In order to identify the secretion pathway the fusion construct sSpAD-N^proEDDIE-pep6His was expressed under control of the strong T5 promoter. This promoter enabled the expression of detectable protein amounts in the presence of toxins, such as carbonyl cyanide m-chlorphenylhydrazone (CCCP) within short incubation times in different host strains.

Several secretion pathways can be analyzed using diverse protonophores and knockout strains. CCCP has the ability to specifically inhibit all proton motive force driven pathways in E. coli [18,19]. To test if sSpAD secretion was driven by any of these pathways the recombinant fusion protein was expressed and cultivated in the presence of CCCP. As shown in Figure 2A, CCCP inhibited secretion into the periplasm. The slight signal in the periplasmic fraction could be considered as an artefact because minor expression of the recombinant protein occurs even before induction and addition of CCCP. Thus the signal in the periplasmic fraction was most likely due to secretion before inhibition of the given secretion pathway. Immuno blots against GroEL were performed to confirm the integrity of the spheroplasts. GroEL is a cytoplasmic chaperone, which was not found in the periplasm (Figure 2B). Anti Maltose binding protein E (MalE) immuno blots were performed to verify the isolated fractions as periplasm of E. coli cells (Figure 2C).

DADE is a MC4100 derived Tat knock-out strain lacking a functional Tat pathway[20]. This strain was used to show if sSpAD secretion was Tat dependent. Subcellular fractions showed that Tat knockout had no effect on the secretion capacity of the fusion protein shown in Figure 2A. Therefore, the involvement of the Tat pathway in the secretion process was excluded.

Sodium-azide is a strong inhibitor of the ATPase SecA and therefore has the ability to specifically inhibit the Sec translocation [21] in E. coli. Preliminary experiments with sodium-azide indicated an inhibition of translocation, although, recombinant expression artefacts were observed in the periplasmic fraction due to the osmotic shock procedure. Therefore, experiments were performed in parallel, using the target protein with and without the secretion tag to determine the background. The densitometric comparison of the amount of recombinant protein with and without sSpAD-tag in the periplasmic fraction confirmed that sodium-azide inhibited secretion. Continuative experiments were performed with the SecE knock-out strain CM124 [21]. As the Sec pathway is crucial for cell viability, SecE was complemented via expression from an L-arabinose induced plasmid. Over-night (including L-arabinose) culture was split, diluted and subsequently cultivated with and without L-arabinose. After dilution the culture without L-arabinose should have no functional Sec pathway. As shown in Figure 3 the knockout of the Sec pathway inhibited secretion of the target protein. These experiments led to the assumption that the Sec pathway seems to be involved in some way in the export of sSpAD fusion proteins to the periplasm; however the SignalP 3.0 prediction tool http://www.cbs.dtu.dk/services/SignalP did not reveal a canonical Sec secretion signal. The recent advances in genomics and proteomics revealed numerous extracellular proteins lacking defined secretion signals [22], often combining functions within the cytoplasm and in the extracellular environment [23]. Possibly, the elevated levels of expression of the sSpAD tagged proteins could lead to an emergency mechanism, which exports proteins through the Sec channel without using a signal sequence.

The fusions of sSpAD to the pestiviral autoprotease N^proEDDIE-pep6His, to green fluorescent protein GFPmut3.1 and to human superoxide dismutase 1 (SOD) produced by shaking flask cultivation were quantified by densitometry of SDS-PAGE as described in Methods. Since the fusion tag was not cleaved after secretion, it was not possible to distinguish easily between secreted and non-secreted proteins. Preliminary experiments showed that overexpressed proteins lacking export signals were detected in the periplasmic fraction. Therefore all recombinant proteins were expressed with and without sSpAD tag and the periplasmic fractions of all samples were isolated. Subsequently, the concentrations of the target proteins in the periplasm of all samples were measured and the amount found in the periplasm without sSpAD subtracted from the amount of proteins secreted with the sSpAD tag. The corrected secretion capacities are given in Table 1. The secretion yields for sSpAD-SOD (16.4 mg/L) and sSpAD-N^proEDDIE-pep6His (11.3 mg/L) were well above the reported yield for Sec mediated secretion [9]. However, in contrast to N^proEDDIE-pep6His (not detectable in the periplamic fraction) and SOD (approximately 30 percent in the periplamic fraction), a major part of GFPmut3.1 was found in the periplasm, although the lack of the cytoplasmic chaperone GroEL indicated intact spheroplasts. Due to expression of GFPmut3.1 from a codon-optimized gene the protein level of GFPmut3.1 was rather high and increased even with the sSpAD tag. Therefore it was assumed, that these high expression levels provoked a passive diffusion of GFPmut3.1 into the periplasm.

