The type VI secretion system protein AsaA in Acinetobacter baumannii is a periplasmic protein physically interacting with TssM and required for T6SS assembly

Type VI secretion system (T6SS) is described as a macromolecular secretion machine that is utilized for bacterial competition. The gene clusters encoding T6SS are composed of core tss genes and tag genes. However, the clusters differ greatly in different pathogens due to the great changes accumulated during the long-term evolution. In this work, we identified a novel hypothetical periplasmic protein designated as AsaA which is encoded by the first gene of the T6SS cluster in the genus Acinetobacter. By constructing asaA mutant, we delineated its relative contributions to bacterial competition and secretion of T6SS effector Hcp. Subsequently, we studied the localization of AsaA and potential proteins that may have interactions with AsaA. Our results showed that AsaA in Acinetobacter baumannii (A. baumannii) localized in the bacterial periplasmic space. Results based on bacterial two-hybrid system and protein pull-down assays indicated that it was most likely to affect the assembly or stability of T6SS by interacting with the T6SS core protein TssM. Collectively, our findings of AsaA is most likely a key step in understanding of the T6SS functions in A. baumannii.


Results
AsaA is Acinetobacter specific and highly conserved in the species A. baumannii. Previous studies have described the gene cluster of T6SS in A. baumannii 12,15 . The cluster is about 20 kb, consisting of more than 20 T6SS genes. asaA is the first gene of the T6SS cluster. In this work, we used strain ATCC 17978 as a representative of A. baumannii. In a BLAST comparison of the amino acid sequences of AsaA homologues, AsaA only present in the genus Acinetobacter, and the homology of the AsaA amino acid sequence differs between different species. The AsaA of A. baumannii shares 87. .6% similarity with that in the species A. pittii, A. calcoaceticus and A. nosocomialis, with a much more lower similarity 46.8% and 57.8% with A. baylyi and A. indicus. In the species A. baumannii, it is highly conserved among different strains with different sequence types. The amino acid sequence of AsaA in ATCC 17978 shares 99.1-100% similarity with the other A. baumannii strains (Table S3).
AsaA is required for bacterial competition. asaA was knocked out from the A. baumannii ATCC 17978 chromosome and the mutant was named ΔasaA. Complementation using pTrc99A which is a useful vector for the expression of unfused and fused proteins in E. coli was performed to generate CΔasaA 16 . To investigate whether asaA knock-out would affect the growth of A. baumannii, the ΔasaA, CΔasaA and wild type (WT) 17978 were tested for the growth rates in LB. As shown in Supplementary Fig. 1, there were no significant difference between different strains. So knocking out asaA did not affect the growth of A. baumannii in LB.
A recent study found that A. baumannii ATCC 17978 can utilize its T6SS to compete with E. coli 15 . To determine whether AsaA is required for the competition, we used E. coli strain JM109/pK18mob, which contained a kanamycin resistant plasmid pK18mob, as a target for bacterial competition assays. TssM is a core component of T6SS. Previous study has shown deletion of tssM in A. baumannii can completely abolish its ability to outcompete E. coli 15 . So we used ΔtssM mutant as a negative control. WT 17978, ΔasaA, CΔasaA or ΔtssM were incubated with E. coli and surviving E. coli were calculated. The results revealed that there were significant differences among the competition ability of these bacteria. Compared with WT 17978 and CΔasaA, the number of surviving E. coli is much higher in both ΔasaA and ΔtssM groups than in WT 17978 and CΔasaA groups, suggesting that deletion of asaA resulted in a decreased competitive ability in WT 17978 (Fig. 1A). Quantitatively, bacterial mixtures were plated onto kanamycin resistant and non-resistant plates to culture E. coli and total bacteria. As shown in Fig. 1B, similar to ΔtssM, ΔasaA mutant has a low ability to kill E. coli and there were no significant differences in the survival of E. coli strains between ΔtssM and ΔasaA mutants. In contrast, the survival of E. coli strains in the WT 17978 and CΔasaA groups decreased significantly (Fig. 1B). At the same time, the survival of A. baumannii www.nature.com/scientificreports www.nature.com/scientificreports/ were not influenced by E. coli and there were no significant differences in the survival of A. baumannii among different groups (Fig. 1C). These results suggest that AsaA is critical for bacterial competition in A. baumannii ATCC 17978.
