5
Synthesis, characterization, and immunogenicity in mice of Shigella sonnei O-specific oligosaccharide-core-protein conjugates John B. Robbins a,1 , Joanna Kubler-Kielb a , Evguenii Vinogradov b , Christopher Mocca a , Vince Pozsgay a , Joseph Shiloach c , and Rachel Schneerson a a National Institute of Child Health and Human Development, National Institutes of Health, 9000 Rockville Pike, Bethesda, MD 20892; b Institute for Biological Sciences, National Research Council, 100 Sussex Drive, Ottawa, ON, Canada K1A 0R6; and c National Institute of Allergy and Infectious Disease, National Institutes of Health, Bethesda, MD 20892 Contributed by John B. Robbins, January 27, 2009 (sent for review November 14, 2008) Shigellosis, an enteric disease, is on the World Health Organiza- tion’s priority prevention list. In one study, the Shigella sonnei O-specific polysaccharide (O-SP)-protein conjugate showed 72% protection against disease in Israeli army recruits exposed to high rates (8 –14%) of infection. The protection was related to vaccine- induced IgG anti-O-SP levels. Synthetic oligosaccharides of Shigella dysenteriae type 1, bound by their reducing ends to a carrier protein (‘‘sun’’-type configuration), induced significantly higher antibody levels than the native O-SP bound to protein by multiple- point attachments (‘‘lattice’’-type configuration). Attempts to syn- thesize the S. sonnei O-SP based oligosaccharides were not suc- cessful. Here, we describe the isolation, characterization, and conjugation of low-molecular-mass O-SP-core (O-SPC) fragments. The O-SPC fragments were bound by their reducing ends similar to the preparation of the synthetic S. dysenteriae type 1 conjugates. The O-SPC conjugates used oxime linkages between the terminal Kdo residues at the reducing ends of the S. sonnei saccharides and aminooxy linkers bound to BSA or a recombinant diphtheria toxin. The coupling reaction was carried out at a neutral pH and room temperature. IgG antibody levels induced in young outbred mice by the S. sonnei O-SPC conjugates were significantly higher then those elicited by the O-SP conjugates. Accordingly, we propose to evaluate clinically these conjugates. lipopolysaccharide glycoconjugate vaccine Kdo IgG S higellae, Gram-negative bacteria, cause endemic and epi- demic diarrhea and/or dysentery worldwide, especially in developing countries. It has been estimated that 160 million cases of shigellosis with 1 million deaths occur worldwide annually (1). At least half of the cases and deaths occur in children 5 years old (1). The symptoms include fever, watery diarrhea, or dysentery (blood and mucus in the stool and cramps). Control of this disease is hampered by the low infec- tious dose of this pathogen (100 bacteria) and lack of safe drinking water and food (2). Despite its discovery over a century ago, there is still no licensed vaccine for shigellae. Residual morbidity, even after effective antibiotic therapy and increasing antibiotic resistance of shigellae, urge the development of vac- cines to prevent disease caused by this pathogen. Lipopolysaccharides (LPSs) of Shigella are both essential virulence factors and protective antigens of this genus. The outer domain of this tripartite molecule, termed O-specific polysaccharide (O-SP), ‘‘shields’’ the bacteria from serum complement killing, similar to the action of capsular polysac- charides (3, 4). We hypothesized that serum antibodies to the