In vivo characterization of the bacterial intramembrane-cleaving protease RseP utilizing the heme binding tag-based assay iCliPSpy
Substrate and product of the response catalyzed by RseP
An efficient manufacturing technique for the heme-binding peptide TNFα-(1-39) as potential substrate of RseP and different I-CLiPs, is its expression as MBP fusion protein in E. coli25. To make sure the proper topology [considering that full-length TNFα is a type II transmembrane protein: N-terminus in (cytoplasm) and C-terminus out (periplasm/lumen of the ER)], MBP was fused to the N-terminus of TNFα-(1-39) with out the sign peptide (residues -24 to -1, Fig. 1b). TNFα-(1-39) accommodates solely the N-terminal 9 amino acid residues L31FLSLFSFL39 of the TNFα TMD and is subsequently present in each the membrane and cytosolic fractions25,27. As an alternative of the MBP assemble described in our latest examine25, we used a mutant MBP protein (MBPmut) containing extra substitutions, together with I317V, which additional will increase the affinity for amylose31 (Fig. 1b). To extend the quantity of heme sure by TNFα, we used the variant peptide TNFα-(1-39)-L31P, which reveals a stronger interplay with heme in comparison with the wildtype peptide25. As well as, MBPmut and TNFα-(1-39)-L31P are linked in MBPmut-TNFα-(1-39)-L31P by a shortened linker peptide (Fig. 1b), which ought to enhance the proteolytic stability of the fusion protein. We expressed the putative product of RseP-mediated intramembrane proteolysis, TNFα-(1-34), in the identical method [MBPmut-TNFα-(1-34)-L31P]. The MBPmut-TNFα-(1-39)-L31P, MBPmut-TNFα-(1-34)-L31P and His-MBP-TNFα-(1-39)-L31P genes (Fig. 1b) have been cloned into the pETDuet-1 vector (Supplementary Desk 1 and Supplementary Observe 1) and the recombinant plasmids have been reworked into E. coli T7 Specific cells. For in vivo experiments, expression of MBP fusions was carried out with out IPTG induction on agar plates (Fig. 2a). A normal drawback of the T7 expression system is the already current low protein expression with out IPTG induction, additionally referred to as leaky expression. Nonetheless, leaky expression might be a bonus, particularly within the expression of membrane proteins, because the low-level expression prevents the formation of inclusion our bodies32,33 and likewise favors heme-binding of TNFα-(1-39)-L31P, because the heme focus in cells is extraordinarily low34.
After a number of days at room temperature cells expressing MBPmut-TNFα-(1-39)-L31P have been ocher inexperienced, whereas cells expressing His-MBP-TNFα-(1-39)-L31P have been solely faintly coloured and cells expressing MBPmut-TNFα-(1-34)-L31P or the empty vector management pETDuet-1 have been white (Fig. 2a, b). SDS-PAGE evaluation of complete cells scraped from agar plates confirmed the manufacturing of the totally different MBP-TNFα fusion proteins and confirmed that diminished expression of His-MBP-TNFα-(1-39)-L31P is answerable for the much less intense coloration of those cells (Fig. 2c). Due to this fact, we chosen the MBPmut-TNFα-(1-39)-L31P, and MBPmut-TNFα-(1-34)-L31P constructs for additional experiments. The membrane-associated MBP fusion proteins have been then solubilized with the non-ionic detergent n-dodecyl-β-D-maltopyranoside (DDM) and purified by amylose affinity chromatography within the presence of DDM. UV/Vis spectroscopy of the purified fusion proteins confirmed the ligation of heme by bis-thiolate in MBPmut-TNFα-(1-39)-L31P and by thiolate in MBPmut-TNFα-(1-34)-L31P (Fig. 2nd-f). MBPmut-TNFα-(1-34)-L31P with thiolate ligated heme reveals an absorbance most at about 392 nm, because the addition of DDM ends in a bathochromic shift of the absorbance most by ∼20 nm (Fig. 2f). Cells expressing MBPmut-TNFα-(1-34)-L31P are white as a result of the heme content material may be very low. Heme reconstitution of MBPmut-TNFα-(1-34)-L31P, nevertheless, resulted in a red-brown coloration of the fusion protein (Fig. 2nd, f). Taken collectively, our outcomes present that recombinantly expressed reporter substrate and mimicked product of I-CliP-mediated proteolysis in E. coli differ of their spectroscopic properties, leading to seen adjustments within the coloring of bacterial colonies. Subsequent, we examined whether or not our reporter MBPmut-TNFα-(1-39)-L31P additionally serves as a substrate for bacterial I-CLiPs.
