Glycosphingolipids Recognized by Acinetobacter baumannii.

Acinetobacter baumannii is an opportunistic bacterial pathogen associated with hospital-acquired infections, including pneumonia, meningitis, bacteremia, urinary tract infection, and wound infections. Recognition of host cell surface carbohydrates plays a crucial role in adhesion and enables microbes to colonize different host niches. Here the potential glycosphingolipid receptors of A. baumannii were examined by binding of 35S-labeled bacteria to glycosphingolipids on thin-layer chromatograms. Thereby a selective interaction with two non-acid glycosphingolipids of human and rabbit small intestine was found. The binding-active glycosphingolipids were isolated and, on the basis of mass spectrometry, identified as neolactotetraosylceramide (Galβ4GlcNAcβ3Galβ4Glcβ1Cer) and lactotetraosylceramide (Galβ3GlcNAcβ3Galβ4Glcβ1Cer). Further binding assays using reference glycosphingolipids showed that A. baumannii also bound to lactotriaosylceramide (GlcNAcβ3Galβ4Glcβ1Cer) demonstrating that GlcNAc was the basic element recognized. In addition, the bacteria occasionally bound to galactosylceramide, lactosylceramide with phytosphingosine and/or hydroxy fatty acids, isoglobotriaosylceramide, gangliotriaosylceramide, and gangliotetraosylceramide, in analogy with binding patterns that previously have been described for other bacteria classified as “lactosylceramide-binding”. Finally, by isolation and characterization of glycosphingolipids from human skin, the presence of neolactotetraosylceramide was demonstrated in this A. baumannii target tissue.


Introduction
Acinetobacter baumannii is emerging as a worldwide problem as a nosocomial pathogen in hospitalized patients. These bacteria primarily cause pneumonia, but they are also frequent causes of wound and burn infections, bacteremia, meningitis, urinary tract infections, and skin and soft tissue infections. The mortality associated with these infections is high. Isolates resistant to almost all available antimicrobials have been found, thus limiting treatment options. In fact, A. baumannii has been classified as highest priority on the recently published WHO list of pathogens needing research and development of new antibiotics [1].
The ability of A. baumannii to survive for an extended period of time on artificial surfaces allows it to persist in the hospital environment. This is due to its ability to form biofilms [2]. Biofilm formation in A. baumannii is phenotypically associated with exopolysaccharide production and pilus formation [3]. The Csu pili are required for biofilm formation in A. baumannii but do not play a role in adherence to human epithelial cells [4]. Recently, the X-ray structure of the CsuC-CsuE chaperone-adhesin complex demonstrated that the tip protein CsuE has three hydrophobic finger-like loops [5]. This unique structural feature mediates bacterial adhesion to abiotic substrates. Several different factors have been shown to be involved in the adherence of A. baumannii to human epithelial cells, as, e.g., the outer membrane protein A (OmpA), the biofilm-associated protein (BAP), the BAP-like proteins 1 and 2 (BLP-1 and BLP-2), the predicted pili subunit encoded by the LH92_11085 gene, and Acinetobacter trimeric autotransporter adhesin (Ata) [6][7][8][9][10]. Several of these factors (OmpA, BAP, BLP-1, BLP-2, and the LH92_11085 gene product) are also involved in biofilm formation, in line with the association found between biofilm production and human epithelial cell adherence [11,12].
Thus, a substantial amount of information about the A. baumannii factors involved in adherence to abiotic surfaces and epithelial cells is available. However, much less is known regarding the factors of epithelial cells that the bacteria bind to. Binding to specific receptors on the target cells allows microorganisms to colonize and cause infection and leads to an efficient delivery of virulence factors. The majority of microbial attachment sites on host cells and tissues identified are glycoconjugates [13,14]. In the present study, the potential carbohydrate recognition by A. baumannii bacterial cells was investigated by binding of A. baumannii bacteria to glycosphingolipids from various sources on thin-layer chromatograms.

Reference Glycosphingolipids
Total acid and non-acid glycosphingolipid fractions were isolated as described [15]. Pure reference glycosphingolipids were isolated by repeated chromatography on silicic acid columns and by HPLC and identified by mass spectrometry [16,17] and 1 H-NMR spectroscopy [18].
Binding of radiolabeled bacteria to glycosphingolipids on thin-layer chromatograms was done as described [20]. Dried chromatograms were dipped in diethylether/n-hexane (1:5 v/v) containing 0.5% (w/v) polyisobutylmethacrylate for 1 min. The chromatograms were blocked with BSA/PBS/TWEEN for 2 h at room temperature. Then the plates were incubated for 2 h at room temperature with 35 S-labeled bacteria (1-5 × 10 6 cpm/mL) diluted in BSA/PBS/TWEEN/MANNOSE. After washing six times with PBS and drying, the plates were autoradiographed for 12-36 h using XAR-5 X-ray films (Carestream; 8941114).