O-SP of shigellae confer immunity to humans against the homologous bacteria (3). To test this hypothesis, experimental vaccines composed of protein conjugates of the O-SP of Shigella dysenteriae type 1, Shigella sonnei, and Shigella flexneri 2a were synthesized and evaluated in young adults (5). The 3 conjugates were safe and elicited specific LPS antibodies. Evaluation of a S. sonnei O-SP/recombinant Pseudomonas aeruginosa Exotoxin A (rEPA) conjugate in Israeli soldiers demonstrated 72% efficacy with vaccine failures occurring in individuals who responded with significantly lower serum antibody levels than those who were protected (6). The highest incidence and severity of S. sonnei shigellosis is in young children. Evaluation of such conjugates in children showed age-related antibody responses and protection (7). A signifi- cant improvement in the immunogenicity of S. dysenteriae type 1 conjugates was achieved by using synthetic oligosaccharides (OS) of defined lengths bound by their reducing ends to a protein at defined densities (8). Synthesis of S. sonnei O-SP oligosaccharides has not been possible to date. We therefore used low-mass O-SP-core (O-SPC) fragments isolated from the LPS to bind to carrier proteins similarly to the preparation of the synthetic S. dysenteriae type 1 oligosacchride-protein conjugates. Results Isolation and Chemical Characterization of O-SPC. LPS was extracted from 18-h cultures of S. sonnei or Plesiomonas shigelloides as described (5). S. sonnei saccharides, released after mild acid hydrolysis from lipid A, were separated into 4 fractions (Fig. 1). The yields of fractions 1–4 were 50%, 17%, 31%, and 2% by weight, respectively. Integration of the FucNAc4N methyl signal in 1 H-NMR spectra (1.34–1.36 ppm) relative to the anomeric signals of core -Gal M (5.82 ppm) and -Gal L (5.62 ppm) (see Table 1 and Scheme 1) showed that fraction F1 contained core with 29 O-SP repeat units (RU), F2 contained core with an average of 3.5 RU, and F3 contained core with an average of 1.3 RU (Fig. 2). Fraction F4 contained various degradation products and was not studied further. MS spectra confirmed that fraction F3 consisted of the core with 1 RU (Fig. 3A). The following species have been detected: core 1 RU without the core GlcN (2,110.6 Da), core 1 RU including GlcN (2,271.7 Da), core 1 RU including GlcN but without phosphate (P) (2,190.9 Da), core 1 RU with GlcN, P and PEtN (2,377.8 Da), and the latter but without P (2,314.1 Da). MS analysis of fraction F2 showed a complex spectrum containing 3,4, and 5 charged ions of the oligosaccharides with 2, 3, and 4 RU, heterogenic in P, PEtN, and GlcN substitutions, respectively. To interpret the spectrum, F2 was partially dephosphorylated. The following species were detected (Fig. 3B): (i) core 2 RU (2,433.9 Da), and partially substituted with P (2,496.9 Da), and core 2 RU with GlcN (2,595.0 Da); Author contributions: J.B.R., J.K.-K., E.V., and R.S. designed research; J.B.R., J.K.-K., E.V., C.M., and R.S. performed research; J.B.R., J.K.-K., E.V., and R.S. contributed new reagents/ analytic tools; J.B.R., J.K.-K., E.V., C.M., and R.S. analyzed data; and J.B.R., J.K.-K., E.V., and R.S. wrote the paper. The authors declare no conflict of interest. 1 To whom correspondence should be addressed. E-mail: [email protected]. 7974 –7978 PNAS May 12, 2009 vol. 106 no. 19 www.pnas.orgcgidoi10.1073pnas.0900891106 Downloaded by guest on March 17, 2020