Visualization of the in vivo exercise of RseP
A well-characterized bacterial I-CLiP is E. coli RseP, which we used to develop the I-CLiP assay iCliPSpy. A beforehand revealed in vivo assay for monitoring RseP exercise9,23 makes use of E. coli rseP rseA double-disrupted cells carrying plasmids expressing RseP-Myc8 and an HA-MBP-tagged, truncated model of RseA, RseA140, as a substrate for RseP. RseA140 cleavage by RseP was monitored by immunoblot evaluation9,23 and this assay has been efficiently used to review the construction, perform and substrate binding of RseP by site-directed mutagenesis19,20,21,35. For the reason that in vivo exercise of RseP shouldn’t be straight seen when taking a look at colonies of E. coli cells over-expressing RseP-Myc and HA-MBP-RseA140, we sought to develop an assay that enables easy detection of I-CLiP exercise primarily based on the colour of the colonies.
To check whether or not MBPmut-TNFα-(1-39)-L31P might be cleaved by RseP, RseP-Myc8 wild kind (wt, Supplementary Fig. 1a) or the catalytically inactive variant RseP-Myc H22F have been expressed along with our I-CliP reporter. Since RseP is crucial for bacterial progress7, we carried out iCliPSpy in presence of the chromosomal copy of rseP. In precept, the RseP-dependence of progress might be circumvented by down-regulating the porins OmpA and OmpC36. Our knowledge, nevertheless, present that the quantity of chromosomally encoded RseP shouldn’t be ample to degrade overproduced MBPmut-TNFα-(1-39)-L31P to any considerable extent (Fig. 3a), in order that extra genetic modification was not required.
Co-expression of the I-CliP reporter with RseP-Myc wt resulted in white cells (Fig. 3a), indicating that the overexpressed RseP had cleaved TNFα-(1-39)-L31P. The H22F substitution in RseP, which results in a lack of zinc binding, inhibited RseP exercise, in order that co-expression of MBPmut-TNFα-(1-39)-L31P with rseP-Myc H22F resulted in inexperienced E. coli colonies (Fig. 3a). SDS-PAGE evaluation of complete cells scraped from the agar plate confirmed comparable expression ranges of the MBPmut-TNFα fusions and confirmed that the RseP cleavage product of MBPmut-TNFα-(1-39)-L31P has an analogous obvious molecular weight as MBPmut-TNFα-(1-34)-L31P, indicating that just a few amino acid residues have been cleaved off by RseP (Fig. 3b). Expression of rseP-Myc wt and rseP-Myc H22F genes was confirmed by western blotting (Fig. 3c). Larger expression ranges of rseP-Myc H22F in comparison with rseP-Myc wt have been much less pronounced in case of co-expression with MBPmut-TNFα-(1-39)-L31P (examine the primary two lanes with lanes 5 and 6 of the immunoblot proven in Fig. 3c). As anticipated and regardless of this greater expression, proteolytic cleavage of MBPmut-TNFα-(1-39)-L31P by RseP-Myc H22F was not noticed. In vivo processing of MBPmut-TNFα-(1-39)-L31P was additionally confirmed for untagged RseP (Supplementary Figs. 1a-c and 2a, b). Nonetheless, when a 10x HisTag (His10) was fused straight with out a linker peptide to the carboxy terminus of RseP, the exercise of RseP was decreased (Supplementary Fig. 1a-c), suggesting that the interplay of the periplasmic His10 Tag with the PDZ domains (Supplementary Fig. 1b) might have an effect on folding, stability or exercise of RseP.