AsaA is required for Hcp secretion. As Hcp secretion reflects the functionality of the T6SS, detection of secreted Hcp in the culture supernatants is a marker of a functional T6SS. In order to explore the mechanisms of AsaA involved in bacterial competition, we examined whether AsaA is involved in the secretion of Hcp. RT-PCR analysis showed that hcp were being actively transcribed in the ΔasaA mutant cultured in LB medium ( Fig. 2A). To further investigate whether AsaA is involved in Hcp secretion, we performed Western blot assays to test the secretion of Hcp in the ΔasaA background. For this purpose, the recombinant plasmid pTHcpH6 containing the hcp gene and the 6× His-tag coding sequence, was introduced into different strains, resulting in the strains named 17978/pTHcpH6, ΔasaA/pTHcpH6 and ΔtssM/pTHcpH6, respectively (Table S1). Hcp were present in the whole cell lysates in all the three strains ( Fig. 2B), at the same time, it is present in the supernatant of 17978/ pTHcpH6 but absent in ΔasaA/pTHcpH6 or ΔtssM/pTHcpH6, suggesting that deletion of asaA affected the secretion of Hcp (Fig. 2B). These results indicated that AsaA is critical for Hcp secretion.
AsaA located in the periplasm of A. baumannii cells. The protein encoded by asaA in A. baumannii ATCC 17978 was annotated as conserved hypothetical protein 17 . It has 230 amino acids (aa). We performed bioinformatics analysis on AsaA and failed to find additional information about its biochemical function and any known domain or motif. According to the amino acid sequence analysis of the SignalP program (http://www.cbs. dtu.dk/services/SignalP/) and the LipoP program (http://www.cbs.dtu.dk/services/LipoP/), AsaA has a signal fragment and it may be a secreted protein, but not a lipoprotein. This suggested that AsaA might be located in the periplasm or secreted outside the cells. To determine the cellular location of AsaA, we constructed a ΔasaA/ pTasaAH6 strain, which expressed AsaA with a 6× His-tag. The total cell lysates, inter membrane, periplasmic, outer membrane and extracellular protein fractions of ΔasaA/pTasaAH6 cultured at logarithmic growth phase were prepared. As shown in Fig. 3, using the inner membrane protein PglC, the periplasm protein DsbA, the outer membrane protein OmpA, the extracellular protein Hcp, and the cytoplasm protein Dnak as controls, we found that AsaA is present in the periplasmic and total cell lysates but absent in the extracellular, inter membrane or outer membrane protein fractions, indicating that AsaA located in the periplasm of A. baumannii.
AsaA interacts with TssM. Based on the facts that AsaA located in the bacterial periplasm and takes part in Hcp secretion but not involved in Hcp expression, we presumed that AsaA might physically interacts with the T6SS components. To this end, we performed bacterial two-hybrid system to investigate the possibility of the physically interactions between AsaA and a series of the T6SS core proteins: TssB, TssL and TssM. In ATCC 17978, the gene encoding TssM was divided into two open reading frames, A1S_1302 and A1S_1303 18 . A1S_1302 encodes the N-terminus and the middle part of TssM (we named this part TssM 1302 ). A1S_1303 encodes the C-terminus of TssM (we named this part TssM 1303 ). In this study, we tested the interactions between AsaA with TssM 1302 (    and TssM 1302(436-1041) were fused to pTRG, yielding recombinant plasmids named as pTB, pTL, pTM 1303 , pTM 1302  and pTM 1302(436-1041) , respectively. Finally, the plasmids were co-introduced into XL1-Blue MRFʹ strain. For confirmation, all the resulting recombinant strains were tested on dual-selective medium containing 3-AT and streptomycin. If interactions between the proteins occur, the recombinant strains can obtain the ability to grow on dual-selective medium due to the activation of HIS3-addA reporter genes. Our results showed that only the X/pBA-pTM 1302(436-1041) can grow well on dual-selective medium, in contrast, the other recombinant strains can't grow on dual-selective medium (Fig. 4A), implying that physical interactions existed between AsaA and the periplasmic domain of TssM in the reporter strain.
