Aspartate tightens the anchoring of staphylococcal lipoproteins to the cytoplasmic membrane

Abstract In gram‐negative bacteria, the ABC transporter LolCDE complex translocates outer membrane‐specific lipoproteins (Lpp) from the inner membrane to the outer membrane. Lpp possessing aspartate (Asp) at position +2 are not translocated because it functions as a LolCDE avoidance signal. In gram‐positive bacteria, lacking an outer membrane and the Lol system, Lpp are only anchored at the outer leaflet of the cytoplasmic membrane. However, the release of Lpp particularly in pathogenic or commensal species is crucial for immune modulation. Here, we provide evidence that in Staphylococcus aureus Asp at position +2 plays a role in withholding Lpp to the cytoplasmic membrane. Screening of published exoproteomic data of S. aureus revealed that Lpp mainly with Gly or Ser at position +2 were found in exoproteome, but there was no Lpp with Asp+2. The occurrence of Lpp with Asp+2 is infrequent in gram‐positive bacteria. In S. aureus USA300 only seven of the 67 Lpp possess Asp+2; among them five Lpp represented Lpl lipoproteins involved in host cell invasion. Our study demonstrated that replacing the Asp+2 present in Lpl8 with a Ser enhances its release into the supernatant. However, there is no different release of Asp+2 and Ser+2 in mprF mutant that lacks the positive charge of lysyl‐phosphatidylglycerol (Lys‐PG). Moreover, substitution of Ser+2 by Asp in SitC (MntC) did not lead to a decreased release indicating that in staphylococci positions +3 and +4 might also be important for a tighter anchoring of Lpp. Here, we show that Asp in position +2 and adjacent amino acids contribute in tightening the anchoring of Lpp by interaction of the negative charged Asp with the positive charged Lys‐PG.

group from phosphatidylglycerol to the prolipoprotein (Sankaran & Wu, 1994), Lsp recognizes the diacylglyceryl modification and cleaves between the amino acid at position -1 and the lipid-modified cysteine (+1) residue (Hussain, Ichihara, & Mizushima, 1982), and Lnt carries out the N-acylation of the free N-terminus of cysteine to form N-acyl diacylglyceryl cysteine (Gan et al., 1995). After the modification, the Lpp of gram-negative bacteria either stay at the inner membrane or are exported to the outer membrane through the Lol (localization of lipoprotein) system (Narita, Matsuyama, & Tokuda, 2004;. However, Lpp that contain aspartate at position +2 (Asp+2) next to cysteine (+1) are hardly recognized by the Lol machinery. Therefore, these Lpp remain anchored at the inner membrane (Narita & Tokuda, 2011;Okuda & Tokuda, 2011). Thus in E. coli, the amino acid at position +2 determines the localization of Lpp either in inner or outer membrane (Yamaguchi, Yu, & Inouye, 1988). By in silico prediction it was assumed that the residues following the +1 cysteine may contain the Lpp sorting information (Zuckert, 2014).
While diacylated Lpp are sensed by the TLR2/TLR6 heterodimer, triacylated Lpp are sensed by the TLR2/TLR1 heterodimer (Jin et al., 2007;Kang et al., 2009;Schenk, Belisle, & Modlin, 2009;Takeda, Takeuchi, & Akira, 2002). As the Lpp receptors TLR2, TLR1, and TLR6 recognize Lpp via their ectodomain (Jimenez-Dalmaroni et al., 2015) it is assumed that soluble Lpp in the bacterial supernatant bind to their receptors. It means that certain Lpp are not permanently anchored in the outer leaflet of the cytoplasmic membrane but are released into the environment during growth. Indeed, it has been shown recently that S. aureus strains that produce the detergent-like phenol-soluble modulins (PSMs) (Cheung, Joo, Chatterjee, & Otto, 2014) release higher amounts of Lpp from the cytoplasmic membrane (Hanzelmann et al., 2016).
As the amount of released Lpp into the environment is correlated with the immune response we asked the question whether negatively charged amino acids (aa) next to the cysteine in position +1 strengthens the anchoring of Lpp at the membrane by ionic interaction with positive charged phospholipids. To strengthen the hypothesis that the negative charged Asp+2 interacts with a positive charged membrane phospholipid, we created a ΔmprF mutant in USA300. MprF lysinylates phosphatidylglycerol to Lys-PG, one of the dominant membrane phospholipids in S. aureus (Ernst et al., 2015;Peschel et al., 2001;Staubitz, Neumann, Schneider, Wiedemann, & Peschel, 2004). In the mprF mutant the positive lysyl group is absent and now we saw that in this mutant Lpl8 +2D is not more retained to the membrane than Lpl8 +2S. This result clearly indicates that the strong retention of Lpl8 +2D to the membrane is most likely due to its ionic interaction with the positive charged Lys-PG. Furthermore, we screened all known Lpp of S. aureus USA300 for the +1 to +3 amino acid of the mature Lpp and found that Lpp with glycine at position +2 (G +2 ) were the most abundant in the supernatant, whereas Lpp with an Asp at position +2 (D +2 ) were hardly found in the supernatant (Nguyen et al., 2015;Stoll, Dengjel, Nerz, & Götz, 2005).

