μg of ampicillin/ml to counterselect bla mutants containing nonsense or frameshift mutations, and grown overnight. Cells then were scraped off the plates and used for two additional rounds of identical labeling and sorting. Finally, single fluorescent clones were spotted onto agar plates by using the single-cell deposition unit of the MoFlo cell sorter. Plasmid DNA was isolated from individual clones and subjected to nucleotide sequence analysis.

Purification of β-lactamase variant proteins.The coding sequences for the Bla-C98Y variant and the wild-type β-lactamase were amplified by PCR with oligonucleotides BlaSigNdeUp and Bla-Bgl12dwn and the respective pASKInt100-Bla plasmid as a template. The PCR product was cloned into pCR4TOPO (Invitrogen, Carlsbad, Calif.). The resulting vector was digested with NdeI and BglII, and the bla-containing fragment was cloned into similarly digested expression vector pBBR22bII, a derivative of pBBR22b (38), yielding plasmids pBBR22bII-Blawt and pBBR22bII-BlaC98Y. BL21(DE3) cells were transformed with the respective pBBR22bII-Bla plasmid, and a single colony was used to inoculate 1 liter of dYT containing the appropriate antibiotic. The culture was grown at 15°C, and the expression of the Bla variant, which is under T7 promoter control, was induced at an OD₆₀₀ of 0.6 by the addition of 1 mM IPTG. The β-lactamase variant proteins were purified by metal chelate affinity chromatography (9).

Measurement of Bla activity.Wild-type Bla and Bla-C98Y were diluted to equimolar concentrations in 50 mM phosphate (pH 8.0)-50 mM NaCl containing either no or 1 M urea. The hydrolysis of the chromogenic substrate PADAC (Calbiochem, San Diego, Calif.) was monitored colorimetrically at 450 nm. The initial velocity of the reaction was measured and expressed as relative enzymatic activity, where the activity of wild-type Bla without the addition of urea was set to 100.

Production of C-terminally truncated intimin passenger fusion proteins in E. coli K-12.Figure 1C shows the organization of plasmid pASKInt100, which was constructed for the cell surface exposure of various passenger domains via fusion to C-terminally truncated intimin. pASKInt100 contains the coding sequence for C-terminally truncated intimin, which lacks the three carboxy-terminal domains of intimin, and a passenger domain, which is flanked by two distinct antigenic tags, an E epitope within the linker region connecting C-terminally truncated intimin and the passenger domain and a C-terminal Send epitope (12). The expression of the gene construct is under tetA promoter-operator control so that the system is inducible with anhydrotetracycline (43). The passenger domains used in this work are listed in Table 2.

To examine whether these fusion proteins accumulate in the outer membrane as full-length proteins and to assess susceptibilities to proteolysis, membrane fractions of E. coli 71-18 cells carrying the respective expression plasmids were analyzed by SDS-PAGE and immunoblotting (Fig. 2). Each fusion protein could be detected in the membrane fraction by Western blot analysis with a monoclonal antibody that recognizes the C-terminal Send epitope, albeit with various yields of net accumulation. This experiment also revealed significant proteolysis of the hybrid proteins, which was not altered in the presence of 2-ME or EGTA. The same results were obtained with a monoclonal antibody to the internal E epitope (data not shown). Since the antibody used in the Western blot analysis recognizes the carboxy-terminal epitope of the fusion protein, the proteolysis products correspond to fusion proteins with a truncated amino-terminal periplasmic domain of intimin.

• Open in new tab
• Download powerpoint

FIG. 2.

Western blot analysis with anti-Send epitope antibody of a whole-membrane preparation of induced recombinant 71-18 cellsharboring pASKInt100-REI, pASKInt100-Ubi, pASKInt100-Bla,pASKInt100-BLIP, pASKInt100-Cal, or pASKInt100-ΔP and grown in the presence or absence of 20 mM 2-ME or 20 mM EGTA. M, marker proteins (pencil marked after Ponceau S staining); sizes (in thousands) are indicated. Arrowheads indicate full-length proteins.

Translocation of passenger domains to the E. coli cell surface and localization studies.To investigate whether the C-terminally truncated intimin domain mediated the translocation of the passenger domains through the outer membrane, mid-exponential-growth-phase cultures of E. coli 71-18 cells expressing the respective fusion proteins were probed for display of the passenger domains by immunofluorescence staining with anti-E epitope and anti-Send epitope monoclonal antibodies, which recognize epitopes flanking the respective passenger domains. Both antibodies are specific for their respective tags and do not recognize C-terminally truncated intimin. As a control, the cell surface display of a C-terminally truncated intimin fusion protein lacking a passenger domain was investigated with 71-18(pASKInt100-ΔP) cells, in which the E epitope is directly fused to the Send epitope. Labeled cells were analyzed by FACS.