Isolation of the A. baumannii Binding Tetraglycosylceramide from Human Small Intestine
A non-acid glycosphingolipid fraction (20 mg) from a human small intestine from our glycosphingolipid collection was first separated by chromatography on a 4 g Iatrobeads (6RS-8060, Iatron Laboratories Inc., Tokyo, Japan;) column eluted with chloroform/methanol/water 60:35:8 (by volume), 27 × 1 mL. The fractions obtained were analyzed by thin-layer chromatography and anisaldehyde staining, and the A. baumannii binding activity was assessed using the chromatogram binding assay. The fractions were pooled according to the mobility on thin-layer chromatograms and their A. baumannii binding activity. This resulted in an A. baumannii binding fraction (8.4 mg) containing mono-to tetraglycosylceramides, which was further separated on a 2 g Iatrobeads column eluted with chloroform/methanol/water 65:25:4 (by volume), 18 × 1 mL. Pooling of the A. baumannii binding subfractions gave a fraction (1.7 mg), which was separated on a 1 g Iatrobeads column eluted with chloroform/methanol/water 60:35:8 (by volume), 10 × 0.5 mL. This gave a fraction containing the A. baumannii binding tetraglycosylceramide (1.4 mg). This fraction was designated fraction HI-4.

Isolation of the A. baumannii Binding Tetraglycosylceramide from Rabbit Intestine
The non-acid glycosphingolipid fraction (34 mg) from rabbit intestine from our glycosphingolipid collection was separated in a similar manner. The first separation was done on a 2 g Iatrobeads column eluted with chloroform/methanol/water 60:35:8 (by volume), 24 × 1 mL. A. baumannii binding fractions were pooled (giving 6.3 mg) and then further separated on a 2 g Iatrobeads column eluted with chloroform/methanol/water 65:25:4 (by volume), 17 × 1 mL. Thereby an A. baumannii binding tetraglycosylceramide fraction (1.1 mg) was obtained (designated fraction RI-4).

Isolation of Glycosphingolipids from Human Skin
Full-thickness skin grafts were collected with ethical approval (Dnr-624-16; decision 2016-11-20) after informed consent, at the Department of Plastic Surgery, Sahlgrenska University Hospital. The material was lyophilized and acid and non-acid glycosphingolipids were thereafter isolated as described [15]. Briefly, the material (24.2 g dry weight) was extracted in two steps in a Soxhlet apparatus with chloroform and methanol (2:1 and 1:9, by volume, respectively). The material thereby obtained was subjected to mild alkaline hydrolysis and dialysis, followed by separation on a silicic acid column. Acid and non-acid glycosphingolipid fractions were obtained by chromatography on a DEAE-cellulose column. In order to separate the non-acid glycosphingolipids from alkali-stable phospholipids, the non-acid fraction was acetylated and separated on a second silicic acid column, followed by deacetylation and dialysis. Final purifications were done by chromatographies on DEAE-cellulose and silicic acid columns. Thereby, 28 mg of total non-acid glycosphingolipids was obtained. The total non-acid glycosphingolipid fraction was separated into subfractions by chromatography on a 1 g Iatrobeads column eluted with increasing amounts of methanol in chloroform (10 mL of 10% of methanol in chloroform (by volume), 10 mL of 25% of methanol in chloroform, 10 mL of 33% methanol in chloroform, 10 mL of 75% methanol in chloroform, and 10 mL of 100% methanol). Thereby five glycosphingolipid containing fractions were obtained which were denoted fractions S-1-S-5.

