Predominant membrane localization is an essential feature of the bacterial signal recognition particle receptor

Abstract

Background: The signal recognition particle (SRP) receptor plays a vital role in co-translational protein targeting, because it connects the soluble SRP-ribosome-nascent chain complex (SRP-RNCs) to the membrane bound Sec translocon. The eukaryotic SRP receptor (SR) is a heterodimeric protein complex, consisting of two unrelated GTPases. The SRbeta subunit is an integral membrane protein, which tethers the SRP-interacting SRalpha subunit permanently to the endoplasmic reticulum membrane. The prokaryotic SR lacks the SRbeta subunit and consists of only the SRalpha homologue FtsY. Strikingly, although FtsY requires membrane contact for functionality, cell fractionation studies have localized FtsY predominantly to the cytosolic fraction of Escherichia coli. So far, the exact function of the soluble SR in E. coli is unknown, but it has been suggested that, in contrast to eukaryotes, the prokaryotic SR might bind SRP-RNCs already in the cytosol and only then initiates membrane targeting.

Results: In the current study we have determined the contribution of soluble FtsY to co-translational targeting in vitro and have re-analysed the localization of FtsY in vivo by fluorescence microscopy. Our data show that FtsY can bind to SRP-ribosome nascent chains (RNCs) in the absence of membranes. However, these soluble FtsY-SRP-RNC complexes are not efficiently targeted to the membrane. In contrast, we observed effective targeting of SRP-RNCs to membrane-bond FtsY. These data show that soluble FtsY does not contribute significantly to cotranslational targeting in E. coli. In agreement with this observation, our in vivo analyses of FtsY localization in bacterial cells by fluorescence microscopy revealed that the vast majority of FtsY was localized to the inner membrane and that soluble FtsY constituted only a negligible species in vivo.

Conclusion: The exact function of the SRP receptor (SR) in bacteria has so far been enigmatic. Our data show that the bacterial SR is almost exclusively membrane-bound in vivo, indicating that the presence of a soluble SR is probably an artefact of cell fractionation. Thus, co-translational targeting in bacteria does not involve the formation of a soluble SR-signal recognition particle (SRP)-ribosome nascent chain (RNC) intermediate but requires membrane contact of FtsY for efficient SRP-RNC recruitment.

Figure 1

Binding of FtsY to ribosome nascent chains requires the presence of signal recognition particles (SRP) and guanosine 5'(β,-γ imido) triphosphate (GMP)-PNP. MtlA189 RNCs were in vitro synthesized in the presence of purified SRP and in vitro synthesized, purified FtsY. When indicated puromycin or GMP-PNP were added. Ribosome-nascent chains (RNCs) were subsequently separated by centrifugation through a sucrose cushion into pellet fraction (P) and supernatant (S), which were analysed on SDS-polyacrylamide gel electrophoresis (SDS-PAGE). Both panels of the figure correspond to the same gel but were separated due to the large size-difference between FtsY and the RNCs. The radioactive material was quantified using a phosphor imager and the Imagequant software. Three independent experiments were performed.

Figure 2

FtsY-depleted inner membrane vesicles are unable to support co-translational targeting. (A) MtlA189 RNCs were incubated with wild-type (wt) INV, FtsY-containing IY28 INV (FtsY⁺) and FtsY-depleted IY28 INV (FtsY) (1 μl, 50 μg protein) and subjected to flotation gradient centrifugation. Subsequently, the gradient was separated into five fractions, which were analysed on SDS-PAGE. Fraction 2 and 3 of the gradient correspond to the membrane fraction, while fraction 4 and 5 reflect the ribosome-nascent chains that did not bind to the inner membrane vesicles [INV]. The sum of the radioactive material in all fractions was set as 100% and the amount of radioactive material in the individual fractions was quantified using a phosphor imager. (B) Flotation gradient analyses as in A, but FtsY-depleted IY28 INV were pre-incubated with purified FtsY (2 μg/25 μl; reconst. IY28). (C) Western blot analyses of the INV (5 μl; 250 μg protein) analysed in A and B using polyclonal antibodies against FtsY and against the integral membrane protein YidC. (D) Western blot analyses of the in vitro transcription/translation system used in this study using polyclonal anti-FtsY antibodies. Two equivalents of the in vitro system components were loaded. One equivalent corresponds to the amount required for in vitro protein synthesis. As control wild-type INV (200 μg protein) and purified FtsY (0.1 μg) were loaded.