As shown in Fig. 3, all fusion protein-producing cells were immunofluorescently stained with the anti-E epitope antibody, while the extent of immunofluorescence obtained with the anti-Send epitope antibody, recognizing the C-terminal epitope of the fusion protein, varied greatly. Both epitope tags flanking ubiquitin and—in accordance with previous results (47)—the immunoglobulin light-chain variable domain REI_v were detected, indicating the cell surface exposure of these passenger proteins. No immunofluorescence was detected upon cell staining with an antibody directed against a Send epitope carboxy terminal to the calmodulin, BLIP, or β-lactamase passenger domain. These data indicate that fusion of the passenger domains to C-terminally truncated intimin prevented neither the outer membrane localization of the C-terminally truncated intimin core domain nor the cell surface exposure of the immunoglobulin-like C-terminally truncated intimin extracellular domain, while the outer membrane translocation of the calmodulin, BLIP, and β-lactamase passenger domains failed.

• Open in new tab
• Download powerpoint

FIG. 3.

C-terminally truncated intimin-mediated cell surface display of passenger proteins. FACS histograms of E. coli 71-18 cells harboring no plasmid (A), pASKInt100-ΔP (B), pASKInt100-REI (C), pASKInt100-Ubi (D), pASKInt100-Bla (E), pASKInt100-Cal (F), or pASKInt100-BLIP (G) are shown. Induced cells were incubated with anti-Send epitope antibody (S) or anti-E epitope antibody (E), biotinylated anti-mouse antibody, and streptavidin-R-PE conjugate. Unlabeled 71-18 cells served as a control (−) in panels B to D.

Since an anti-E epitope immunofluorescent label was detected for all fusion proteins investigated, it can be assumed that the Cal, Bla, and BLIP passenger domains are trapped in the periplasmic space or the outer membrane. In order to test this assumption, cells expressing the REI_v, Bla, and Cal fusion proteins were incubated with trypsin prior to being stained with anti-E epitope antibody. After trypsin treatment, periplasmic proteins were isolated by osmotic shock treatment, and the cleared osmotic shock fluid was subjected to SDS-PAGE and Western blot analysis with anti-E epitope antibody (Fig. 4B). Only for cells expressing Cal and Bla fusion proteins could bands be detected. Cell permeabilization by the addition of EDTA prior to trypsin treatment resulted in the complete digestion of the periplasmic fragments (data not shown), corroborating the notion that these fragments are susceptible to trypsin cleavage but are shielded from enzyme attack by the presence of an intact outer membrane. The sizes of these bands corresponded to those of fragments that would result from cleavage of the proteins in surface-exposed domain D0 (Fig. 4C). Our findings indicate that these passenger domains reside within the periplasmic space or the outer membrane, are detached from the membrane anchor by trypsin cleavage, and are subsequently found within the periplasm. For REI_v, however, no periplasmic fragment could be detected, in compliance with the finding that REI_v is located on the cell surface and therefore is susceptible to trypsin cleavage.

• Open in new tab
• Download powerpoint

FIG. 4.

(A) Model of fusion of C-terminally truncated intimin with a trapped passenger domain. Truncated intimin (intimin′; broken lines) is shown as a dimer, where the left extracellular domain is omitted for the sake of clarity. Domain D0 and the E epitope are located on the cell surface, whereas the passenger and the Send epitope are trapped within the periplasm. Residue 550 is the last amino acid of the transmembrane region, and amino acid 659 is the last amino acid of truncated intimin. Dimensions are not proportional. (B) Western blot analysis with anti-E epitope antibody of periplasmic protein preparations from induced recombinant 71-18 cells that harbored pASKInt100-REI, pASKInt100-Cal, or pASKInt100-Bla and that were treated with trypsin prior to osmotic shock. M, marker proteins (pencil marked after Ponceau S staining); sizes (in thousands) are indicated. (C) Linear representation of truncated intimin-Bla (top) and truncated intimin-Cal (bottom) fusion proteins. The putative trypsin cleavage site, as deduced from panel B, is indicated by scissors. tm, transmembrane region; E, E epitope; S, Send epitope.