Basically, our outcomes counsel that green-colored colonies seem when ample MBPmut-TNFα-(1-39)-L31P is current. White colonies, nevertheless, are noticed when MBPmut-TNFα-(1-39)-L31P has been cleaved or when the reporter is expressed at very low concentrations. Due to this fact, an expression charge of the I-CLiP reporter that’s impartial of the sequence of the co-expressed rseP gene (Fig. 3b and Supplementary Fig. 2b) is a necessary requirement for this assay.
Processing adjustments heme binding properties of the substrate
Our knowledge counsel that intramembrane proteolysis catalyzed by RseP impacts the dimerization motif S34LFS37 of TNFα-(1-39) and consequently alters the heme binding mode of the TNFα peptide, in step with the colour change. To check whether or not MBPmut-TNFα-(1-39)-L31P certainly serves as a substrate for RseP cleavage, membrane-associated proteins solubilized with DDM from IPTG-induced E. coli cells co-expressing MBPmut-TNFα-(1-39)-L31P with both rseP wt or rseP H22F have been enriched by way of amylose affinity chromatography (Fig. 4a), analyzed by SDS-PAGE (Fig. 4b) and the heme coordination mode was decided by UV/Vis spectroscopy (Fig. 4c). The ratio of energetic and accurately folded protease to substrate could differ after IPTG-induced overexpression in comparison with leaky expression within the in vivo assay. This might result in totally different proportions of absolutely processed substrate. Certainly, we noticed that each amylose affinity columns have been inexperienced, however the coloration was extra intense within the case of co-expression with rseP H22F (Fig. 4a). MBPmut-TNFα-L31P proteins purified from rseP wt over-expressing cells confirmed a barely decrease obvious molecular weight than MBPmut-TNFα-L31P proteins purified from rseP H22F over-expressing cells (Fig. 4b). The UV/Vis spectrum of MBPmut-TNFα-L31P purified from rseP H22F over-expressing cells confirmed absorbance maxima at 371 nm and 451 nm, demonstrating bis-thiolate ligation and confirming that RseP H22F doesn’t cleave MBPmut-TNFα-(1-39)-L31P. For MBPmut-TNFα-L31P purified from rseP wt over-expressing cells, nevertheless, the absorbance at 451 nm was considerably diminished, the absorbance most at 371 nm was shifted to 373 nm and the absorbance at 392 nm was elevated [Fig. 4c, see also Fig. 2f, which shows a UV/Vis absorbance spectrum of the putative RseP cleavage product MBPmut-TNFα-(1-34)-L31P, which would be expected if proteolysis was complete and the dimerization motif disrupted]. As anticipated, unprocessed and RseP-processed MBPmut-TNFα-(1-39)-L31P fusion proteins have been additionally current within the cytosolic fraction (Supplementary Fig. 3a-e; examine additionally Kupke et al.25).
Solubilized membrane-associated MBPmut-TNFα-L31P proteins purified from E. coli co-overexpressing both inactive rseP H22F or energetic rseP wt have been then separated by gel filtration in presence of DDM into heme-binding (monomeric or dimeric) MBPmut-TNFα-L31P and heme-free monomeric MBPmut-TNFα-L31P fusion proteins (Fig. 4d). Heme-free monomers are at all times current as heme shouldn’t be stoichiometrically sure by TNFα-ICDs25. The upper content material of heme-free monomers of MBPmut-TNFα-L31P within the case of RseP wt in comparison with RseP H22F might be associated to a decrease heme binding affinity of the processed fusion protein (examine UV/Vis spectra in Fig. 4c), as a result of hydrophobic residues throughout the heme-binding motif SRRCPFLSLFSFL of TNFα might also contribute to heme binding and since bis-thiolate ligation stabilizes heme binding25. The heme-binding MBPmut-TNFα-L31P fusion protein obtained within the case of RseP H22F eluted at 15.40 ml, whereas the protein noticed for RseP wt eluted at 16.05 ml (comparable to a decrease obvious molecular weight) from the scale exclusion column (Fig. 4d). Since MBPmut-TNFα-(1-39)-L31P accommodates solely a single cysteine residue [Cys30 of TNFα-(1-39)-L31P,] bis-thiolate ligation of heme requires dimerization25 (Fig. 1a). Due to this fact, the eluted proteins have been once more characterised by UV/Vis spectroscopy to differentiate between monomeric and dimeric heme-binding proteins (Fig. 4e). The UV/Vis spectra of the heme-binding proteins purified by gel filtration verify that dimeric MBPmut-TNFα-(1-39)-L31P is current when MBPmut-TNFα-(1-39)-L31P was co-expressed with rseP H22F however not with rseP wt. Within the latter case, the absorbance most at about 392 nm (measured within the presence of DDM) reveals that processed MBPmut-TNFα-L31P coordinates heme by way of a single cysteine residue and is monomeric. It seems that RseP-processed (monomeric) heme binding MBPmut-TNFα-L31P shouldn’t be uniform, because the noticed elution quantity is as talked about both 16.05 ml (when measuring the absorbance at 280 nm) or 15.85 ml (absorbance at 371 nm; Fig. 4d). We assume that no less than two totally different processing merchandise with totally different heme binding affinities (and thus totally different heme content material) are current. At greater protein concentrations, a processing product that also accommodates a part of the dimerization motif could dimerize to a small extent and subsequently additionally reveals an absorbance most at about 450 nm (and 373 nm; examine spectrum “+ RseP wt” of Fig. 4e with the corresponding spectrum of Fig. 4c).