Figure 3

Soluble FtsY- signal recognition particle (SRP) - MtlA189 ribosome-nascent chains (RNCs) are not efficiently targeted to inner membrane vesicles. (A) In vitro synthesized MtlA189 RNCs were incubated with purified SRP, FtsY or both in the presence of guanosine 5'(β,-γ imido) triphosphate (GMP)-PNP. The RNCs were subsequently isolated by centrifugation through a sucrose cushion and resuspended in buffer (50 mM triethanolamine acetate, pH 8; 50 mM potassium acetate; 5 mM magnesium acetate; 1 mM DTT). One volume was directly trichloroacetic acid precipitated and five volumes were subjected to immune precipitation using sepharose-bound polyclonal antibodies against either Ffh or FtsY. As control pre-immune serum (Pre-IS) was used. (B) The MtlA189 RNCs shown in A, and containing either SRP or SRP and FtsY, were incubated with the indicated inner membrane vesicles (1 μl, 50 μg protein) and subjected to flotation gradient centrifugation.

Figure 4

Cellular localization of FtsY in vitro and in vivo. (A) Escherichia coli cells were grown on LB medium up to mid-exponential phase (OD₆₀₀ 1.2). Cell breakage was performed using a French pressure cell in the presence of protease inhibitors. Unbroken cells and large cell fragments were removed by centrifugation and the supernatant of this centrifugation was then separated by ultracentrifugation into the soluble fraction (S) and the pellet fraction (P). Conditions indicate the centrifugation time in a Ti50.2 rotor at 45,000 rpm. After western transfer, the different fractions were analysed using antibodies against FtsY, against the integral membrane protein YidC and against the soluble protein Hsp60 (GroEL). (B) The functionality of the FtsY-green fluorescent protein (GFP) constructs was analysed by expressing plasmid-borne copies in the conditional FtsY depletion strain IY28. IY 28 containing either no plasmid (IY28) or the indicated plasmids was grown on LB-plates in the presence or absence of arabinose. Wt FtsY corresponds to untagged FtsY, wt FtsY-GFP corresponds to full length FtsY fused C-terminally to GFP. FtsY(B3)-GFP and FtsY(NG+1)-GFP correspond to GFP-tagged FtsY mutants which exhibit reduced activity due to impaired membrane binding. The growth experiments were performed in the absence of IPTG for preventing high-level expression of the plasmid-borne FtsY derivatives. (C) Western blot analyses of IY28 cells containing either no or the indicated plasmids. Cells were grown in the presence of arabinose or fructose but in the absence of IPTG.

Figure 5

FtsY is predominantly membrane bound in vivo. (A) Wild-type Escherichia coli cells carrying different FtsY-green fluorescent protein derivatives as described in Figure. 4 were analysed by fluorescence microscopy. The upper panel displays the original image and the lower panel a processed image after a three dimensional deconvolution of Z-stacks. (B) The localization of FtsY and the integral membrane protein SecY were analysed in wild-type E. coli cells containing only the endogenous amounts of FtsY and SecY. For this purpose, immune-fluorescence was carried out using polyclonal SecY and FtsY antibodies and Alexa fluor 555 labelled secondary antibodies. (C) Bacillus subtilis cells expressing FtsY-YFP as sole source of the protein growing exponentially; foci along the lateral cell membrane are indicated by white triangle. White bar 2 μm.