Endoglycoceramidase Digestion and Liquid Chromatography/Electrospray Ionization Mass Spectrometry
Endoglycoceramidase II from Rhodococcus spp. (Takara Bio Europe S.A., Gennevilliers, France) was used for hydrolysis of glycosphingolipids, and the oligosaccharides obtained were analyzed by liquid chromatography-electrospray ionization-mass spectrometry (LC-ESI/MS) [17]. In brief, the oligosaccharides were separated on a column (200 × 0.180 mm) packed in-house with 5 µm porous graphite particles (Hypercarb, Thermo Fischer Scientific, Waltham, MA, USA) and eluted with an acetonitrile gradient (A: 8 mM ammonium bicarbonate; B: 100% acetonitrile). The saccharides were analyzed in the negative ion mode on an LTQ linear quadrupole ion trap mass spectrometer (Thermo Electron, San José, CA, USA). The IonMax standard ESI source on the LTQ mass spectrometer was equipped with a stainless-steel needle kept at −3.5 kV. Compressed air was used as a nebulizer gas. The heated capillary was kept at 270 • C, and the capillary voltage was −50 kV. Full-scan (m/z 380-2000, 2 microscans, maximum 100 ms, target value of 30,000) was performed, followed by data dependent MS 2 scans of the three most abundant ions in each scan (2 microscans, maximum 100 ms, target value of 10,000). The threshold for MS 2 was set to 500 counts. Normalized collision energy was 35%, and an isolation window of 3 u, an activation q = 0.25, and an activation time of 30 ms, was used. The conditions for MS 3 and MS 4 were the same, except that the thresholds were set to 300 and 100 counts, respectively.

Inhibition Experiments
A. baumannii 890 cells were preincubated with PBS buffer or 1:100 dilutions of antibodies raised against the N-terminal domain of CsuE (αEN) [5] for 2 h, and then assayed for binding to pure reference glycosphingolipids on thin-layer chromatograms, as described above.

Screening for A. baumannii Carbohydrate Recognition
In the initial screening for carbohydrate recognition by A. baumannii, mixtures of glycosphingolipids from various sources were used in order to expose the bacteria to a large number of variant carbohydrate structures. Thus, the binding of the A. baumannii strains to acid and non-acid glycosphingolipid mixtures isolated from the small intestine of different species (human, rat, cat, rabbit, dog, monkey, and pig), erythrocytes of different species (human, cat, rabbit, dog, horse, chicken, and sheep), human cancers (lung cancer, kidney cancer, colon cancer, liver cancer, and gastric cancer), and rabbit thymus was tested. There was no binding of A. baumannii to any acid glycosphingolipids. However, in most non-acid glycosphingolipid fractions a binding of the bacteria to compounds migrating in the mono-and diglycosylceramide regions was observed (exemplified in Figure 1, lanes 1-4). In addition, a number of more slow-migrating glycosphingolipids in the non-acid fractions of rabbit small intestine, rabbit thymus, rat intestine, and human small intestine were distinctly recognized by the bacteria (Figure 1, lanes 1-4). graphite particles (Hypercarb, Thermo Fischer Scientific, Waltham, MA, USA) and eluted with an acetonitrile gradient (A: 8 mM ammonium bicarbonate; B: 100% acetonitrile). The saccharides were analyzed in the negative ion mode on an LTQ linear quadrupole ion trap mass spectrometer (Thermo Electron, San José, CA, USA). The IonMax standard ESI source on the LTQ mass spectrometer was equipped with a stainless-steel needle kept at −3.5 kV. Compressed air was used as a nebulizer gas. The heated capillary was kept at 270 °C, and the capillary voltage was -50 kV. Full-scan (m/z 380-2000, 2 microscans, maximum 100 ms, target value of 30,000) was performed, followed by data dependent MS 2 scans of the three most abundant ions in each scan (2 microscans, maximum 100 ms, target value of 10,000). The threshold for MS 2 was set to 500 counts. Normalized collision energy was 35%, and an isolation window of 3 u, an activation q = 0.25, and an activation time of 30 ms, was used. The conditions for MS 3 and MS 4 were the same, except that the thresholds were set to 300 and 100 counts, respectively.

Inhibition Experiments
A. baumannii 890 cells were preincubated with PBS buffer or 1:100 dilutions of antibodies raised against the N-terminal domain of CsuE (αEN) [5] for 2 h, and then assayed for binding to pure reference glycosphingolipids on thin-layer chromatograms, as described above.