In abstract, gel filtration reveals that cleavage by RseP ends in MBPmut-TNFα fusion proteins that differ in molecular weight and heme binding properties from the substrate MBPmut-TNFα-(1-39)-L31P.
RseP processing websites within the MBPmut-TNFα substrate
The UV/Vis absorbance spectrum of the MBPmut-TNFα-(1-39)-L31P fusion protein purified from cells co-overexpressing rseP wt reveals that the heme-coordinating residue Cys30 continues to be current within the proteolytic cleavage product, however that the bis-thiolate ligation of heme has been largely misplaced. Consequently, cleavage occurred inside TNFα TMD residues 31-39 of the reporter protein, disrupting the dimerization motif (examine Fig. 1). To find out the precise RseP processing website(s) throughout the C-terminal TNFα TMD residues P31FLSLFSFL39 by mass spectrometry, MBPmut-TNFα-(1-39)-L31P was once more co-expressed with both rseP wt, empty vector management pCOLADuet-1, rseP-Myc wt or rseP-Myc H22F. The MBP-TNFα fusion proteins produced have been purified from the membrane fraction by amylose affinity chromatography after which analyzed by SDS-PAGE (Fig. 5a, b) and UV/Vis spectroscopy (Fig. 5c) to verify the experimental knowledge proven in Fig. 4 and to make sure RseP catalyzed processing inside TNFα-(1-39)-L31P previous to mass spectrometric evaluation. As beforehand noticed, the MBPmut-TNFα fusion proteins that also bind heme by way of bis-thiolate ligation have a barely greater molecular weight than the fusion proteins that bind heme by way of mono-thiolate ligation (Fig. 5a, b), however the decision of SDS-PAGE is just too low to separate otherwise truncated proteins. Two complementary mass spectrometric analyses have been then carried out to find out the proteolytic cleavage websites: (i) evaluation of the intact mass of the MBP-TNFα fusion proteins (Fig. 5d, e and Supplementary Fig. 4a-d). and (ii) sequence evaluation and quantification of their tryptic peptides (Supplementary Fig. 4e). For the primary strategy, MBP-TNFα fusion proteins obtained from the 4 totally different genetic backgrounds have been separated by reversed-phase (RP)-liquid chromatography (LC) and the intact mass data was obtained by coupled electrospray ionization mass spectrometry (ESI-MS) and in contrast with the calculated common molecular weights of C-terminally truncated MBPmut-TNFα-(1-39)-L31P fusion proteins (Fig. 5d, e). The detected main degradation product MBPmut– … CP31 [lacking the terminal eight amino acid residues of TNFα-ICD-(1-39)] shouldn’t be an RseP processing product, as it’s also current within the absence of over-expressed energetic protease (Fig. 5e). Residues F32LSLFSFL39 may very well be cleaved off by unspecific proteolytic processing (impartial of RseP), as to our data there isn’t any E. coli K12 endopeptidase that particularly cleaves Professional-X peptide bonds. We assume that MBPmut– … CP31 is detectable, regardless of the absence of TNFα TMD residues, as a result of the membrane fraction has not been washed and subsequently additionally accommodates soluble proteins (but additionally the hydrophobic heme cofactor sure to the CP motif would possibly goal this fusion protein to the membrane).