| Bacterial strains and growth conditions
Bacterial strains and plasmids used in this study are listed in Table 1. S. aureus strains were grown aerobically in basic medium, BM (1% soy peptone, 0.5% yeast extract, 0.5% NaCl, 0.1% glucose, and 0.1% K 2 HPO 4 , pH 7.4) at 37°C. For strains containing the plasmid pCtuf, the media were supplemented with 10 μg/ml of chloramphenicol.

| Creation of a double mutant S. aureus USA300∆lpl∆spa::erm by phage transduction
In order to avoid unnecessary binding of IgG to protein A, the spa gene was deleted by phage transduction using Φ11 with SA113 spa::erm (Schlag et al., 2010) as donor strain; as a result strain USA300∆lpl∆spa::erm was created. Briefly, phage Ф11 was used to produce a phage lysate of SA113∆spa::erm. The lysate was filtered through a 0.2 μm pore-size filter and used to infect strain USA300∆lpl at a low multiplicity of infection (phage-to-recipient ratio of 1:10).
Transducants carrying ∆spa::erm were selected on tryptic soya agar (TSA) supplied with erythromycin 2.5 μg/ml. As a control, the phage lysate was plated alone to avoid reisolating the donor strain. Positive clones containing the spa deletion were confirmed by DNA sequencing. For that, genomic DNA was isolated from clones using Quick-gDNA™ Miniprep Kit (Zymo Research Europe GmbH) and used as template for PCR, using the primers For.spa.seq (5'-AAGACCATG CTGAACAATTATTAGCTCA-3') and Rev.spa.seq (5'-TGCAGGTGG TGTAGCAGCGAAAC-3'). The PCR products were purified using illustra GFX PCR DNA and Gel Band Purification Kits (GE healthcare) and send for sequencing (GATC Biotech AG) to confirm spa deletion.
All primers were purchased from integrated DNA technologies (Idt, Illinois).
Briefly, the mixture of purified PCR flanking fragments and linearized plasmid were mixed 2xHiFi DNA Assembly Master Mix (New England Biolabs Inc., UK) at ratio 1:1 and incubated at 50°C for 1 hr. Later the ligation mixture was transformed into chemo-competent E. coli DC10B and plated onto BM Ampicillin (100 μg/ml) plate and incubated at 37°C. The clones containing plasmid pBASE-mprF were screened using colony PCR and confirmed by DNA sequencing. The correct plasmid pBASE-mprF subsequently transformed by electroporation into USA300∆lpl∆spa::erm. The deletion procedure of mprF gene was followed as described previously (Bae & Schneewind, 2006).
The final gene deletion was checked and confirmed by PCR and DNA sequencing.

| Construction of pCtuf-Lpl8-strep and pCtuf-SitC-strep
To construct the plasmid pCtuf for expression of Lpl8, the lpl8 +2D sequence was amplified from S. aureus USA300 genomic DNA, using the primers: Forward primer F_lpl8 +2D (5'-AATATTT AATTAATGAAGTCTATAAAAAGGATTGGATTG-3') and Reverse and restriction site PacI in underline. The lpl8 +2S sequence was amplified with a pair of primers F_lpl8 +2S and R_lpl8-strep. The amplified fragments and plasmid pCtuf were cut using the same digestion enzymes (PacI and HindIII) and subsequently ligated together (T4 Rapid ligation Kit by Thermo Scientific). The ligation products were transformed into S. aureus RN4220 by electroporation to yield the plasmid pCtuf-lpl8 +2D and pCtuf-lpl8 +2S , which were then transformed into S.
Restriction sites PacI and HindIII are underlined, respectively, and in the reversed primer strep-tag nucleotide sequence was added (bold letters). To create SitC sequence with Asp+2, the glycine at the 19th codon was replaced by the aspartate codon (G19D) using the forward primer F_SitC +2D (5'-ATATTAATTAATGAAAAAATTAGTACCTTTA TTATTAGCCTTATTACTTCTAGTTGCTGCATGTGATACTG-3') with the substituted nucleotide sequence (bold letters) and restriction site PacI underlined. The sitC +2D sequence was amplified with a pair of primers F_SitC +2D and R_SitC-strep. Two resulting plasmid pCtuf-SitC +2G and pCtuf-SitC +2D were cloned into S. aureus RN4220 and later in USA300Δlpl∆spa::erm (Figure 1c and d). The strain carrying the pCtuf plasmid without inserted fragment was used as negative control.