To validate the interaction between AsaA and TssM 1302(436-1041) , the pull-down biotinylated protein-protein assay was further performed. The AsaA  (from the 25 th to 230 th aa) and TssM 1302(436-1041) encoding sequences were cloned into the expression vector pET-30a to produce the recombinant His-tagged proteins AsaA-His 6 and TssM-His 6 (Fig. 4B, full-length gel was presented in Supplementary Fig. 2). AsaA-His 6 was immobilized on sepharose beads, and then pull-down assays for TssM-His 6 was performed. As shown in Fig. 4C (full-length gel was presented in Supplementary Fig. 3), the AsaA-His 6 protein was able to capture TssM-His 6 protein (lane 1). Vice versa, TssM-His 6 protein immobilized on streptavidin sepharose beads was also able to capture AsaA-His 6 protein (lane 3). So, AsaA-His 6 and TssM-His 6 can capture each other. The results indicated that AsaA interacts directly with TssM 1302(436-1041) .
We then tried to find the exact domains of AsaA that interacts with TssM 1302(436-1041) . Since the first 24 amino acids of AsaA encode a signal peptide, we used the truncated AsaA excluding the N-terminal signal peptide sequence for testing. DNA fragments as shown in Fig. 5A encoding the different domains of AsaA (the 25 th -70 th aa, 25 th -110 th aa, 25 th -150 th aa, 25 th -190 th aa, 70 th -230 th aa, 110 th -230 th aa, 150 th -230 th aa and 190 th -230 th aa) were amplified and fused to pBT, respectively. Then, the recombinant plasmids and pTM 1302(436-1041) were co-introduced into XL1-Blue MRFʹ strain, respectively. All the resulting recombinant strains were tested on dual-selective medium. Only the strains which harbored the plasmid pBA 25-150 -pTM 1302(436-1041) , pBA 25-190 -pTM 130225-190 -pTM (436-1041 and pBA 70-230 -pTM 1302(436-1041) could grow on the dual-selective medium but others could not (Fig. 5B). This indicated that the 70 th -150 th aa of AsaA were crucial for the interaction. We further tested the interactions of the 70 th -150 th aa, 70 th -110 th aa, and 110 th -150 th aa of AsaA with TssM 1302(436-1041) . Only the strain containing pBA 70-150 -pTM 1302(436-1041) could grow on the selective agar plates as shown in Fig. 5B. The pull-down biotinylated protein-protein assay was performed for further validation. As shown in Fig. 5C, TssM-His 6 protein was able to capture AsaA 70-150 -His 6 protein (lane 1). Vice versa, AsaA 70-150 -His 6 was also able to capture TssM-His 6 protein (lane 3). Our results indicated that the peptide of AsaA consisting of aa from the 70 th to the 150 th is sufficient to interact with the periplasmic domain of TssM. www.nature.com/scientificreports www.nature.com/scientificreports/

Discussion
Our knowledge about the molecular architecture and the function of T6SS has great strides. For A. baumannii, several studies have been performed to elucidate its functions. It has been proven that the T6SS of A. baumannii is responsible for bacterial competition and is also implicated in host colonization 12,15,19-21 . However, for the www.nature.com/scientificreports www.nature.com/scientificreports/ functions of the proteins in T6SS, only a few were approved experimentally, including VgrG which contributes to both virulence and antimicrobial resistance in A. baumannii ATCC 19606 21 and apeptidoglycan hydrolase, named TagX, which is conserved in the genus Acinetobacter and is essential for Hcp secretion 12 . There were still a lot need to be elucidated about the T6SS in A. baumannii.