Screening for A. baumannii Carbohydrate Recognition
In the initial screening for carbohydrate recognition by A. baumannii, mixtures of glycosphingolipids from various sources were used in order to expose the bacteria to a large number of variant carbohydrate structures. Thus, the binding of the A. baumannii strains to acid and non-acid glycosphingolipid mixtures isolated from the small intestine of different species (human, rat, cat, rabbit, dog, monkey, and pig), erythrocytes of different species (human, cat, rabbit, dog, horse, chicken, and sheep), human cancers (lung cancer, kidney cancer, colon cancer, liver cancer, and gastric cancer), and rabbit thymus was tested. There was no binding of A. baumannii to any acid glycosphingolipids. However, in most non-acid glycosphingolipid fractions a binding of the bacteria to compounds migrating in the mono-and diglycosylceramide regions was observed (exemplified in Figure 1, lanes 1-4). In addition, a number of more slow-migrating glycosphingolipids in the non-acid fractions of rabbit small intestine, rabbit thymus, rat intestine, and human small intestine were distinctly recognized by the bacteria (Figure 1, lanes 1-4).  Screening for Acinetobacter baumannii carbohydrate recognition by binding to glycosphingolipids on thin-layer chromatograms.

Isolation of A. baumannii Binding Tetraglycosylceramides from Human and Rabbit Small Intestine
The A. baumannii binding compounds migrating in the tetraglycosylceramide region from rabbit and human small intestine ( Figure 1, lanes 1 and 4) were both isolated by chromatography on Iatrobeads columns. The fractions obtained were examined for A. baumannii binding on thin-layer chromatograms and pooled according to the mobility on thin-layer chromatograms and their A. baumannii binding activity. After several separation steps, 1.1 mg of the A. baumannii binding tetraglycosylceramide fraction from rabbit intestine (designated fraction RI-4) (Figure 2A During these initial experiments, we also examined the potential carbohydrate binding of A. baumannii CsuC-CsuE protein [5], using 125 I-labeled protein in chromatogram binding assays. However, no glycosphingolipid binding was observed.

Isolation of A. baumannii Binding Tetraglycosylceramides from Human and Rabbit Small Intestine
The A. baumannii binding compounds migrating in the tetraglycosylceramide region from rabbit and human small intestine (Figure 1, lanes 1 and 4) were both isolated by chromatography on Iatrobeads columns. The fractions obtained were examined for A. baumannii binding on thin-layer chromatograms and pooled according to the mobility on thin-layer chromatograms and their A. baumannii binding activity. After several separation steps, 1.1 mg of the A. baumannii binding tetraglycosylceramide fraction from rabbit intestine (designated fraction RI-4) (Figure 2A  Aliquots of fractions HI-4 and RI-4 were digested with endoglycoceramidase II from Rhodococcus spp., and the oligosaccharides obtained were analyzed by LC-ESI/MS, as described below. Aliquots of fractions HI-4 and RI-4 were digested with endoglycoceramidase II from Rhodococcus spp., and the oligosaccharides obtained were analyzed by LC-ESI/MS, as described below.

LC-ESI/MS of the Human Small Intestinal Glycosphingolipid Recognized by A. baumannii
LC-ESI/MS of oligosaccharides using porous graphitized carbon (PGC) columns gives resolution of isomeric oligosaccharides, and the carbohydrate sequence can be deduced from series of C-type fragment ions obtained by MS/MS (MS 2 ) [17,22,23]. Furthermore, MS 2 spectra of oligosaccharides with a Hex or HexNAc substituted at C-4 have diagnostic cross-ring 0,2 A-type fragment ions ( 0,2 A and 0.2 A-H 2 O), which allow differentiation of linkage positions. Thus, such 0,2 A-type fragment ions are found in MS 2 spectra of oligosaccharides with globo (Galα4Gal) or neolacto/type 2 (Galβ4GlcNAc) core structures but are absent in the spectra obtained from oligosaccharides with isoglobo (Galα3Gal) or lacto/type 1 (Galβ3GlcNAc) core chains.
The base chromatogram from LC-ESI/MS of fraction HI-4A had molecular ion at m/z 706, demonstrating an oligosaccharide with one HexNAc and three Hex. MS 2 of this ion ( Figure 3A) gave a C-type fragment ion series (C 2 at m/z 382 and C 3 at m/z 544) identifying an oligosaccharide with Hex-HexNAc-Hex-Hex sequence. The spectrum had a distinct 0,2 A 2 -H 2 O fragment ion at m/z 263 and a 0,2 A 3 fragment ion at m/z 281, demonstrating a terminal Hex-HexNAc sequence with a 4-substitution of the HexNAc, i.e., a type 2 chain [17,22,23]. The 0,2 A 4 ion at m/z 646 and the 0,2 A 3 -H 2 O ion at m/z 628 were most likely derived from cross-ring cleavages of the 4-substituted Glc of the internal lactose (Galβ4Glc) part. Taken together, a neolacto tetrasaccharide (Galβ4GlcNAcβ3Galβ4Glc) was thus tentatively identified.