Klauser et al. showed that cell surface exposure of CtxB—a subunit of cholera toxin which contains a single intramolecular disulfide bond—fused to the N. gonorrhoeae IgA protease autotransporter domain was observed only when disulfide bond formation of CtxB in the E. coli periplasm prior to translocation was interfered with either by the addition of 2-ME to the growth medium or by the use of an E. coli dsbA mutant lacking a major periplasmic oxidoreductase which promotes periplasmic disulfide bond formation (23). In order to investigate the influence of periplasmic disulfide bond formation on the cell surface exposure of passenger domains fused to C-terminally truncated intimin, strains 71-18 and 71-18dsbA harboring the respective pASKInt100 derivatives were grown in the absence or presence of 20 mM 2-ME, induced with anhydrotetracycline, and analyzed for cell surface exposure of the respective passenger domains by immunofluorescence staining as described above. With the exception of BLIP (Fig. 5), none of the passenger domains displayed elevated levels of cell surface exposure (data not shown). The addition of 20 mM 2-ME to the growth medium of strain 71-18(pASKInt100-BLIP) resulted in a slight improvement in the extracellular exposure of BLIP, a protein that contains two intramolecular disulfide bonds. The same results were obtained when 1 mM tris(2-carboxyethyl)phosphine (TCEP) was used as a reducing agent (data not shown). The use of 71-18dsbA as an expression host resulted in a further increase in the number of surface-exposed BLIP molecules per cell, and with the addition of 20 mM 2-ME, the surface display of BLIP was substantially improved (Fig. 5).

• Open in new tab
• Download powerpoint

FIG. 5.

C-terminally truncated intimin-based cell surface display of BLIP. A FACS histogram of recombinant 71-18 or 71-18dsbA cells harboring pASKInt100-BLIP is shown. Induced cells were harvested and successively incubated with anti-Send epitope antibody or anti-E epitope antibody, biotinylated anti-mouse antibody, and streptavidin R-PE conjugate. I^S, 71-18(pASKInt100-BLIP) cells grown in dYT and labeled with anti-Send epitope antibody. II^S, 71-18(pASKInt100-BLIP) cells grown in dYT containing 20 mM 2-ME and labeled with anti-Send epitope antibody. III^S, 71-18dsbA(pASKInt100-BLIP) cells grown in dYT and labeled with anti-Send epitope antibody. IV^S, 71-18dsbA(pASKInt100-BLIP) cells grown in dYT containing 20 mM 2-ME and labeled with anti-Send epitope antibody; V^E, 71-18(pASKInt100-BLIP) cells grown in dYT and labeled with anti-E epitope antibody. Unlabeled 71-18 cells served as a control (−).

Export of calmodulin.The results obtained with BLIP could be interpreted as indicating that there is a correlation between export and prevention of periplasmic folding of the passenger domain. To test this supposition, we chose as a passenger domain calmodulin (Cal), which does not contain disulfide bonds but whose folding stability is strongly dependent on the presence of calcium ions. The apo form of calmodulin is significantly unfolded at normal temperatures, while the calcium-loaded form of calmodulin has been found to be exceptionally stable (17), because it can be exposed to temperatures of >90°C or to a 9 M urea solution without a marked change in its tertiary structure. It is important to note that the secretion of calmodulin through the cytoplasmic membrane is unaffected by the presence or absence of calcium ions (33). This property allowed us to probe the influence of periplasmic folding of the calmodulin passenger domain on surface display by varying the periplasmic concentrations of Ca²⁺ ions.

For this experiment, E. coli cells containing pASKInt100-Cal were grown in the presence or absence of 20 mM EDTA. As shown in Fig. 2, the expression of a full-length C-terminally truncated intimin-Cal fusion protein could be detected in the presence or absence of EGTA, a result which could be verified by immunofluorescence staining with an anti-E epitope antibody; however, calmodulin surface display was observed only in the presence of 20 mM EDTA. The same results were obtained with EGTA, a chelator that is more Ca²⁺ specific, or by growing cells in M9 mininal medium supplemented with 20 mM EDTA (data not shown). Furthermore, when cells that produced the C-terminally truncated intimin-Cal fusion protein in the presence of 20 mM EGTA were washed and resuspended in buffer containing 50 mM calcium chloride, the surface-exposed calmodulin regained its native fold and was able to bind the natural substrate molecule Myk (20), a short peptide from light-chain myosin kinase, since calmodulin-presenting cells were immunofluorescently stained with a Myk-Bla^Send fusion protein (Fig. 6D). No enhancement of cell surface net accumulation was observed upon cell growth in the presence of 20 mM EDTA for any of the other passenger domains (data not shown), supporting the theory that the successful export of calmodulin in the absence of calcium ions is due to the low folding stability of apo-calmodulin rather than to the destabilization of the outer membrane by chelating compounds.