The total-length MBPmut-TNFα-(1-39)-L31P fusion protein characterised by the C-terminal sequence … C30PFLSLFSFL39, nevertheless, was recognized each by its intact mass and by sequence evaluation of tryptic peptides (see Supplementary Observe 2), for “Empty vector management” and “RseP-Myc H22F” (inactive RseP protein) purifications (Fig. 5e, and Supplementary Fig. 4a, b, e). This full-length MBPmut-TNFα-(1-39)-L31P protein is ready to bind heme by bis-thiolate ligation, ensuing within the attribute absorbance maxima of 370 nm, 450 nm and 550 nm25,37,38.
Nonetheless, when MBPmut-TNFα-(1-39)-L31P was co-expressed with rseP wt or rseP-Myc wt, MBPmut-TNFα-(1-39)-L31P ( = MBPmut– … CPFLSLFSFL39) was cleaved between residues F36 and S37 in addition to between L33 and S34 of TNFα-(1-39)-L31P in each circumstances, ensuing within the main processing merchandise MBPmut– … CPFLSLF36 and MBPmut– … CPFL33 for each wt proteases (Fig. 5e, and Supplementary Fig. 4c-e). Upon co-expression with rseP wt, we discovered small quantities of three extra cleavage merchandise, MBPmut– … CPF32, MBPmut– … CPFLS34 (Fig. 5e, and Supplementary Fig. 4c) and MBPmut– … CPFLSLFS37 (Supplementary Desk 2). We have no idea whether or not these truncated MBPmut-TNFα-(1-39)-L31P proteins are additionally RseP processing merchandise or whether or not the 2 main RseP processing merchandise are additional degraded by different E. coli proteases. MBPmut– … CPFL33 can now not bind heme by bis-thiolate ligation, for the reason that lack of the C-terminal pentapeptide L35FSFL39 already results in mono-thiolate ligated heme25 (Fig. 2). MBPmut– … CPFLSLF36, which lacks the second serine residue Ser37 vital for (heme promoted) dimerization25, might nonetheless dimerize to a small extent and bind heme by bis-thiolate ligation (examine absorbance of the processing merchandise of RseP at 450 nm, Fig. 4c).
Small quantities of the key processing product of RseP, MBPmut– … SRRCPFLSLF36, have been additionally recognized for the “Empty vector management” however not for “RseP-Myc H22F” by LC-MS/MS evaluation (Supplementary Desk 2). It’s probably that minor cleavage of the substrate protein by chromosomally encoded RseP (and different E. coli proteases) happens, which might be detected by MS/MS after tryptic digest of purified MBP fusion proteins. This processing of MBPmut-TNFα-(1-39)-L31P was not detected in intact mass evaluation (proven in Fig. 5e and Supplementary Fig. 4a).
In abstract, mass spectrometric evaluation confirms that RseP cleaves the TNFα substrate throughout the TMD residues 31-39. The totally different heme binding modes of substrate and processing product(s) decided by UV/Vis spectroscopy are demonstrably associated to C-terminal truncation of the substrate. The mass spectrometric evaluation suggests a processive cleavage mechanism of RseP and likewise offers insights into the specificity of RseP.