In this study, we made a detailed study on the function of AsaA. AsaA is encoded by the first gene within the T6SS cluster. By comparative bioinformatics analysis, we found AsaA is specific to the genus Acinetobacter and is highly conserved in the species A. baumannii. Studies on the function of the highly conserved and Acinetobacter specific T6SS protein AsaA can help us learn more about the pathogenesis of A. baumannii. We used ATCC 17978 as a representative of A. baumannii to study whether AsaA is critical for bacterial competition. The competition ability of ATCC 17978 varied with different prey cells. Guillermo et al. showed ATCC 17978 can compete against www.nature.com/scientificreports www.nature.com/scientificreports/ E. coli DH5α 15 , while Weber et al. showed that ATCC 17978 is unable to utilize its T6SS for antibacterial activity against E. coli MG1655 18 . We postulated that different E. coli strains may have different immunity proteins and thus resulted in different competition abilities. In this study, we found ATCC 17978 can well compete E. coli JM109 and AsaA is critical for the competition.
Our results demonstrated that AsaA localizes in the periplasm, where it interacts with the periplasmic domain of TssM by a 80 aa peptide from the 70 th to 150 th aa, forming a two protein complex. AsaA is not participated in the regulation of the Hcp expression but is essential for Hcp secretion in A. baumannii ATCC 17978. At this stage, AsaA played an important role in the T6SS, but its precise roles remain to be determined. However, based on the facts that (1) mutants of AsaA and TssM showed similar phenotypes in bacterial competition and Hcp secretion; (2) similar to tssM, asaA is within the T6SS gene cluster; and (3) the TssM 1302(436-1041) extends into the periplasm and interact with AsaA physically, we presume that AsaA is most likely to be a structural component of the T6SS. The interaction with TssM may consolidate the structure of T6SS and thus facilitate the transportation of effectors.
The T6SS structural components are encoded by T6SS cluster, which are probably acquired by horizontal gene transfer during evolution. It has been shown that A. baumannii T6SS clusters appear to be conserved amongst sequenced A. baumannii strains with the exception that a portion of the gene cluster is inverted 15 . However, the clusters have undergoing great changes during the long-term evolution. A main difference from other bacteria is that the genus Acinetobacter do not produce a readily identifiable homolog of core-components TssJ 18 , which has been shown to interact with the extreme C-terminus of TssM. Whether AsaA can replace TssJ is of interest. In this study, we found that: (1) AsaA does not interact with TssM 1303 , which encode the C-terminus of TssM, but interact with TssM 1302, which encode the N-terminus and the middle part of TssM; (2) AsaA localizes in periplasm, while TssJ localizes in outer membrane; and (3) secondary structure of AsaA is predicted to be alpha helix, while TssJ is mostly beta-folded. Base on these facts, we presume that AsaA is unlikely a replacement protein for TssJ 22 .
Contact-dependent growth inhibition (CDI) system are alternative war machines doing the same job as T6SS, i.e. fight against neighbouring cells and oucompeting them. They are found on the cell surface of both A. baumannii and A. baylyi ADP1 cells [23][24][25] . Only a relatively few STs host CDI system in A. baumannii, the same holds for T6SS cluster and asaA gene, whether there is a correlation between the presence/absence of T6SS and CDI system need further investigation.
In conclusion, we confirmed that AsaA is required for the secretion of Hcp and it most likely affects the assembly or stability of the T6SS by interacting with the periplasmic domain of the core T6SS protein TssM in the periplasmic space. We cannot exclude the possibility that AsaA associates with assisting the apparatus assembling or affecting the apparatus stability rather than as a T6SS structural component. Nonetheless, given the fact that AsaA is an Acinetobacter genus-specific protein, the results suggest that the T6SS structural components of Acinetobacter is distinctive from other Gram-negative pathogens.

Methods
Strains and culture conditions. All strains and plasmids are summarized in Table S1. The reporter strains were grown in M9 His-dropout medium, other strains were grown in LB medium. Antibiotics, when appropriate, were added to bacterial cultures medium at the following final concentrations in μg/mL: chloramphenicol (Cm) at 12.5, streptomycin (Sm) at 10, ampicillin (Amp) at 100, tetracycline (Tc) at 20, and kanamycin (Kan) at 50.

