Identification of Promoters for Efficient Gene Expression in Magnetospirillum gryphiswaldense

Abstract

To develop an expression system for the magnetotactic bacterium Magnetospirillum gryphiswaldense, we compared gene expression from the widely used Escherichia coli P_lac promoter with that from known and predicted genuine M. gryphiswaldense promoters. With the use of green fluorescent protein as a reporter, the highest expression level was observed with the magnetosomal P_mamDC promoter. We demonstrate that this promoter can be used for the expression of modified magnetosome proteins to generate “antibody-binding” magnetosomes.

The formation of magnetosomes, which are subcellular compartments consisting of membrane-bounded magnetite (Fe₃O₄) nanocrystals in magnetotactic bacteria, is of interdisciplinary interest in fields of microbial cell biology, biotechnology, and nanotechnology. Despite considerable efforts by many researchers, the genetic analysis of magnetotactic bacteria is still cumbersome and many genetic tools are lacking (7). In this study we developed an expression system for Magnetospirillum gryphiswaldense, which is one of the most widely studied magnetotactic organisms (5, 20).

To identify genuine promoters for gene expression in M. gryphiswaldense, we investigated the expression of green fluorescent protein (GFP) from the Escherichia coli P_lac promoter and from putative and previously identified M. gryphiswaldense promoters such as P_mamDC and P_mamAB, which are highly transcribed under magnetite-inducing conditions, i.e., at microaerobiosis and in the presence of micromolar amounts of iron (19). In addition, genomic regions upstream of large ribosomal gene clusters, which potentially encode the strong P_rpsJ (rpsJ, MGR3815 ribosomal protein S10) and P_rplK (rplK, MGR3801 ribosomal protein L11) promoters, and sequences that putatively encode the P_apdA (P_mms16 in “Magnetospirillum magneticum”) and the P_msp3 promoters, which were described previously in M. magneticum (26), were analyzed.

For the construction of GFP reporter vectors, the egfp gene encoding the GFPmut1 variant (2, 10) was PCR amplified (primers are shown in Table 1), cloned into pGEM-T Easy (Promega), sequenced, and subcloned into pBBR1MCS2 (8) downstream of the P_lac promoter to generate pBBRegfp (Table 2). The vector was digested with NsiI and ApaI, blunted with mung bean nuclease (New England Biolabs), and religated to remove the P_lac promoter and yield the promoterless GFP reporter plasmid pBBRpl (Table 2). Putative promoter regions that included the intergenic region upstream from the start codon to the next open reading frame were PCR amplified from genomic DNA of M. gryphiswaldense R3/S1 (21). The PCR products were cloned into pGEM-T Easy, sequenced, and subcloned into pBBRpl, resulting in the plasmids pBBRPmamDC, pBBRPmamAB, pBBRPmsp3, pBBRPapdA, pBBRPure, pBBRPrplK, and pBBRPrpsJ (Table 2). The plasmids were transferred into M. gryphiswaldense R3/S1 by conjugation from E. coli BW29427 (K. Datsenko and B. L. Wanner, unpublished data) as described previously (18, 22).

TABLE 1.

Primers used in this study^a

Primer name	Sequencea	Restriction site
egfpfw	CATATGGTGAGCAAGGGCGAGGAG	NdeI
CL2	GTGGATCCTTACTTGTACAGCTCGTC	BamHI
papdAfw	CTCGAGGAGCCTCTCCATTAAACAATG	XhoI
papdArev	CATATGCTTGAATTCCTCCAACCGGGGGTATG	NdeI
prplKfw	AAGCTTGGCATCAAGGTTTCGGAAG	HindIII
prplKrev	CATATGTTTACCCTACCTCTGGTCG	NdeI
prpsJfw	CTCGAGGTCCTTCG GGATCGCTTG AC	XhoI
prpsJrev	CATATGATTCACGTCATCCGTTAAATC	NdeI
purerev	CATATGGTGGTTATGCGCTGCTCAAAATC	NdeI
purefw	CTCGAGCTTTTCTCGATCCGGGAAAAATAC	XhoI
PmamDCrev	CATATGCTGATCTCCGGCAAGTGTATG	NdeI
PmamDCfw	CTCGAGCAATGACCACCACCACCTTA AAC	XhoI
PmamABfw	CTCGAGATGGCGCAAAGATGTGACGT C	XhoI
PmamABrev	CATATGTCCCGTCACAATTCACCTCC	NdeI
pmsp3fw	CTCGAGGAACTCCAAAAGCAAGGCTATTTAC	XhoI
pmsp3rev	ATTAATCCGAAAGCTCCTTGAATCAAAAG	VspI
PmamDCxlrev	CTCGAGGATCTCCGGCAAGTGTATGCAC	XhoI
zzfw	GCTGCACATATGGCGCAACACGATGAAGCC	NdeI
zzrev	CCATCTAGAAATATTACCGCCAGCCATTG	XbaI