Combining error-prone mutagenesis of RseP with iCliPSpy
To establish residues in RseP which can be crucial for protease exercise and/or substrate binding, we mixed the in vivo assay iCliPSpy with error-prone mutagenesis of the rseP gene (Fig. 6a, b). We chosen seven inexperienced E. coli clones co-producing an inactivated RseP enzyme and the substrate MBPmut-TNFα-(1-39)-L31P for sequencing of the rseP gene. In six out of seven clones, we discovered untimely cease codons and frameshift mutations along with a number of amino acid exchanges (Fig. 6c), and subsequently excluded them from additional evaluation. In distinction, the mutant designated RseP G1 contained eleven non-synonymous mutations, however neither untimely cease codons nor frameshift mutations. We mapped the eleven amino acid exchanges onto the AlphaFold wt construction of RseP and located that the Y69H substitution is positioned throughout the fringe of the β-MRE at a place near the active-site throughout the substrate binding groove (Fig. 6d). The sting construction was proven to be vital for each protease exercise and substrate binding19,20. For instance, the Y69P substitution in RseP considerably impairs the cleavage of the mannequin substrate HA-MBP-RseA14819. Nonetheless, the introduction of a Professional residue can considerably alter the construction of the β-MRE19,20. To additional examine the function of Tyr69 in protease perform with out altering the general construction of the β-MRE, we examined whether or not the only substitution Y69H, during which Tyr69 is exchanged for a extra hydrophilic histidine residue, reduces RseP exercise in vivo. Assuming that the colour of colonies on agar plates displays the proteolytic exercise of RseP, we discovered a decreased in vivo exercise of RseP Y69H in comparison with RseP wt (Fig. 6e), which was extra pronounced for RseP G1 and the active-site variant RseP H22F (Fig. 6e). Subsequent, we centered on a extra detailed characterization of residues Tyr69 and Tyr428 by MD simulations and site-directed mutagenesis. Residue Tyr428 of TMD 4 was included in our research as a result of it’s localized on the alternative facet of Tyr69 on the putative substrate binding groove, and Tyr69 and Tyr428 collectively may match in gating substrates22 into the binding groove (Fig. 6d).
MD simulations of membrane-embedded RseP
To additional examine the function of Tyr69 within the perform of RseP, we carried out MD simulations of RseP primarily based on the construction predicted by AlphaFold29,30 (with a Zn2+ ion manually added to the energetic website), because the pdb file 7W6X for the crystal construction of full-length RseP22 was not publicly accessible. To imitate the native setting, RseP was embedded in a synthetic membrane resembling the lipid composition of E. coli membranes (see Supplementary Desk 3 and Supplementary Observe 3) and MD simulations (Fig. 6f and Supplementary Fig. 5a-e) have been carried out on a time scale of two microseconds. The constructions obtained confirmed solely minor adjustments within the RseP energetic site-containing membrane core in comparison with the AlphaFold prediction used for the simulations (and in comparison with the RseP crystal construction). Nonetheless, the MD simulations indicated that the periplasmic tandem PDZ area is versatile (Supplementary Fig. 5e). The MD simulations of membrane-embedded RseP predict that Tyr69 and Tyr428 are localized on the cytosol/membrane interface adjoining to the energetic website of RseP as a part of the substrate binding groove (Fig. 6f and Supplementary Fig. 5). The physiological significance of a lipid molecule positioned by MD simulation subsequent to/within the substrate binding groove, nevertheless, is much less clear. Within the case of the MD simulation proven in Fig. 6f and Supplementary Fig. 5 this lipid was phosphatidylethanolamine (PE), in different simulations we discovered cardiolipin, suggesting that total hydrophobicity is the key determinant for insertion into the substrate binding pocket of RseP.
In our in vivo assay (Fig. 6g), the Y428H substitution had no important affect on the proteolytic exercise of RseP. Moreover, the double substitution Y69H/Y482H was indistinguishable from the only substitution Y69H (Fig. 6g), suggesting that Tyr428 neither performs a task in binding the substrate MBPmut-TNFα-(1-39)-L31P neither is vital for catalytic exercise. To help this outcome, MBP fusion proteins purified from E. coli cells co-overexpressing MBPmut-TNFα-(1-39)-L31P with wt or mutant rseP genes have been analyzed by SDS-PAGE (Supplementary Fig. 6a) and UV/Vis spectroscopy (Supplementary Fig. 6b). When MBPmut-TNFα-(1-39)-L31P was co-expressed with rseP Y69H or with rseP Y69H/Y428H, heme remained sure to the reporter by way of bis-thiolate ligation (Supplementary Fig. 6b), confirming that intramembrane proteolysis by RseP is impaired when Tyr69 is exchanged for His. Nonetheless, RseP Y428H behaved just like the wt enzyme, processing the C-terminal residues of the reporter substrate (examine Fig. 5) and shifting heme binding in direction of a monomeric, thiolate-bound kind.