Lipid A Structural Divergence in Rickettsia Pathogens

Spikes in rickettsioses occur as deforestation, urbanization, and homelessness increase human exposure to blood-feeding arthropods. Still, effective Rickettsia vaccines remain elusive.

IMPORTANCE Spikes in rickettsioses occur as deforestation, urbanization, and homelessness increase human exposure to blood-feeding arthropods. Still, effective Rickettsia vaccines remain elusive. Recent studies have determined that Rickettsia lipopolysaccharide anchors the protective S-layer to the bacterial surface and elicits bactericidal antibodies. Furthermore, growing immunological evidence suggests vertebrate sensors (MD-2/TLR4 and noncanonical inflammasome) typically triggered by the lipid A portion of lipopolysaccharide are activated during Rickettsia infection. However, the immunopotency of Rickettsia lipid A is unknown due to poor appreciation for its structure. We determined lipid A structures for four distinct rickettsiae, revealing longer acyl chains relative to highly inflammatory bacterial lipid A. Surprisingly, lipid A of the Rocky Mountain spotted fever agent deviates in structure from other rickettsiae. Thus, lipid A divergence may contribute to variable disease phenotypes, sounding an alarm for determining its immunopotency and possible utility (i.e., as an adjuvant or anti-inflammatory) for development of more prudent rickettsiacidal therapies. L ipopolysaccharide (LPS), an amphipathic molecule comprising the majority of the outer leaflet of the Gram-negative bacterial outer membrane, is composed of extracellular polysaccharide chains (O-antigen) linked to a membrane phosphoglycolipid (lipid A) by a core oligosaccharide. Depending on its structure, lipid A can be a potent activator of the mammalian immune system through detection by MD-2/TLR4 (1, 2) and the noncanonical inflammasome (3,4). However, not all lipid A structures are equal in their ability to activate these mammalian cellular receptors (5); for instance, some bacterial pathogens employ lipid A modification as a mechanism of immune evasion when infecting a mammalian host (6,7). Given the importance of lipid A for membrane integrity, resistance to antibiotics, and use as a vaccine adjuvant, it is crucial to understand the structure and immunostimulatory potential of lipid A from Gram-negative pathogens.
Species of Rickettsia, Gram-negative obligate intracellular Alphaproteobacteria, are metabolic parasites of a wide range of eukaryotic hosts (8). Across the Rickettsia tree, agents of human disease from the transitional group (TRG), typhus group (TG), and spotted fever group (SFG) rickettsiae are interspersed with numerous invertebrate and protist endosymbionts, most with unknown pathogenicity. All described rickettsioses are vector-borne diseases that differ in their severity of illness and clinical manifestations (9), facts undoubtedly linked to variability in the Rickettsia secretome (10) and Oantigen epitopes (11,12). Rickettsia O-antigen contains the sugar quinovosamine (11)(12)(13), and transposon-mediated disruption of an epimerase involved in its production abrogates S-layer formation, dampens pathogenicity, and abolishes recognition by bactericidal antibodies (14). While the immunopotency of the lipid A moiety of Rickettsia LPS remains unknown, it could be proinflammatory given MD-2/TLR4 and noncanonical inflammasome activation during infection (15)(16)(17)(18)(19)(20). However, the lone published Rickettsia lipid A structure (an unreported strain of Rickettsia typhi, the agent of murine typhus) revealed a bisphosphorylated hexa-acylated structure with acyl chains ranging from C14 to C18 in length (21), much longer than the C12-C14 chains of the highly inflammatory hexa-acylated lipid A of Escherichia coli (Fig. 1A). These collective observations warrant determining Rickettsia lipid A structures for phenotypically diverse species to assess if any structural variability correlates with disease severity.
To investigate possible lipid A diversity between rickettsiae, we isolated and analyzed lipid A from three human pathogens (R. akari, R. typhi, and R. rickettsii) and a nonpathogen (R. montanensis), all grown in Vero cell cultures. Lipid A extracted directly from infected-Vero cells for R. akari, R. typhi, and R. montanensis produced major ions at m/z 1,936 that are largely consistent with the published R. typhi lipid A structure (21) and likely represent bisphosphorylated hexa-acyl structures (Fig. 1B). However, there is considerable variance between species in the number and intensity of minor lipid A ions (m/z 1,963, 1,909, and 1,882) that likely represents heterogeneity in fatty acid (FA) incorporation (see Fig. S1 in the supplemental material) known to occur in some bacteria (22). In contrast, the Rocky Mountain spotted fever agent, R. rickettsii, produces a lower-molecular-weight lipid A molecule (major ion, m/z 1,882) (Fig. 1C), corresponding to a loss of four carbons from one or more FA chains (Fig. 1D). Sequential FA release by alkaline treatment indicates a laurate (C12) on 29-hydroxypalmitate (C16-OH) as opposed to palmitate/stearate (C16/C18) in the other Rickettsia structures (Fig. S2).