In order to exemplify the quick and easy purification of fusion proteins the periplasmic fraction of a 10 ml shaking flask cultivation of sSpAD-SOD was purified. The sample was applied on an IgG column as described in Methods. After a single purification step 16.9 mg/L of the purified fusion protein could be obtained (Figure 4). This was in good agreement to the amount found by determination of the secretion capacity (16,4 mg/L).

The pET30a T7 promoter (T7p) was replaced by three alternative promoters e.g. T5, the artificial Tac promoter and the lacUV5 promoter resulting in the three vectors pT5, pTac and pLacUV5. Two oligonucleotides corresponding to the given promoter sequence with complementary bases, prom lacUV5 SphI F and prom lacUV5 XbaI R, were directly ligated into an SphI and XbaI digested pET30a plasmids. The Tac promoter, consisting of two oligonucleotides with 65 complementary bases, prom Tac SphI F and prom Tac XbaI R, was directly ligated into SphI and XbaI digested pET30a plasmids. The T5 promoter was generated by PCR with the given primers in Table 2 and subsequently cloned into SphI and XbaI digested pET30a plasmids.

Codon optimized GFPmut3.1 gene was amplified using the primers sGFP F NdeI start and sGFP R (Table 2). The amplified fragment was ligated into NdeI and SalI digested pLacUV5 vector, resulting in pLacUV5-GFPmut3.1 plasmid.

The sSpAD signal sequence was codon optimized (sequence given in the appendix 1) and amplified using the primers pET30 sSpAD NdeI F and pET30 sSpAD NdeI R. Subsequently sSpAD was subcloned into the pLacUV5 GFPmut3.1 vector resulting in the pLacUV5-sSpAD-GFPmut3.1.

The sSpAD-N^pro-EDDIE-pep6His construct was generated by digestion with NdeI of the pET30 N^pro-EDDIE-pep6His vector and subsequent ligation with the same insert generated for the pLacUV5-sSpAD-GFPmut3.1 construct. Subsequently the promoter of the pET30-sSpAD-N^pro-EDDIE-pep6His construct was replaced with the promoter LacUV5.

The pLacUV5-sSpAD-SOD plasmid was generated in two steps. First sSpAD was amplified using the primers pET 30 sSpAD NdeI F and pLacUV5 sSpAD NheI R and cloned into the pLacUV5 vector, which resulted in the pLacUV5 vector with an additional NheI restriction site. Subsequently the codon optimized SOD gene was amplified using the primers SOD NheI F and SOD SalI R (Table 2). The amplified fragment was ligated into the pLacUV5-sSpAD-Nhe vector, which was digested with NheI and SalI, resulting in pLacUV5 sSpAD-SOD plasmid. A list of the constructs is given in Table 3.

Expression plasmids were transformed into E. coli BL21(DE3), subsequently a single colony was picked to inoculate overnight cultures. These cultures were diluted 1:20 with fresh TY-medium, supplemented with 0.5% glucose and grown to a density of OD₆₀₀: 1.0 at 37°C/225 rpm. Recombinant protein synthesis was induced adding 1 mM IPTG. Expression with T5 promoter was carried out for 1 h at 37°C/225rpm in BL21(DE3), BL21(DE3) with 2 mM NaN₃, BL21(DE3) 50 μM CCCP. For the shake flask cultivation with E. coli DADE experimental procedure was similar to E. coli BL21(DE3) but without addition of glucose during cultivation.

In contrast overnight cultures of CM124 cells carrying pET30-Tac-promoter-plasmids were grown in the presence of 0.2% L-arabinose. These overnight cultures were diluted 1:20 and split. The divided cultures were grown with and without 0.2% L-arabinose in parallel. Expression was induced at OD₆₀₀: 0.5 with 1 mM IPTG and the cultures were incubated for 1 h at 37°C/225 rpm.