Synthesis, characterization, and immunogenicity in mice of ...Synthesis, characterization, and immunogenicity in mice of Shigella sonneiO-specific oligosaccharide-core-protein conjugates

  • Upload
    others

  • View
    11

  • Download
    0

Embed Size (px)

Citation preview

Page 1: Synthesis, characterization, and immunogenicity in mice of ...Synthesis, characterization, and immunogenicity in mice of Shigella sonneiO-specific oligosaccharide-core-protein conjugates

Synthesis, characterization, and immunogenicity inmice of Shigella sonnei O-specificoligosaccharide-core-protein conjugatesJohn B. Robbinsa,1, Joanna Kubler-Kielba, Evguenii Vinogradovb, Christopher Moccaa, Vince Pozsgaya, Joseph Shiloachc,and Rachel Schneersona

aNational Institute of Child Health and Human Development, National Institutes of Health, 9000 Rockville Pike, Bethesda, MD 20892; bInstitute for BiologicalSciences, National Research Council, 100 Sussex Drive, Ottawa, ON, Canada K1A 0R6; and cNational Institute of Allergy and Infectious Disease, NationalInstitutes of Health, Bethesda, MD 20892

Contributed by John B. Robbins, January 27, 2009 (sent for review November 14, 2008)

Shigellosis, an enteric disease, is on the World Health Organiza-tion’s priority prevention list. In one study, the Shigella sonneiO-specific polysaccharide (O-SP)-protein conjugate showed 72%protection against disease in Israeli army recruits exposed to highrates (8–14%) of infection. The protection was related to vaccine-induced IgG anti-O-SP levels. Synthetic oligosaccharides of Shigelladysenteriae type 1, bound by their reducing ends to a carrierprotein (‘‘sun’’-type configuration), induced significantly higherantibody levels than the native O-SP bound to protein by multiple-point attachments (‘‘lattice’’-type configuration). Attempts to syn-thesize the S. sonnei O-SP based oligosaccharides were not suc-cessful. Here, we describe the isolation, characterization, andconjugation of low-molecular-mass O-SP-core (O-SPC) fragments.The O-SPC fragments were bound by their reducing ends similar tothe preparation of the synthetic S. dysenteriae type 1 conjugates.The O-SPC conjugates used oxime linkages between the terminalKdo residues at the reducing ends of the S. sonnei saccharides andaminooxy linkers bound to BSA or a recombinant diphtheria toxin.The coupling reaction was carried out at a neutral pH and roomtemperature. IgG antibody levels induced in young outbred miceby the S. sonnei O-SPC conjugates were significantly higher thenthose elicited by the O-SP conjugates. Accordingly, we propose toevaluate clinically these conjugates.

lipopolysaccharide � glycoconjugate � vaccine � Kdo � IgG

Shigellae, Gram-negative bacteria, cause endemic and epi-demic diarrhea and/or dysentery worldwide, especially in

developing countries. It has been estimated that �160 millioncases of shigellosis with �1 million deaths occur worldwideannually (1). At least half of the cases and deaths occur inchildren �5 years old (1). The symptoms include fever, waterydiarrhea, or dysentery (blood and mucus in the stool andcramps). Control of this disease is hampered by the low infec-tious dose of this pathogen (�100 bacteria) and lack of safedrinking water and food (2). Despite its discovery over a centuryago, there is still no licensed vaccine for shigellae. Residualmorbidity, even after effective antibiotic therapy and increasingantibiotic resistance of shigellae, urge the development of vac-cines to prevent disease caused by this pathogen.

Lipopolysaccharides (LPSs) of Shigella are both essentialvirulence factors and protective antigens of this genus. Theouter domain of this tripartite molecule, termed O-specificpolysaccharide (O-SP), ‘‘shields’’ the bacteria from serumcomplement killing, similar to the action of capsular polysac-charides (3, 4). We hypothesized that serum antibodies to theO-SP of shigellae confer immunity to humans against thehomologous bacteria (3). To test this hypothesis, experimentalvaccines composed of protein conjugates of the O-SP ofShigella dysenteriae type 1, Shigella sonnei, and Shigella flexneri2a were synthesized and evaluated in young adults (5). The 3conjugates were safe and elicited specific LPS antibodies.

Evaluation of a S. sonnei O-SP/recombinant Pseudomonasaeruginosa Exotoxin A (rEPA) conjugate in Israeli soldiersdemonstrated 72% efficacy with vaccine failures occurring inindividuals who responded with significantly lower serumantibody levels than those who were protected (6). The highestincidence and severity of S. sonnei shigellosis is in youngchildren. Evaluation of such conjugates in children showedage-related antibody responses and protection (7). A signifi-cant improvement in the immunogenicity of S. dysenteriae type1 conjugates was achieved by using synthetic oligosaccharides(OS) of defined lengths bound by their reducing ends to aprotein at defined densities (8). Synthesis of S. sonnei O-SPoligosaccharides has not been possible to date. We thereforeused low-mass O-SP-core (O-SPC) fragments isolated fromthe LPS to bind to carrier proteins similarly to the preparationof the synthetic S. dysenteriae type 1 oligosacchride-proteinconjugates.

ResultsIsolation and Chemical Characterization of O-SPC. LPS was extractedfrom 18-h cultures of S. sonnei or Plesiomonas shigelloides asdescribed (5). S. sonnei saccharides, released after mild acidhydrolysis from lipid A, were separated into 4 fractions (Fig. 1).The yields of fractions 1–4 were 50%, 17%, 31%, and 2% byweight, respectively. Integration of the FucNAc4N methyl signalin 1H-NMR spectra (1.34–1.36 ppm) relative to the anomericsignals of core �-Gal M (5.82 ppm) and �-Gal L (5.62 ppm) (seeTable 1 and Scheme 1) showed that fraction F1 contained corewith �29 O-SP repeat units (RU), F2 contained core with anaverage of 3.5 RU, and F3 contained core with an average of 1.3RU (Fig. 2). Fraction F4 contained various degradation productsand was not studied further.