• Open in new tab
• Download powerpoint

FIG. 6.

C-terminally truncated intimin-mediated cell surface display of human calmodulin. (A and B) E. coli 71-18 cells harboring pASKInt100-Cal were grown in the absence (A) or presence (B) of 20 mM EDTA. Induced cells were harvested, successively incubated with anti-Send epitope antibody (S) or anti-E epitope antibody (E), biotinylated anti-mouse antibody, and streptavidin-R-PE conjugate, and analyzed by FACS. Unlabeled 71-18 cells served as a control (−). (C and D) Binding of an MykBla^Send fusion protein to surface-exposed calmodulin. Induced cells grown in the absence (C) or presence (D) of 20 mM EGTA were washed, resuspended in 50 mM CaCl₂, and incubated with MykBla^Send, consisting of a calmodulin-binding segment of myosin kinase fused to β-lactamase and tagged with a Send epitope. Binding of the fusion protein to calmodulin was detected by fluorescence microscopy with anti-Send epitope antibody, biotinylated anti-mouse antibody, and streptavidin-R-PE conjugate for cell staining.

Surface display of β-lactamase variants.β-Lactamase (Bla) has been used as a reference passenger domain for surface display based on the IgA protease (Kolmar, unpublished) and AIDA (28) autotransporter domains or a shortened OmpA porin (15). Surface display of Bla fused to the C-terminally truncated intimin translocator, however, was unsuccessful under all conditions tested. As a working hypothesis, we assumed that periplasmic folding and/or disulfide bond formation might prevent the outer membrane translocation of Bla. In this situation, it might be possible to isolate Bla variants that have folding characteristics different from those of the wild-type enzyme and that are able to pass through the outer membrane barrier when fused to C-terminally truncated intimin.

In a search for such Bla variants, we randomly mutagenized pASKInt100-Bla by propagating the plasmid in strain 71-18 mutS(pZA22-mutD5). This hypermutator strain is deficient in DNA mismatch repair because it carries an insertion of Tn10 (tetracycline resistance) in the mutS locus. In addition, it contains the low-copy-number plasmid pZA22-mutD5, carrying the dnaQ gene and a mutD5 mutation encoding a dominant-negative ε subunit of the proofreading exonuclease of DNA polymerase III under lac promoter control (42). Cell surface-exposed Bla variants were isolated from the mutant cell population by cell staining with mouse anti-Send epitope antibody and FACS with the single-cell deposition unit of the MoFlo cell sorter. After verification of the surface display of Bla by immunostaining with anti-Send epitope antibody (Fig. 7), the complete bla genes of seven clones were sequenced. All clones carried a mutation at either codon 52 or codon 98, which encodes a cysteine residue in wild-type Bla (Table 3). Notably, we found the cysteine residues exchanged with comparably bulky arginine and tyrosine residues, a finding which may mean that apart from eliminating disulfide bonds, further destabilization of the protein fold may be required for Bla surface display.

• Open in new tab
• Download powerpoint

FIG. 7.

C-terminally truncated intimin-based cell surface display of β-lactamase variants (Bla). A FACS histogram of recombinant 71-18 cells harboring wild-type plasmid pASKInt100-Bla (I) or the C52R (II), C98R (III), or C98Y (IV) variant is shown. Induced cells were harvested and successively incubated with anti-Send epitope antibody, biotinylated anti-mouse antibody, and streptavidin-R-PE conjugate.

E. coli cells exposing the Bla variants on their cell surface displayed markedly reduced Bla activity and were able to grow only at ampicillin concentrations of <100 μg/ml. Wild-type Bla and the variant Bla-C98Y were further characterized by cloning of the respective genes into expression vector pBBR22bII, which allows the expression of β-lactamase with six additional carboxy-terminal histidine residues under T7 promoter control. Purified Bla-C98Y displayed markedly reduced β-lactamase activity on the chromogenic substrate PADAC (Table 3). Furthermore, while the relative enzymatic activity of wild-type β-lactamase was reduced to 80% in the presence of 1 M urea, the activity of Bla-C98Y was reduced to 10% under these conditions; these results indicate drastically reduced stability of Bla-C98Y compared to that of wild-type Bla. Taken together, these data suggest that β-lactamase can be translocated across the bacterial outer membrane via fusion to C-terminally truncated intimin only under conditions in which folding and disulfide bond formation of the β-lactamase passenger domain in the periplasm are prevented.