| Harvesting of extracellular proteins
For the detection of extracellular proteins S. aureus clones were grown aerobically in BM with 10 μg/ml chloramphenicol at 37°C. Overnight T A B L E 1 Strains and plasmids used in this study

| Western blot
For Western blot analysis, the supernatants were thawed on ice and Membranes were incubated with primary antibody (anti-strep-tag rabbit IgG, Abcam) for 1 hr and subsequently with secondary antibodies (anti-rabbit IgG goat IgG alkaline phosphatase conjugated, Sigma) for 1 hr each at room temperature under gentle shaking. Detection was carried out using BCI/NBP (Sigma, Munich); blots were scanned by Epson scanner. with Asp+2 was very low ranging from 0% to 6% per species/strain (Table S1).

| The lipoproteins with Asp at position +2 are not frequent in gram-positive bacteria
To illustrate the low distribution of Lpp with Asp+2 we show a selection of gram-positive species representative ( Figure 2). As can be seen in the average the occurrence of Lpp with (Asp+2) ranged from 0% to 6%. Exceptionally high percentage (7 Lpp ≈ 10%) was the num- S. aureus USA300 as a prototype of pathogenic S. aureus strains we found that among the 67 Lpp the most abundant amino acid at position +2 was glycine (39 Lpp), followed by serine (12 Lpp), aspartate (7 Lpp), alanine (3 Lpp), and each one with arginine, threonine, glutamine, asparagine, glutamate, and valine (Table 2).

| Evaluation of S. aureus Lpp detected in exoproteome
Next we analyzed the published exoproteomic data of S. aureus to see which Lpp with amino acid in position +2 are preferentially found in the supernatant. It turned out that in the supernatant particularly Lpp with uncharged or positively charged aa in position 2 were found: such as glycine, serine, alanine, arginine, and threonine. Lpp with aspartate in position 2 (Asp+2) were not detected in the exoproteome of various S. aureus strains. The results are summarized in Table 2, and the full detailed list of Lpp is shown in Table 3.

| The positive charged lysyl-phosphatidylglycerol (Lys-PG) enhances Lpl8 retention
In S. aureus, most of the Lpp containing Asp+2 belong to the Lpl lipoproteins. Previously it has been demonstrated that the Lpl lipoproteins F I G U R E 1 Schematic representation of plasmid constructs used in the study for lipoprotein release in S. aureus. The initial plasmid pCtuf were inserted with either (a) lpl8 +2D , (b) lpl8 +2S , (c) SitC +2G , and (d) SitC +2D , all with c-terminal strep-tag sequence (See methods for detail). The expression of Lpl8 and SitC variant were under the constitutive control of elongation tufA promoter are expressed both at transcriptional level and that they are associated with the cell envelope; their biological function is that they contribute to host cell invasion (Nguyen et al., 2015;. To evaluate the effect of Asp+2 on Lpl release we substituted in Lpl8 Lpl8 +2D and Lpl8 +2S . However, after 8 and 12 hr Lpl8 +2S was much more released than Lpl8 +2D (Figure 3b). The release of Lpl8 +2S occurred mainly in the late exponential and stationary growth phase.