MS spectra confirmed that fraction F3 consisted of the corewith 1 RU (Fig. 3A). The following species have been detected:core � 1 RU without the core GlcN (2,110.6 Da), core � 1 RUincluding GlcN (2,271.7 Da), core � 1 RU including GlcN butwithout phosphate (P) (2,190.9 Da), core � 1 RU with GlcN, Pand PEtN (2,377.8 Da), and the latter but without P (2,314.1 Da).

MS analysis of fraction F2 showed a complex spectrumcontaining 3�, 4�, and 5� charged ions of the oligosaccharideswith 2, 3, and 4 RU, heterogenic in P, PEtN, and GlcNsubstitutions, respectively. To interpret the spectrum, F2 waspartially dephosphorylated. The following species were detected(Fig. 3B): (i) core � 2 RU (2,433.9 Da), and partially substitutedwith P (2,496.9 Da), and core � 2 RU with GlcN (2,595.0 Da);

Author contributions: J.B.R., J.K.-K., E.V., and R.S. designed research; J.B.R., J.K.-K., E.V.,C.M., and R.S. performed research; J.B.R., J.K.-K., E.V., and R.S. contributed new reagents/analytic tools; J.B.R., J.K.-K., E.V., C.M., and R.S. analyzed data; and J.B.R., J.K.-K., E.V., andR.S. wrote the paper.

The authors declare no conflict of interest.

1To whom correspondence should be addressed. E-mail: [email protected].

7974–7978 � PNAS � May 12, 2009 � vol. 106 � no. 19 www.pnas.org�cgi�doi�10.1073�pnas.0900891106

Dow

nloa

ded

by g

uest

on

Mar

ch 1

7, 2

020

Page 2: Synthesis, characterization, and immunogenicity in mice of ...Synthesis, characterization, and immunogenicity in mice of Shigella sonneiO-specific oligosaccharide-core-protein conjugates

(ii) core � 3 RU (2,837.1 Da), and partially substituted with aphosphate group (2,918.1 Da), or core � 2 RU with GlcN(2,999.1 Da); and (iii) core � 4 RU (3,240.6 Da). Integration ofNMR data indicated that even longer oligosaccharides, with 5RU were present in F2 but their signals were not observed in themass spectra, probably because they showed multiply-chargedions mixed with background signals. Similar results had beenobtained previously (9, 10). The F2 and F3 oligosaccharidefractions were used for conjugation.

Preparation and Characterization of Conjugates. Three conjugateswere prepared by binding O-SPC-F2 to either BSA (conjugate1; BSA/O-SPC-F2) or recombinant diphtheria (rDT) toxin(conjugates 2 and 3: rDT/O-SPC-F2). One conjugate wasprepared by binding O-SPC-F3 to BSA (conjugate 4: BSA/O-SPC-F3). The conjugation was based on formation of stableoxime linkages between the Kdo residue present at the O-SPCreducing end and an aminooxy linker bound to the carrierprotein (12). This procedure yielded high molecular massconjugates, revealed by MALDI-TOF MS, SDS/PAGE, andprotein and sugar colorimetric assays; all methods provided

comparable results. The number of O-SPC chains per proteinwas calculated from the molecular mass of the conjugate, thecarrier protein, and the O-SPC. Thus, conjugate 1 contained�7 O-SPC chains/BSA molecule, conjugate 2 contained �6O-SPC/rDT, conjugate 3 contained �12 O-SPC chains/rDT,and conjugate 4 contained �11 O-SPC chains/BSA (Table 2).An excess of saccharide was used for conjugation to ensuremaximal binding. The yield of the protein in the conjugate was65–75%, and the yield of the saccharide was 30–35%. Conju-gates 1, 2, and 3 prepared with O-SPC-F2 reacted by doubleimmunodiffusion with rabbit anti-S. sonnei and anti-proteinsera by a line of identity. Conjugate 4 prepared with O-SPC-F3precipitated with the anti-BSA serum but not with the anti-S.sonnei serum. Only conjugates of O-SPC-F2 were used forimmunization.

IgG Anti-LPS Responses (Table 2). Conjugates 1, 2, and 3 elicitedlow levels of IgG anti-LPS after the second injection with abooster response after the third. The geometric means (GM)of IgG anti-LPS after the third injection were 366 ELISA units(EU) for conjugate 1 and 392 EU for conjugate 2. Conjugate3, which contained twice as much of O-SPC-F2 chains per rDTthan conjugate 2 (12 vs. 6), induced statistically-lower GMantibody levels (150 EU vs. 392 EU; P � 0.01). All 3 O-SPC-F2conjugates induced statistically-higher antibody levels than the‘‘lattice’’-type conjugate (a clinical lot) prepared with thefull-length O-SP (366 vs. 67, P � 0.0001; 392 vs. 67, P � 0.0001;150 vs. 67: P � 0.05).