For the determination of the solubility and the secretion capacity overnight cultures of the host strain BL21(DE3) carrying the pET30-LacUV5-promoter plasmids were diluted 1:20. Expression was induced at OD₆₀₀: 0.5 with 1 mM IPTG and the cultures were incubated for 2 h at 37°C/225 rpm.

Isolation of the periplasm was performed at 24°C with a gentle osmotic shock procedure to minimize E. coli cell disruption during preparation (modified from [15]). After expression, a culture volume corresponding to 10 ml of OD₆₀₀ 2.0 was centrifuged at 3000 g for 10 min. The pellet was completely suspended in one culture volume of osmolysis buffer (100 mM TRIS-HCl pH 7.8, 15.4% sucrose, 3 mM EDTA) and incubated at 50 rpm for 10 min, followed by centrifugation. The supernatant was discarded and the pellet was resuspended in one culture volume water and incubated for 10 min at 50 rpm. Afterwards the suspension was centrifuged for 10 min at 3000 g. The supernatant, containing periplasmic proteins, was decanted and filtrated (membrane, 0.45 μm pore size). The spheroplasts were kept as a cellular fraction. For SDS-PAGE analysis periplasmic and cytoplasmic fraction samples were precipitated with TCA and all pellets solubilised in loading buffer (62.5 mM Tris-HCl pH 6.8, 10% glycerol, 2% SDS, 0.0025% bromophenol blue, 50 mM DTT).

For the determination of the solubility the cells were taken up in a culture volume lysis buffer (20 mM Na₂HPO₄ pH 8.0, 75 mM NaCl, 5 mM EDTA) and disrupted with a French press (American Instruments Co., Inc). Aliquots of the lysate were collected and centrifuged at 14000 rpm for 15 min. The supernatant contained soluble cytoplasmic protein, whereas the pellet represented the insoluble protein fraction. For SDS-PAGE analysis the samples were precipitated with TCA and all pellets solubilized in loading buffer (62.5 mM Tris-HCl pH 6.8, 10% glycerol, 2% SDS, 0.0025% bromophenol blue, 50 mM DTT).

To the filtrated periplasm fraction Na₂HPO₄/NaH₂PO₄ and NaCl were added to a final concentration of 20 mM Na₂HPO₄/NaH₂PO_4, 500 mM NaCl, pH 8. 10 ml of this solution were concentrated to 2 ml with ultrafiltration devices. After removal of precipitated proteins by centrifugation, the supernatant was applied to a pre-equilibrated 500 μl gravity flow IgG-NHS-Sepharose column and purified. The column was washed with 2.5 ml 20 mM Na₂HPO₄/NaH₂PO₄ 500 mM NaCl pH 8 buffer. The recombinant protein was eluted with 5 column volumes of 0.2 glycine buffer pH 3. The eluted fractions were pooled and the total concentration of the purified protein was quantified.

Cellular integrity after expression of the recombinant proteins and subsequent osmotic shock treatment was surveyed by Immuno blot analysis with antibodies against the periplasmic Maltose binding protein MalE, and the cytoplasmic chaperone GroEL. Cell fraction samples were separated on 4-20% Tris-Glycine gels and the proteins electrophoretically transferred onto nitrocellulose membranes. Incubation times of the antibodies were carried out according the instruction manuals.

Quantification of the secretion capacity was carried out by densitometric analysis of target proteins in comparison to BSA standards on Coomassie stained gels. Gels were photographed and analyzed with AlphaEaseFC software (Alpha Innotech Corporation).

GFPmut3.1 [P42212]; SOD [BT008028.1]; Staphylococcal Protein A[P38507]:Domain D:

ADAQQNKFNKDQQSAFYEILNMPNLNEEQRNGFIQSLKDDPSQSTNVLGEAKKLNESQAPK N^proEDDIE [7]

DNA sequence sSpAD (codon optimized):

GCAGACGCACAACAGAATAAGTTTAACAAAGACCAGCAGAGCGCATTCTACGAAATTCTGAACAT

GCCGAATCTGAATGAGGAACAACGTAATGGCTTTATTCAGTCTTTAAAAGACGACCCATCTCAGA

GCACCAACGTTCTGGGCGAAGCAAAGAAACTGAACGAATCTCAGGCACCAAAA