SAUSA300_1884
CamS sex pheromone biosynthesis

C GNHK
No TMD

C GKKE
No TMD

C GNKE
No TMD

C STTN
No TMD

MRSA clinical isolates
Detected (Herbert, Ziebandt et al., 2010) 63. Detected (Becher et al., 2009;Hanzelmann et al., 2016) T A B L E 3 (Continued) supplied with C-terminal strep-tag were constitutively expressed on pCtuf vector in S. aureus USA300∆lpl∆spa::erm: the original SitC with glycine at position +2 (SitC +2G ) and the mutated SitC with aspartate at position +2 (SitC +2D ) (Figure 1c and d). The release of SitC +2G and SitC +2D into the supernatant was monitored by Western blot in the same way as described for the Lpl8 variants. The protein samples were adjusted to the same A280/260 values to obtain an equal amount of extracellular proteins (Figure 4a). In 4-hr culture supernatant, neither SitC +2G nor SitC +2D were detected in the Western blot, but after 8-hr and 12-hr cultivation SitC was clearly detected (Figure 4b). However, there was no difference observed in the release of both SitC variants.

| DISCUSSION
Lipoproteins (Lpp) in gram-positive bacteria have two major functions. They play an important role in physiology by being involved in transport of diverse nutrients, by acting as chaperons, by being involved in respiration, or by contributing to host cell invasion Shahmirzadi et al., 2016). These functions are exerted by the protein part of the individual Lpp. Their second function is related to their interaction with the host's immune system. Here, it is not so much the protein part, but the lipid moiety plays the crucial role as potent activators of the innate and adaptive immune response by interacting with TLR2/1 or TLR2/6 receptors (Schenk et al., 2009;Takeda et al., 2002). Some Lpp were also considered as vaccine candidates. For example, immunization with FhuD2, a Lpp involved in ferric-hydroxamate uptake, alone or together with hydroxamate siderophores, were protective in a murine staphylococcal infection model (Mishra et al., 2012). Recently, the combination of five antigens provided close to 100% protection against four different S. aureus strains (Bagnoli et al., 2015). Among them, were two Lpp, FhuD2 and the conserved staphylococcal antigen 1A (Csa1A).
Lpp in gram-positive bacteria anchored in the outer leaflet of the cytoplasmic membrane. However, for the interaction with TLR2 receptor they must be released from the membrane to be able to expose We therefore asked the question how aa next to the invariable cysteine in position +1 contributes to the holding of Lpp to the membrane.
Here, we investigated the effect of aa at position +2 of the S. aureus Lpp in their release into the environment. In E. coli the significance of the aa at position +2 in withholding Lpp to the cytoplasmic membrane was well documented (Narita & Tokuda, 2016;Okuda & Tokuda, 2011). In several studies it has been reported that Asp+2 facilitates anchoring of Lpp at the inner membrane because Asp+2 functions as a Lol avoidance signal (Poquet, Kornacker, & Pugsley, 1993;Seydel, Gounon, & Pugsley, 1999;Terada, Kuroda, Matsuyama, & Tokuda, 2001;Yamaguchi et al., 1988). On the other hand, the outer membrane-specific Lpp stimulated ATP hydrolysis by LolCDE but not the inner membrane-specific Lpp (Masuda, Matsuyama, & Tokuda, 2002). The Lol avoidance mechanism was based on the strength of the hydrogen bonds between the negative charged Asp+2 and the positive charged phosphatidyl ethanolamine (PE) of the membrane phospholipids; glutamate at +2 position, with its longer side chain, interacts differently with PE (Hara, Matsuyama, & Tokuda, 2003). Therefore, the formation of a tight Lpp-PE complex causes the Lol avoidance signal.
Like in E. coli there is in S. aureus also a predominant positive charged phospholipid, the Lys-PG (Gould & Lennarz, 1970), which is synthesized by MprF (Ernst et al., 2015;Peschel et al., 2001). Lys-PG, which yields 20%-40% of staphylococcus total membrane phospholipids, causes resistance against cationic antimicrobial compounds through ionic repulsion (Slavetinsky, Kuhn, & Peschel, 2016). Given that, gram-positive bacteria lack the Lol system it is still possible that Asp+2 strengthens the anchoring of the corresponding Lpp to the membrane via the interaction with the positively charged Lys-PG. Indeed, that is the case as in a mprF mutant there was no difference in retention between Asp+2 or Ser+2 in our model Lpl8.
Screening of Lpp of gram-positive bacteria for Asp+2 revealed that this aa is very rare at this position, and, if at all, it is mainly found in pathogenic species/strains of Bacillus, Clostridium, Mycobacterium, Staphylococcus aureus, and some streptococci (Table S1). To verify the question, we screened 13 publications containing exoproteomic data of S. aureus and found out that none of seven Lpp with Asp+2 were detected in supernatant (Table 2 and 3). These seven Lpp with Asp+2 consist of; SAUSA300_0175 with a putative function as nitrate ABC transporter substrate-binding protein, YidC (OxaI), and 5 Lpl lipoproteins. YidC, is an essential Lpp in bacteria acting as a membrane protein translocase and chaperone for membrane protein folding (Kuhn & Kiefer, 2017). Lpl lipoproteins contribute to S. aureus invasion to the host cells and G2/M transition delay (Nguyen et al., 2015;. We assume that for the function of these seven Lpp with Asp+2, a tighter anchoring to the cytoplasmic membrane is necessary for their function. Particularly the epidemic S. aureus strains, such as USA300, which contain seven Lpp with Asp+2. Five of these Lpp are encoded in the lpl operon of the νSAα genomic island (Diep et al., 2006). This lpl operon contributes to host cell invasion via the protruding protein part (Nguyen et al., 2015;. It makes sense that these Lpl proteins are especially tightly anchored to the cytoplasmic membrane to facilitate host cell invasion by interacting with the proposed target molecule at the host cell surface. To experimentally verify the role of Asp+2 in retaining of Lpp at the cytoplasmic membrane we substituted the Lpl8 +2D by Lpl8 +2S . We have chosen Ser as many Lpp with Gly or Ser in position +2 are found in the exoproteome (Table 2). However, when we substituted SitC +2G by SitC +2D ; there was no difference in release into the supernatant ( Figure 4b). Apparently, aa in positions downstream of +2 might play a role. In Lpl8 for instance there is in position +4 a second aspartate C-DGDN, whereas in SitC (MntC) C-GTGG there is no Asp (Table 3).
These results suggest that aspartate in position +2 plays a role but is not sufficient to withhold Lpp tightly at the cytoplasmic membrane and aa in position +3 and +4 might also contribute to this function.
In E. coli it has been shown that besides the +2, aa at position +3 also contributes in sorting of Lpp to the inner or outer membrane. Glu, Asp, Gln, or Asn at +3 position enhanced the retention to the inner membrane, whereas His, Lys, Val, Ile, Ala, Cys, or Thr decreased it (Terada et al., 2001). In Pseudomonas aeruginosa it was shown that Lys and Ser at positions +3 and +4 play a critical role for retaining Lpp in the inner membrane (Narita & Tokuda, 2007) (Lewenza, Mhlanga, & Pugsley, 2008). In P. aeruginosa Lpp that are located in the inner membrane have Gly+2 followed by Asp/Glu +3 (Remans, Vercammen, Bodilis, & Cornelis, 2010). Recently, it has been shown that aa variations in +2 of the alkaline phosphatase (PhoA) expressed in Mycoplasma gallisepticum did not affect the retention of PhoA to the membrane (Panicker, Kanci, Markham, & Browning, 2016). Finally, in Bacillus subtilis it was found that Gly+2 facilitates release of Lpp while a Ser+2 favors withholding in the membrane (Tjalsma & van Dijl, 2005).