IgG Anti-O-SP Responses (Table 3). Coating the ELISA plates withS. sonnei or P. shigelloides LPS of identical O-SP but different

35 45 55 65 75 85 95 105 115 125 135 145

Effluent [ml]

RefractionIndex

F1 F2 F3 F4

Fig. 1. BioGel P-10 gel filtration of S. sonnei LPS after 1% acetic acidhydrolysis. F1, O-SP, �29 RU; F2, core � average 3.5 RU of O-SP; F3, core �average 1.3 RU of O-SP; F4, degradation products, no core or O-SP.

Table 1. NMR analyses of S. sonnei O-SPC fragment (�, ppm)

Residue, compound Nucleus

Fragment

1 2 3 4 5 6a 6b

�-Glc H H 5.28 3.68 4.06 3.79 3.86C 101.6 71.6 77.4 71.6 73.4

�-Glc K H 5.83 3.89 4.19 3.58 4.13 3.75 3.75C 95.7 73.8 79.0 69.0 72.5 62.3

�-Gal L H 5.62 3.99 4.21 4.01 4.13C 92.5 73.5 69.3 70.4 72.4

�-Gal M H 5.33 3.86 3.98 4.00 4.12C 96.7 69.4 70.5 70.5 72.5

�-Glc T H 4.74 3.40 3.70 3.50 3.45 3.78 3.90C 103.7 74.0 85.4 69.3 76.7 61.4

�-FucNAc4N W H 4.77 3.84 4.19 3.86 4.08 1.34C 102.8 52.6 77.0 55.9 68.3 16.7

�-AltNAcA X H 4.79 3.84 3.76 4.47 4.60C 102.3 52.4 68.9 78.3 78.1 174.7

�-FucNAc4N Y H 4.76 3.88 4.19 3.96 4.08 1.34C 104.2 52.1 77.0 56.1 68.3 16.7

�-AltNAcA Z H 4.90 4.00 3.69 4.40 4.52C 102.1 52.5 69.1 69.9 78.4 175.1

N G α-GlcN-7-α-Hep-7 M L K |α-Gal-2-α-Gal-2-α-Glc-3-α-Glc-3-α-Hep4P-3-α-Hep4P(PEtA)-5-Kdo | H F E C R-3-β-Glc-3 T

R = α-L-AltNAcA-3-β-FucNAc4N-4-α-L-AltNAcA-3-β-FucNAc4N- Z Y X W

Scheme 1. Structure of S. sonnei O-SPC fragment.

Robbins et al. PNAS � May 12, 2009 � vol. 106 � no. 19 � 7975

MED

ICA

LSC

IEN

CES

Dow

nloa

ded

by g

uest

on

Mar

ch 1

7, 2

020

Page 3: Synthesis, characterization, and immunogenicity in mice of ...Synthesis, characterization, and immunogenicity in mice of Shigella sonneiO-specific oligosaccharide-core-protein conjugates

core structures (10, 12) yielded similar antibody levels. Further,sera induced by either O-SPC or O-SP conjugates were inhibitedsimilarly by O-SPs of S. sonnei and of P. shigelloides with orwithout the core.

DiscussionThe highest incidence and severity of S. sonnei shigellosisoccur in young children. The experimental O-SP-protein con-jugate vaccine revealed age-related immunogenicity and pro-tection (7). To enhance antibody responses of young childrento S. sonnei O-SP, and based on the statistically higherantibody levels induced by the synthetic S. dysentariae type 1‘‘sun’’ configuration conjugates than those induced by thefull-length bacterial lattice-type conjugates (8), we preparedsun configuration S. sonnei conjugates. Because the synthesisof S. sonnei O-SP-based oligosaccharides did not extendbeyond a disaccharide unit (13) we isolated low-molecular-mass oligosaccharides containing LPS core plus several O-SPRU (O-SPC) from the bacterial LPS and used them forconjugate preparation. An oligosaccharide fraction, F2 con-taining the core plus an average of 3.5 RU of O-SP, was boundto the protein carriers by oxime formation (11). These con-jugates were antigenic and induced significantly higher IgGantibody levels then those induced by conjugates preparedwith full-length O-SP. The antibodies induced by the O-SPCconjugate were directed to the RU of the O-SP and not to thecore region based on the similar antibody levels obtained byusing S. sonnei or P. shigelloides LPS for ELISA and the close

to a 100% inhibition of O-SPC-induced antibodies by O-SP ofboth organisms. On the basis of these data we propose toevaluate clinically the O-SPC protein conjugates.