Construction of mutant and complement strains. All primers used in this study are summarized in
Tables S2. asaA and tssM gene knock out mutants of A. baumannii ATCC 17978 were constructed with the method described by Tracy and associates 26 , with modifications. Briefly, a kanamycin resistance cassette was amplified with a pair of 76 bp oligonucleotide primers with 56 nucleotides of homology to the flanking regions of the targeted gene and an additional 20 nucleotides of homology to kanamycin resistance cassette from plasmid pKD4 26 . Subsequently, the PCR products were electroporated into A. baumannii ATCC 17978 carrying pKD46 plasmid, then clones were screened on kanamycin containing LB agar plates 12 . Genomic DNA was isolated from kanamycin resistant clones, amplicons were prepared using primer set 2 and confirmed by sequencing. To remove the kanamycin resistance cassette, pCP20 expressing recombinase was transformed into the mutants. The asaA and tssM gene knock out mutants were confirmed by polymerase chain reaction (PCR) and sequencing.
For genetic complementation of mutant strains, the ORF (open reading frame) sequence of asaA (gene number: A1S_1292) and tssM (gene number: A1S_1302) was amplified from A. baumannii ATCC 17978. Subsequently, the PCR products were cloned into pTrc99A with confirmation by sequencing.
Bacterial growth assays. Bacterial growth assays was performed as previously described by Salomon and associates 27 . Triplicates of A. baumannii strains grown overnight in LB were normalized to OD 600 of 0.01, and the growth was monitored in LB of cultures incubated at 37 °C. Experiments were repeated at least twice with similar results. A representative experiment was shown.
Bacterial competition assays. Bacterial competition assays was performed as previously described by Brent and associates 12 . The pk18mob plasmid was transformed into E. coli strain JM109, generating the strain named JM109/pK18mob. Cultures of A. baumannii and JM109/pK18mob were grown overnight, and the JM109/ pK18mob was washed three times with PBS to remove kanamycin 12 . Cultures were diluted to OD 600 of 1.0. Then 100 µl E. coli was mixed with 10 µl A. baumannii, and 10 µl of the mixture was spotted onto a LB agar plate. After 12 h incubation at 37 °C, spots were excised from the LB agar, the bacteria were diluted serially with 10-fold. Dilutions were plated onto kanamycin containing LB agar plates to select for E. coli and onto non-antibiotic www.nature.com/scientificreports www.nature.com/scientificreports/ resistant LB agar plates to determine the number of total bacteria 20 . Experiments were repeated at least twice with similar results. A representative experiment was shown.

RNA isolation and semi-quantitative reverse transcription PCR (semi-quantitative RT-PCR).
All strains were grown to mid-logarithmic phase and total RNA was extracted using the RNAprep pure (Tiangen) and treated with RNase-free DNase according to the manufacturer's protocol (Tiangen) 28 . Extracted total RNA was frozen in RNase-free water at −70 °C. RNA purity and concentration were determined using gel electrophoresis and spectrophotometer (NanoDrop, Thermo Scientific) 29 . Semi-quantitative RT-PCR was performed according to the One Step RNA PCR Kit instructions (Takara). Two sets of primers against hcp (A1S_1296) and 16S rRNA (A1S_2837) were designed using Vector NTI. For all primer sets, the following cycling parameters were used: 50 °C for 30 min, 94 °C for 2 min followed by 20 cycles of 94 °C for 30 sec, 55 °C for 30 sec, 72 °C for 40 sec. For standardization of results, the relative abundance of 16S rRNA was used as the internal standard 30 .
Hcp secretion assays. pTrc99A-hcp recombinant vector expressing His-tagged Hcp was transformed into A. baumannii ATCC 17978. Cultures were grown in LB with ampicillin to exponential phase. Expression of Hcp was induced by 1 mM IPTG for 2 h. Then 10 ml cultures were centrifuged at 10 000 × g for 10 min to obtain whole cell pellets. The supernatant containing secreted Hcp protein were filtered through a 0.22 μm pore filter. A mixture of 8.5 ml of supernatant and 1.5 ml of 100% trichloroacetic acid (TCA) was placed on ice for 4 h and centrifuged at 10 000 × g for 20 minutes at 4 °C. Precipitated proteins were washed with 100% acetone and then re-suspended in 5 ml PBS. Subsequently, protein concentrations were measured using a BCA protein assay kit. Equal amounts of total protein (25 μg) were used for polyacrylamide gel electrophoresis and proteins were analyzed by Western blot 31 . Experiments were repeated at least twice with similar results. A representative experiment was shown.