| CONCLUSION
An evaluation of literature data shows that in S. aureus the majority (75%) of Lpp found in the exoproteome carry Gly or Ser at position +2, whereas no Lpp with Asp in position +2 was found in the exoproteome.
F I G U R E 5 Model for the enforced interaction of Lpp Asp+2 (negative charged) with the Lys-PG (positive charged). Phosphoglycerol (PG) in blue, Lys-PG in red, D is Aspartate, C is Cysteine The role of Asp+2 in withholding Lpp to the cytoplasmic membrane was also confirmed by Lpl8 +2D and Lpl8 +2S in wild type but not in the mprF mutant. This suggests that the negative charged Asp withholds Lpp at the membrane by interacting with the positive charged Lys-PG ( Figure 5). On the other hand, substitution of SitC +2G by SitC +2D did not lead to a decreased release into the supernatant, suggesting that in staphylococci positions +3 to +5 might also be important for a more tightly anchoring of Lpp in the membrane. In gram-positive bacteria the release of Lpp into the supernatant is crucial for the immune modulation via TLR2 activation thus contributing to inflammation and infection. On the other hand, tightly anchored Lpp such as Lpls' or YidC is prerequisite for their function in host cell invasion and membrane protein sorting. Therefore, finding out sequence motives that modulate the strength of membrane anchoring is important.