^aRestriction sites that were incorporated in the primer are indicated in bold.

TABLE 2.

Plasmids used in this study

Plasmid name	Description	Source or reference
pEGFPN1	GFP expression vector; Ap	BD Biotech
pGEM-T Easy	Cloning vector; Ap	Promega
pBBR1MCS-2	Mobilizable broad-host-range vector; Km	8
pGEMegfp	pGEM-T Easy + egfp PCR product	This study
pBBRegfp	pBBR1MCS-2 + egfp from pGEMegfp	This study
pBBRpl	Promoterless GFP reporter vector based on pBBRegfp	This study
pGEMPmamDC	pGEM-T Easy + PmamDC-PCR product	This study
pGEMPmamAB	pGEM-T Easy + PmamAB-PCR product	This study
pGEMPapdA	pGEM-T Easy + PapdA-PCR product	This study
pGEMPmsp3	pGEM-T Easy + Pmsp3-PCR product	This study
pGEMPrplK	pGEM-T Easy + PrplK-PCR product	This study
pGEMPrpsJ	pGEM-T Easy + PrpsJ-PCR product	This study
pBBRPmamDC	pBBRpl + PmamDC from pGEMPmamDC cloned in XhoI and NdeI sites	This study
pBBRPmamAB	pBBRpl + PmamAB from pGEMPmamAB cloned in XhoI and NdeI sites	This study
pBBRPapdA	pBBRpl + PapdA from pGEMPapdA cloned in XhoI and NdeI sites	This study
pBBRPmsp3	pBBRpl + Pmsp3 from pGEMPmsp3 cloned in XhoI and VspI sites	This study
pBBRPrplK	pBBRpl + PrplK from pGEMPrplK cloned in HindIII and NdeI sites	This study
pBBRPrpsJ	pBBRpl + PrpsJ from pGEMPrpsJ cloned in XhoI and NdeI sites	This study
pEZZ18	Protein A gene fusion vector; Ap	GE Healthcare
pGEMZZ	pGEM-T Easy + ZZ protein domain PCR product	This study
pCL6	pBBR1MCS-2 + MamC-GFP	9
pBBRCZZ	The ZZ protein domain from pGEMZZ was cloned into the NdeI and BamHI sites of pCL6 to replace GFP with the ZZ domain	This study
pJETPdcx1	pJET1.2 + PmamDC-PCR product	This study
pBBRPdcCZZ	pBBRCZZ + PmamDC from pGEMPdcx1	This study

For promoter activity assays, M. gryphiswaldense strains expressing GFP from different promoters were cultivated in triplicate microaerobically in 3-ml culture volumes in six-well culture plates under a microoxic atmosphere (1% oxygen, 99% nitrogen) for 20 to 22 h in FSM medium (6). Cells were washed and resuspended in phosphate-buffered saline to an optical density at 565 nm of 0.5. The expression of GFP was quantified from 100-μl aliquots of the cell suspension with an Infinite 500 plate reader (Tecan). The native promoters were considerably more active than the E. coli P_lac promoter that has been used in previous studies with this organism (9, 16-18). Fluorescence quantification and immunoblot analysis (see Fig. S1 in the supplemental material) have shown that the strongest promoter in M. gryphiswaldense was P_mamDC, followed by P_msp3, P_apdA, P_mamAB, P_rpsJ, P_rplK, and P_lac (Fig. 1). Similarly to other alphaproteobacterial promoters (12, 13, 24), the tested M. gryphiswaldense promoters were inactive in E. coli (see Fig. S2 in the supplemental material).