Cellular localization of AsaA. The cell fractionations were prepared as described by Zang et al. 32 and Deng et al. 33 , with minor modifications. Briefly, overnight cultures of the cells were re-cultured with 1:10 and grown until an OD 600 of 0.5 was reached. And then, the cells were collected by centrifugation at 4 000 × g for 10 min and the supernatant was filtered through a 0.22 μm pore filter and reserved as the extracellular proteins. The cell pellets were washed three times and re-suspended in 1 ml of periplasting buffer (30 000 U of lysozyme, 20% sucrose, 1 mM EDTA) and incubated on ice for 10 min. After centrifugation, the pellet was re-suspended in 10 mM magnesium chloride and incubated at 30 °C for 5 min, and then incubated at 0 °C for 10 min. The sample was freeze-thawed and pelleted by centrifugation, and then the supernatant was reserved as the periplasmic proteins. The cell was disrupted by sonication and further centrifuged at 11 000 × g for 30 min. The supernatant was then centrifuged at 130 000 × g for 1.5 h to isolate total membranes. The pellet was suspended in 0.25% (w/v) sodium lauryl sarcosine and centrifuged at 130 000 × g for 1.5 h. The supernatant was reserved as the inner membrane proteins. The pellet was re-suspended in TM buffer (8 mM MgSO 4 , 10 mM Tris) and reserved as the outer membrane proteins. All samples were analyzed by Western blot analysis by probing with antibodies to the 6× His tag (Hcp, AsaA, OmpA, DsbA and PglC) and DnaK 34 .
Western blot. Proteins were resolved onto PVDF membranes. Membranes were blocked in 5% non-fat milk in 1× TBST buffer for 1 h at room temperature. Membranes were incubated overnight at 4 °C with anti-His 6 or anti-DanK at a 1:1000 dilution. After washed with 1× TBST buffer for five times, the membrane was incubated with a HRP-conjugated secondary antibody for 50 min at room temperature. Membranes were then washed five times in 1× TBST buffer. Hybridizing bands were detected using the ECL kit.
Bacterial two-hybrid assays. Bacterial two-hybrid assays in vivo were tested using the BacterioMatch ® II two-hybrid system (Stratagene). The truncated asaA, tssM and other genes were amplified by PCR with corresponding primers in Table S2. The amplified asaA gene excluding the N-terminal leader was fused to the vector pBT, yielding the plasmid pBA. The amplified tssM and other genes and their truncated derivatives were individually fused to the vector pTRG, respectively, yielding the plasmids pTM 1302  , pTM 1302(436-1041) , pTB, pTM 1303 and pTL (Table S1).
In order to explore the interaction between AsaA and TssM, 120, 240, 360, 480 bp and other fragments containing partial asaA were fused to pBT, respectively. Pairs of plasmids were co-transformed into XL1-Blue MRF′, and the transformants were cultured on M9 His-dropout medium plate containing Sm (12.5 μg/ml) and 3-AT (5 mM) at 30 °C for 24 h.

Overproduce and purification of recombinant protein.
To overproduce the truncated peptides of AsaA and TssM, fragments coding sequence were fused to pET-30a. Recombinant plasmids were transformed into BL21(DE3) cell (Table S1). The recombinant strains were grown to an OD 600 of 0.4, and then 1 mM IPTG was added to induce the expression of individual proteins. Cells were harvested and washed with PBS buffer, and broken by sonication. The recombinant proteins were purified by Ni-NTA His-bind ® (Novagen), as described in the manual.
Pull-Down assays. To analyze the interaction between AsaA with TssM, pull-down assays were tested using the ProFound TM pull-down biotinylated protein-protein interaction kit (Pierce). As described by li and associates 35 , the AsaA protein was biotinylated by sulfo-NHS-LC-biotin. Subsequently, the biotinylated AsaA was incubated with streptavidin sepharose TM beads. The beads were washed four times and sample containing 60 µg of TssM protein was added and incubated. The beads were washed four times with washing buffer. Subsequently, the final protein was eluted by elution buffer. Finally, the final protein was separated by electrophoresis followed by coomassie blue staining.