The underlying mechanisms for acyl chain heterogeneity within species and divergence between R. rickettsii and other rickettsiae are not readily apparent. The enzymes for lipid A biosynthesis are highly conserved across rickettsiae (Fig. S3), including the motifs that function as a hydrocarbon ruler in late acyltransferase LpxJ (23). However, inspection of analogous motifs in late acyltransferase LpxL revealed an active-site substitution (Leu-Ile) conserved in SFG rickettsiae that diverge after R. gravesii (Fig. 2,  Fig. S4). The potential for this modified LpxL active site to incorporate shorter 29 acyl chains into Rickettsia lipid A awaits experimentation. Despite conserved FA synthesis machinery, malonyl-ACP is synthesized in rickettsiae using host-derived precursors and cofactors (24), making FA availability and incorporation contingent on host metabolic burden during infection. Variance in FA acyl chain length may also reflect engrained flexibility for adapting to divergent cellular environments encountered across invertebrate and vertebrate hosts, making it a priority to assess lipid A structures directly from arthropod and mammalian infections.
The membrane or immunological function of shorter 29 secondary acyl chains in derived SFG rickettsiae remains to be determined. If other human pathogens in this lineage diverging after R. montanensis (Fig. 2, red star) have acyl chain lengths similar to those of R. rickettsii, this lipid A structure may serve as a candidate drug target or rickettsii lipid A shown as an example. Lipid A from R. rickettsii strain Sheila Smith was subjected to sequential release of fatty acids and analyzed at each step with MALDI-TOF analysis (for full spectra, see Fig. S2). m/z, mass-to-charge ratio of lipid A ions identified during MALDI-TOF analysis.
vaccine adjuvant for treatment of these particular SFG rickettsioses. Furthermore, determining lipid A structures for other species throughout the rickettsial tree stands to illuminate additional structural diversity that bolsters joining lipid A with the secretome and O-antigen as variable factors defining phenotypically diverse rickettsioses.
Rickettsiae are rare among obligate intracellular bacteria in that they lyse the phagosome and colonize the host cytosol, risking inordinate metabolic thievery in the face of host surveillance systems (24). The pilfering of host metabolites for cell envelope synthesis entangles rickettsial growth and virulence. Further characterization of rickettsial lipid A, particularly its immunopotency, is imperative for advancing knowledge on host-pathogen interactions and a highly unique mode of obligate intracellular parasitism. Genome-based phylogeny on the left was estimated as previously described (25). Mass spectra on the right depict MALDI-TOF analyses of lipid A extracted from R. akari strain Hartford, R. typhi strain Wilmington, R. montanensis strain M5/6, and R. rickettsii strains Sheila Smith and Iowa (for full spectra, see Fig. S1 in the supplemental material). Peak labels are masses of singly charged ions; arrows denote major peaks. Insets show typical arthropod vectors. The star on the phylogeny indicates the earliest point in SFG rickettsia evolution where a switch from palmitate/stearate (C16/18) to laurate (C12) on lipid A 29 hydroxypalmitate could have occurred. The node reflecting a switch to a conserved Ile in position 17 of block I of the LpxL active site is also noted (see Fig. S4 for more details).
Bacterial strains and cell culture. Vero 76 cells (African green monkey kidney; ATCC CRL-1587) were maintained in Dulbecco's modification of Eagle's medium (DMEM with 4.5 g/liter glucose and 480 L-glutamine) supplemented with 10% heat-inactivated fetal bovine serum (FBS) at 37°C with 5% CO 2 . Rickettsiae were propagated in Vero 76 cells grown in DMEM (supplemented with 5% FBS at 34°C with 5% CO 2 ) for 48 to 72 h until confluence before harvesting. Rickettsial cultures were partially purified by mild sonication (one 10-s pulse; power output 6) followed by 5.0-mm filtration and collected by gentle centrifugation (2,000 Â g; 15 min). Pellets were washed once in ultrapure water prior to lipid A extraction.
Lipid A extraction. Lipid A microextraction from rickettsial cultures initiated by placing pellets from 1 to 4 confluent T75 cm 2 culture flasks (maintained as described above) in 400ml of solution (5 parts isobutyric acid, 3 parts 1 M NH 4 OH) and heated at 100°C for 1 h followed by a 15-min incubation on ice and centrifugation at 2,000 Â g for 15 min. The lipid A-containing bottom layer of supernatant was collected, mixed in equal parts with H 2 O, frozen, and then lyophilized. Contaminants were washed from the dried material by two rounds of methanol (MeOH) washes (1 ml MeOH with sonicating and pelleting at 10,000 Â g for 5 min). The final product was reconstituted in 50ml chloroform-MeOH-H 2 O (2:1:0.25) along with 4 to 8 grains of Dowex ion exchange resin (ThermoFisher), incubated at room temperature with vortexing (5 min). Solubilized lipid A (1 to 2ml) was spotted onto a stainless steel target plate with 1ml of Norharmane matrix (10 mg/ml in 2:1 chloroform:methanol) for matrix-assisted laser desorption ionization (MALDI) analysis on a Bruker microflex MALDI-time of flight mass spectrometry (TOF MS) instrument in negative-ion mode calibrated with Agilent tuning mix (Santa Clara, CA), and data were processed in flexAnalysis (Bruker Daltonics). All microextraction chemicals were obtained from Sigma-Aldrich (St. Louis, MO).
Sequential fatty acid release by alkaline treatment. To liberate primary esterlinked fatty acids, lipid A samples were suspended in 100 ml of 28% NH 4 OH and incubated at 50°C for 5 h with occasional vortexing. To liberate secondary fatty acids, lipid A or treated derivatives were suspended in 40% methylamine and incubated at 50°C for 3 h. Ultrapure H 2 O (100 ml) was added before samples were frozen and lyophilized. Final products were reconstituted in chloroform-MeOH-H 2 O (2:1:0.25) and analyzed by MALDI-TOF. The entire workflow described above was performed in triplicate for all five strains.

SUPPLEMENTAL MATERIAL
Supplemental material is available online only.