LC-ESI/MS of the Human Small Intestinal Glycosphingolipid Recognized by A. baumannii
LC-ESI/MS of oligosaccharides using porous graphitized carbon (PGC) columns gives resolution of isomeric oligosaccharides, and the carbohydrate sequence can be deduced from series of C-type fragment ions obtained by MS/MS (MS 2 ) [17,22,23]. Furthermore, MS 2 spectra of oligosaccharides with a Hex or HexNAc substituted at C-4 have diagnostic cross-ring 0,2 A-type fragment ions ( 0,2 A and 0.2 A-H2O), which allow differentiation of linkage positions. Thus, such 0,2 A-type fragment ions are found in MS 2 spectra of oligosaccharides with globo (Galα4Gal) or neolacto/type 2 (Galβ4GlcNAc) core structures but are absent in the spectra obtained from oligosaccharides with isoglobo (Galα3Gal) or lacto/type 1 (Galβ3GlcNAc) core chains.
The base chromatogram from LC-ESI/MS of fraction HI-4A had molecular ion at m/z 706, demonstrating an oligosaccharide with one HexNAc and three Hex. MS 2 of this ion ( Figure 3A) gave a C-type fragment ion series (C2 at m/z 382 and C3 at m/z 544) identifying an oligosaccharide with Hex-HexNAc-Hex-Hex sequence. The spectrum had a distinct 0,2 A2-H2O fragment ion at m/z 263 and a 0,2 A3 fragment ion at m/z 281, demonstrating a terminal Hex-HexNAc sequence with a 4-substitution of the HexNAc, i.e., a type 2 chain [17,22,23]. The 0,2 A4 ion at m/z 646 and the 0,2 A3-H2O ion at m/z 628 were most likely derived from cross-ring cleavages of the 4-substituted Glc of the internal lactose (Galβ4Glc) part. Taken together, a neolacto tetrasaccharide (Galβ4GlcNAcβ3Galβ4Glc) was thus tentatively identified.

LC-ESI/MS of the Rabbit Small Intestinal Glycosphingolipid Recognized by A. baumannii
The base chromatogram from LC-ESI/MS of fraction RI-4 also had a molecular ion at m/z 706 demonstrating an oligosaccharide with one HexNAc and three Hex. MS 2 of this ion ( Figure 3B) also gave a C-type fragment ion series (C 2 at m/z 382 and C 3 at m/z 544) identifying a Hex-HexNAc-Hex-Hex sequence. However, this MS 2 spectrum had a prominent D 1-2 ion at m/z 202, which is obtained by a C 2 -Z 2 double cleavage, and diagnostic for a 3-substituted HexNAc [22]. No 0,2 A 2 -H 2 O fragment ion at m/z 263, or 0,2 A 3 fragment ion at m/z 281, indicating 4-substitution of the HexNAc were present. The 0,2 A 4 -H 2 O fragment ion at m/z 628 and the 0,2 A 4 fragment ion at m/z 646 were derived from cross-ring cleavages of the 4-substituted Glc of the lactose at the reducing end. No 0,2 A 3 fragment ion at m/z 484 was present, demonstrating that the Gal of the lactose unit was 3-substituted. Taken together, these spectral features identified a lacto tetrasaccharide (Galβ3GlcNAcβ3Galβ4Glc).
Thus, the human small intestinal glycosphingolipid recognized by A. baumannii was tentatively identified as neolactotetraosylceramide (Galβ4GlcNAcβ3Galβ4Glcβ1Cer), while the A. baumannii binding glycosphingolipid of rabbit small intestine was tentatively identified as lactotetraosylceramide (Galβ3GlcNAcβ3Galβ4Glcβ1Cer).