FIG. 1.

Analysis of GFP expression from different promoters in M. gryphiswaldense by fluorometry. The excitation wavelength was 485 nm (20-nm bandwidth), and emission was recorded at 535 nm (25-nm bandwidth). The value for each sample was averaged from 10 reads over an integration period of 20 μs. The error bars reflect the standard deviations calculated from three independent experiments. pl, promoterless control.

Using GFP as a reporter, we estimated gene expression in individual cells by flow cytometry according to a previously described procedure (9). Comparison of the average fluorescence intensities confirmed the bulk measurements (Fig. 1) and showed that cells containing the P_mamDC-GFP construct fluoresced on average twice as much as did cells expressing GFP from P_msp3 or P_apdA (P_mms16) (Fig. 2). In a study of gene expression in M. magneticum, P_msp3 was identified in a luciferase-based assay as the strongest promoter with an activity more than threefold higher than that of the P_mms16 promoter (26). Even though in our study a different reporter and the homologous promoters from M. gryphiswaldense were used, it was unexpected that the activities of P_msp3 and P_apdA were twofold lower than the activity of P_mamDC and that P_msp3 activity was almost identical to P_apdA (P_mms16) activity in M. gryphiswaldense. It is possible that P_msp3, which is a putative peroxiredoxin promoter, was downregulated due to the microaerobic growth conditions, which are required for the production of magnetosomes.

FIG. 2.

Flow cytometry of M. gryphiswaldense (a) and strains expressing GFP from different plasmids: pBBRpl (promoterless) (b), pBBRegfp (P_lac) (c), pBBRPmamDC (d), pBBRPmamH (e), pBBRPmsp3 (f), pBBRPapdA (g), pBBRPrpsJ (h), and pBBRPrplK (i). The proportion of fluorescent cells is shown in bold, and the average fluorescence intensity is displayed in the upper right corner. To estimate the proportion of fluorescent cells, a threshold for fluorescence was set to the fluorescence intensity below which 99% of untransformed M. gryphiswaldense cells, which served as a nonfluorescent standard, were detected. Fifty thousand events were analyzed from each sample.

Flow cytometry also revealed that GFP was not expressed homogeneously in all cells but that a variable fraction of the cells was nonfluorescent (Fig. 2). For the weak and intermediate-strength promoters P_lac, P_rplK, P_rpsJ, and P_mamAB, the proportions of fluorescent cells (from 8.52% to 56.5%) correlated well with the increasing average fluorescence intensities, which indicates that intermediate-strength promoters activate GFP expression in a higher proportion of cells than do weak promoters. In contrast, use of the strong P_msp3, P_mamDC, and P_apdA promoters did not result in a further increased proportion of fluorescent cells (56.3 to 50.3%). It is unclear why a large proportion of cells was inactive with respect to GFP expression. However, inhomogeneous gene expression in an isogenic population of cells can be frequently observed in bacteria (3, 23) and might be caused by cell cycle-dependent effects, stochasticity of gene expression, or variations of growth rates and protein synthesis between individual cells (4, 15, 25). The heterogeneity of gene expression from strong promoters in M. gryphiswaldense is of relevance for the genetic engineering of magnetosomes for biotechnological applications, as the heterogeneity of gene expression will be reflected by heterogeneously modified magnetosomes. However, a proportion of 50 to 60% of cells that express a gene of interest is sufficient for many practical purposes.