Binding to Reference Glycosphingolipids
Thereafter the binding of A. baumannii to defined amounts of a number of reference glycosphingolipids structurally related to neolactotetraosylceramide and lactotetraosylceramide was evaluated. The results are exemplified in Figure 4 and summarized in Table 1. Thereby we found that A. baumannii bound to lactotriaosylceramide ( Figure 4E,F, lane 2), in addition to neolactotetraosylceramide and lactotetraosylceramide. The binding of the bacteria was abolished by substitution of neolactotetraosylceramide or lactotetraosylceramide with an αFuc in the 2-position of the terminal Gal (creating the H type 2 ( Figure 4A,B, lane 1, and Figure 4E,F, lane 4) or H type 1 ( Figure 4A,B, lane 5) pentaosylceramides). The same effect was obtained by an αFuc in the 3or 4-position of the internal GlcNAc position (creating the Le a pentaosylceramide ( Figure 4A,B, lane 6) or the Le x pentaosylceramide (no. 12 in Table 1). Substitution in the 3-position of the terminal Gal of neolactotetraosylceramide with an αGal (Figure 2A,C, lane 4), or a Neu5Ac or Neu5Gc ( Figure 4C,D, lanes 1-3, 5) also abolished the binding. Thus, the minimal requirement for A. baumannii glycosphingolipid binding is a GlcNAc, which may be substituted with Galβ3 (lactotetraosylceramide) or Galβ4 (neolactotetraosylceramide). However, further substitutions of the Galβ3GlcNAc or Galβ4GlcNAc sequences abrogates the binding of the bacteria.
Further binding assays using reference glycosphingolipids showed that A. baumannii also bound to galactosylceramide ( Figure 5A-E, lane 1), lactosylceramide with phytosphingosine and/or hydroxy fatty acids (lane 2), isoglobotriaosylceramide (no. 28 in Table 1), gangliotriaosylceramide (lane 3), and gangliotetraosylceramide (no. 7 in Table 1). While the binding of A. baumannii to lactotriaosylceramide, neolactotetraosylceramide, and lactotetraosylceramide was highly reproducible, binding to these compounds was only occasionally obtained. This is exemplified in Figure 2B where galactosylceramide (lane 2) and gangliotetraosylceramide (lane 7) were non-binding, and the bacteria only bound to gangliotriaosylceramide (lane 3), isoglobotriaosylceramide (lane 5), and lactotetraosylceramide from rabbit intestine (lane 6). Binding to lactosylceramide, isoglobotriaosylceramide, gangliotriaosylceramide, and gangliotetraosylceramide has previously been described for other bacteria classified as "lactosylceramide-binding" [14]. The binding of A. baumannii to fast-migrating compounds in glycosphingolipid mixtures (exemplified in Figure 1B,C) is thus most likely due to recognition of galactosylceramide and lactosylceramide.    Binding is defined as follows: + denotes a highly reproducible binding of A. baumannii CCUG 890 when 4 µg of the glycosphingolipid was applied on the thin-layer chromatogram, while (+) denotes an occasional binding, and -denotes no binding at 4 µg. b In the shorthand nomenclature for fatty acids and bases, the number before the colon refers to the carbon chain length and the number after the colon gives the total number of double bonds in the molecule. Fatty acids with a 2-hydroxy group are denoted by the prefix h before the abbreviation, e.g., h16:0. For long chain bases, d denotes dihydroxy and t trihydroxy. Thus, d18:1 designates sphingosine (1,3-dihydroxy-2-aminooctadecene) and t18:0. phytosphingosine (1,3,4-trihydroxy-2-aminooctadecene).
estimated by binding to serial dilutions of glycosphingolipids on thin-layer chromatograms the bacteria bound to gangliotriaosylceramide, neolactotetraosylceramide, and lactotetraosylceramide with a detection limit of 1-2 μg, while for galactosylceramide 4 μg was required for binding to occur. Thus, there was no clear preferential binding to any of these glycosphingolipids. It should be noted that the binding patterns obtained with the four A. baumannii strains differed to some extent (exemplified in Figure 5 and summarized in Table 2). Lactosylceramide and gangliotriaosylceramide were recognized by all four strains, while only the 890 and 60611 strains bound to galactosylceramide. The most interesting difference was found for lactotetraosylceramide and neolactotetraosylceramide. Here, the 890 strain bound consistently to both compounds, the When the relative affinity of A. baumannii for the binding-active glycosphingolipids was estimated by binding to serial dilutions of glycosphingolipids on thin-layer chromatograms the bacteria bound to gangliotriaosylceramide, neolactotetraosylceramide, and lactotetraosylceramide with a detection limit of 1-2 µg, while for galactosylceramide 4 µg was required for binding to occur. Thus, there was no clear preferential binding to any of these glycosphingolipids.
It should be noted that the binding patterns obtained with the four A. baumannii strains differed to some extent (exemplified in Figure 5 and summarized in Table 2). Lactosylceramide and gangliotriaosylceramide were recognized by all four strains, while only the 890 and 60611 strains bound to galactosylceramide. The most interesting difference was found for lactotetraosylceramide and neolactotetraosylceramide. Here, the 890 strain bound consistently to both compounds, the 60611 strain bound only to lactotetraosylceramide, the 68164 strain bound only to neolactotetraosylceramide, while the 19096 strain did not recognize any of these two glycosphingolipids.

Gangliotri
GalNAcβ4Galβ4Glcβ1Cer a Binding is defined as follows: + denotes a highly reproducible binding of A. baumannii when 4 µg of the glycosphingolipid was applied on the thin-layer chromatogram, while − denotes no binding at 4 µg. b In the shorthand nomenclature for fatty acids and bases, the number before the colon refers to the carbon chain length and the number after the colon gives the total number of double bonds in the molecule. Fatty acids with a 2-hydroxy group are denoted by the prefix h before the abbreviation, e.g., h16:0. For long chain bases, d denotes dihydroxy and t trihydroxy. Thus d18:1 designates sphingosine (1,3-dihydroxy-2-aminooctadecene) and t18:0 phytosphingosine (1,3,4-trihydroxy-2-aminooctadecene).

Isolation and Characterization of an A. baumannii Binding Glycosphingolipid from Human Skin
In order to approach target cell A. baumannii binding glycosphingolipids, we next isolated glycosphingolipids from human skin. The major glycosphingolipids of normal human skin were characterized during the 1980s as glucosylceramide, lactosylceramide, globotriaosylceramide, globoside, and the gangliosides GM3 and GD3 [24]. However, a characterization with the methods available today has not been done. Here we isolated a total non-acid fraction from human skin and separated this into five subfractions, which were denoted fractions S-1-S-5 ( Figure 6A, lanes 1-5).

Inhibition Studies
Preincubation of A. baumannii CCUG 890 with polyclonal antibodies against the N-terminal domain of CsuE (αEN) [5] did not affect the glycosphingolipid binding, i.e., binding to galactosylceramide, lactosylceramide, gangliotriaosylceramide, gangliotetraosylceramide, lactotetraosylceramide, and neolactotetraosylceramide was still obtained (exemplified in Figure 5F). Thus, the hydrophobic three-finger loops at the tip of the Csu pilus are not involved in A. baumannii Binding of A. baumannii to the five non-acid glycosphingolipid subfractions gave a distinct binding in the tetraosylceramide region of fraction S-4 ( Figure 6B, lane 4). Chromatogram binding assays with the Galβ4GlcNAc-recognizing lectin from E. crista-galli [21] also gave a binding in the tetraosylceramide region of fraction S-4 ( Figure 6C, lane 4), indicating the presence of neolactotetraosylceramide.
LC-ESI/MS of the oligosaccharides obtained from fraction S-4 by hydrolysis with endoglycoceramidase II gave a base chromatogram with molecular ion at m/z 706, demonstrating an oligosaccharide with one HexNAc and three Hex. A C-type fragment ion series (C 2 at m/z 382 and C 3 at m/z 544) identifying a Hex-HexNAc-Hex-Hex sequence was obtained by MS 2 of the ion at m/z 706 ( Figure 6D). Moreover, this spectrum had the characteristic 0,2 A 2 -H 2 O fragment ion at m/z 263 and the 0,2 A 3 fragment ion at m/z 281, demonstrating a terminal Hex-HexNAc sequence with a 4-substitution of the HexNAc, i.e., a type 2 chain [17,22,23]. Thus, a neolacto tetrasaccharide (Galβ4GlcNAcβ3Galβ4Glc) was again identified.

Inhibition Studies
Preincubation of A. baumannii CCUG 890 with polyclonal antibodies against the N-terminal domain of CsuE (αEN) [5] did not affect the glycosphingolipid binding, i.e., binding to galactosylceramide, lactosylceramide, gangliotriaosylceramide, gangliotetraosylceramide, lactotetraosylceramide, and neolactotetraosylceramide was still obtained (exemplified in Figure 5F). Thus, the hydrophobic three-finger loops at the tip of the Csu pilus are not involved in A. baumannii glycosphingolipid binding.