To test the applicability of P_mamDC for the expression of fusion proteins in the magnetosome membrane, we expressed a MamC-ZZ fusion protein from P_mamDC. The ezz gene, which codes for the antibody-binding ZZ-protein domain (11), was PCR amplified from pEZZ18 (GE Healthcare), cloned into pGEM-T Easy, sequenced, and cloned into the pCL6 vector (9) to create the mamC-ezz fusion construct pBBRCZZ (Table 2). The P_mamDC promoter was PCR amplified, cloned into pJET1.2/blunt (Fermentas), sequenced, and cloned into pBBRCZZ to yield pBBRPdcCZZ. This plasmid was transferred to M. gryphiswaldense, and MamC-ZZ magnetosomes were purified as described previously (9) with the only exception that the cells were lysed with a bench top constant cell disruptor (135,000 kPa) (Constant Systems) instead of a French press. After incubation of MamC-ZZ modified, unmodified, and GFP-displaying magnetosomes with rabbit anti-GFP antibody, we detected the highest abundance of the rabbit anti-GFP antibody with a shrimp alkaline phosphatase-labeled goat anti-rabbit antibody on the surface of MamC-ZZ magnetosomes (Fig. 3a and b). The signal was substantially stronger than the signal observed for MamC-GFP modified magnetosomes, which were produced by expression of the MamC-GFP fusion from the P_lac promoter (9) (Fig. 3b). These results demonstrate that MamC-ZZ modified magnetosomes efficiently bind rabbit antibodies. In addition, MamC-ZZ magnetosomes that were incubated only with alkaline phosphatase-labeled goat antibody yielded a weak signal (Fig. 3b). This indicates that MamC-ZZ modified antibodies also interact weakly with goat antibody. This was expected, as the staphylococcal protein A, from which the ZZ protein domain was derived, interacts weakly with goat antibodies but strongly with rabbit antibodies (14). While ZZ-modified magnetosomes have been produced in M. magneticum via genetic engineering of the MamC homolog Mms13 previously (27), this is the first time that antibody-binding magnetosomes from M. gryphiswaldense were generated by genetic engineering. One particular advantage of M. gryphiswaldense as a host for production of such genetically functionalized magnetosomes is that several mutant strains with an average magnetosome size of between 24 and 37 nm are available (16), which makes it possible to produce a wide range of magnetosomes with engineered bio- and physicochemical characteristics. The purified ZZ-modified magnetosomes can be used for magnetoimmunoassays (27) or for purification of antibodies, as described with ZZ-modified bacterial polyester granules (1).

FIG. 3.

Antibody-binding assay of MamC-ZZ magnetosomes. (a) Schematic illustration of the antibody-binding assay procedure. 1. MamC-ZZ modified magnetosomes and controls were diluted to a concentration of 1 mM Fe in 500 μl blocking solution (TBS [20 mM Tris, 0.5 M NaCl, pH 7.5] plus 0.5% [wt/vol] milk powder). The samples were incubated for 30 min at room temperature before rabbit anti-GFP antibody was added at a 1:2,000 dilution. 2. After incubation for 45 min, magnetosomes were magnetically collected and resuspended in TBS. 3. After a second magnetic separation step, a conjugate of shrimp alkaline phosphatase and goat anti-rabbit antibody was added in a 1:2,000 dilution in TBS. 4. After an incubation period of 45 min, the magnetosomes were magnetically separated and washed with TBS three times. 5. The particles were resuspended in 200 μl TBS, of which 100 μl was incubated with 100 μl Attophos shrimp alkaline phosphatase detection reagent (Roche) for 5 min. Fluorescence was detected with an Infinite 500 plate reader (Tecan). The excitation wavelength was 430 nm (20-nm bandwidth), and emission was recorded at 535 nm (25-nm bandwidth). (b) Antibody-binding assay of MamC-ZZ modified magnetosomes. The assay was performed with magnetosomes purified from M. gryphiswaldense pBBRPdcCZZ (ZZ), with MamC-GFP modified magnetosomes (GFP) (from M. gryphiswaldense pCL6 [9]), or with wild-type magnetosomes (wt). The magnetosomes were treated either with a primary rabbit anti-GFP antibody (RG), with a shrimp alkaline phosphatase conjugate of a goat anti-rabbit antibody (AR), or with both antibodies (RG, AR). Standard deviations calculated from three replicates are indicated.

In conclusion, we demonstrate that P_mamDC is a powerful tool for the genetic engineering of magnetosome proteins to generate functionalized magnetic nanoparticles for bio- and nanotechnological applications.