Discussion
To establish infections, pathogenic bacteria need to adhere to host cells and tissues, and the majority of microbial attachment sites identified are glycoconjugates [13,14]. In the present study, the carbohydrate recognition by A. baumannii bacterial cells was characterized by binding of A. baumannii bacteria to glycosphingolipids on thin-layer chromatograms. Two A. baumannii binding glycosphingolipids of human and rabbit small intestine were isolated and characterized by mass spectrometry as neolactotetraosylceramide (Galβ4GlcNAcβ3Galβ4Glcβ1Cer) and lactotetraosylceramide (Galβ3GlcNAcβ3Galβ4Glcβ1Cer), respectively. Since A. baumannii also bound to reference lactotriaosylceramide (GlcNAcβ3Galβ4Glcβ1Cer), the basic binding element recognized is GlcNAc.
Glycosphingolipids from human skin were isolated and A. baumannii binding neolactotetraosylceramide was characterized in this target tissue for the bacteria. In addition, as previously described for several other bacteria [14]. A. baumannii bound to galactosylceramide, lactosylceramide with phytosphingosine and/or hydroxy fatty acids, isoglobotriaosylceramide, gangliotriaosylceramide, and gangliotetraosylceramide. However, while the binding of A. baumannii to lactotriaosylceramide, neolactotetraosylceramide, and lactotetraosylceramide was very reproducible, binding to these compounds was only occasionally obtained. The detection limit for binding of A. baumannii to galactosylceramide gangliotriaosylceramide, neolactotetraosylceramide, and lactotetraosylceramide, estimated by binding to serial dilutions of glycosphingolipids on thin-layer chromatograms, was at the microgram level (1-4 µg). Thus, there was no clear preference for any of these glycosphingolipids.
Interestingly, the binding patterns obtained with the four A. baumannii strains were not entirely identical. All four strains bound to lactosylceramide and gangliotriaosylceramide, but only the 890 and 60611 strains bound to galactosylceramide. Furthermore, while the 890 strain bound consistently to both lactotetraosylceramide and neolactotetraosylceramide, the 60611 strain bound only to lactotetraosylceramide, the 68164 strain bound only to neolactotetraosylceramide, and the 19096 strain did not recognize either these glycosphingolipids. This indicates that the adhesins of the A. baumannii strains have differences in the architecture of their carbohydrate binding site(s).
The potential carbohydrate binding of A. baumannii CsuC-CsuE protein [5] was also tested during the initial binding experiments, but no glycosphingolipid binding was observed. Furthermore, the binding of A. baumannii to glycosphingolipids was not affected by preincubation with antibodies against the pilus tip protein CsuE. These antibodies, on the other hand, completely blocked A. baumannii biofilm formation [5]. Thus, the hydrophobic three-finger loops at the tip of the Csu pilus are essential for the formation of biofilms but are not involved in A. baumannii glycosphingolipid binding.
The ability of A. baumannii isolates to adhere to human epithelial cells has been investigated in cell culture experiments, and several proteins, e.g., the outer membrane protein A (OmpA) and the biofilm-associated protein (BAP), have been shown to play a role in the interaction of bacteria with human cells (reviewed in ref. [25]). Despite the multitude of candidate adhesins, no cellular ligands of these proteins have been characterized to date. Further studies of A. baumannii potential adhesins, including the adhesin(s) mediating glycosphingolipid binding, are thus needed.
Interestingly, there is a predicted A. baumannii protein termed lecA, which is described as a galactose binding protein the UniProt database (entry A0A1G5LY11). A. baumannii lecA has very high homology to the lecA (PA-IL) lectin of Pseudomonas aeruginosa, which preferentially binds to glycoconjugates with terminal Galα4Gal sequences and to some extent also to terminal Galα3Gal sequences [26]. Apart from isoglobotriaosylceramide, we did not observe binding of A. baumannii to any glycosphingolipids with terminal Galα4Gal and Galα3Gal sequences (e.g., globotriaosylceramide (no. 28 in Table 1) https://susy.mdpi.com/user/manuscripts/resubmit/0b5776af599a41dc5d4837e5ff40be13) or B penta (no. 12). However, in P. aeruginosa the lecA lectin is located mainly in the cytoplasm with only a small fraction on the cytoplasmic membrane [27]. If this is also the case with A. baumannii lecA, it might explain the absence of binding of the bacterial cells. Obviously the role of this putative lectin needs further investigations.
Binding of G fimbria of human uropathogenic Escherichia coli, and fimbriae belonging to the F17 family of bovine enterotoxigenic and invasive E. coli strains, to terminal GlcNAc has been reported, and the lectin domains of the G and F17 fimbriae have been co-crystallized with N-acetylglucosamine [28,29]. However, the target cell receptors for these fimbriae have not yet been identified.
The identification of the A. baumannii binding glycosphingolipids, lactotetraosylceramide, and neolactotetraosylceramide may further contribute to our understanding of the molecular mechanisms by which A. baumannii establish successful infection in human hosts and could hopefully guide the development of novel high-affinity ligands that may be used as anti-adhesive compounds against A. baumannii infections.