The finding that the identity of the sugar residue positioned atthe nonreducing end of the synthetic S. dysentariae type 1oligosaccharide conjugates is an important variable for theimmunogenicity of these conjugates (14) points to the impor-tance of the end groups in general in the immune response andmay provide an explanation for the superior immunogenicity ofthe O-SPC and synthetic oligosaccharide conjugates over O-SPconjugates; the former have more end groups. The importanceof the terminal saccharide was also demonstrated for Bordetellaeconjugates prepared the same way, where the terminal reducingend was immunodominant (11). This method is being studiedwith S. flexnerii 2a and 6 LPS. The mild reaction conditions andrelative ease of preparation provided by O-SPC may be appli-cable for the preparation of conjugates against other Gram-negative pathogens such as nontyphoidal Salmonella and Esch-erichia coli O157.

Materials and MethodsGrowth of Bacteria and Isolation of LPS. S. sonnei strain 53G and P. shigelloidesstrain 7–63 (serotype O17) were obtained from Sam Formal (Walter ReedArmy Institute of Research, Silver Spring, MD) and cultivated as described (5).LPS from either strain was extracted by the hot phenol method and purifiedas described (15).

1.52.02.53.03.54.04.55.05.5 ppm

1.0 1.2 96.0

1.0 1.1 10.7

1.0 1.1 4.0

Fig. 2. Integration of the 1H NMR spectra of S. sonnei O-SP (Top), O-SPC-F2(Middle), and O-SPC-F3 (Bottom) for the determination of the number of RUattached to the core. Signals at 5.82 and 5.62 ppm belonging to core �-Gal Mand �-Gal L (1 proton) were integrated with signals at 1.34–1.36 ppm belong-ing to FucNAc4N methyl group (3 protons) of O-SP (see Table 3).

700 800 m/z, amu

756.2

791.6770.2729.2

764.2750.2

723.2702.4

785.6696.4 797.2 832.6

826.6

863.6

2110.6

2190.9

2271.7

2314.1

800 900 1000 1100m/z, amu

804.4810.3

831.3 944.7938.8864.0

837.1

858.2

998.7971.7 1079.2

2433.9

2514.32595.0

2837.1

2918.1 2999.1 3240.6

2RU 3RU 4RUB

A 1RU

2377.8

Fig. 3. ESI mass spectra of S. sonnei O-SPC-F3 (A) and dephosphorylatedO-SPC-F2 (B). Triple charged ions are shown. Components of fraction 2 withmasses of 2514.3 and 2918.1 Da contain one phosphate because of incompletedephosphorylation. Assignments of molecular masses are in Results.

7976 � www.pnas.org�cgi�doi�10.1073�pnas.0900891106 Robbins et al.

Dow

nloa

ded

by g

uest

on

Mar

ch 1

7, 2

020

Page 4: Synthesis, characterization, and immunogenicity in mice of ...Synthesis, characterization, and immunogenicity in mice of Shigella sonneiO-specific oligosaccharide-core-protein conjugates

Isolation of Oligosaccharides. S. sonnei LPS (200 mg) was treated with 1% aceticacid at 100 °C for 1.5 h. Lipid A was removed by ultracentrifugation at142,000 � g for 5 h at 4 °C, and the soluble product subjected to gel chroma-tography on a BioGel P-10 (1 � 100 cm) column in pyridine/acetic acid/waterbuffer (4:8:988 mL), monitored with a Knauer differential refractometer.

Analytical Methods. Protein concentration was measured by the method ofLowry et al. (16), and sugar concentration was measured by the anthrone assay(17). SDS/PAGE used 14% gels according to the manufacturer’s instructions(Bio-Rad). Immunodiffusion was performed in 1% agarose in PBS.

Spectroscopy. 1H and 13C NMR spectra were recorded by using a Varian Inova500-MHz spectrometer for samples in D2O solutions at 35 °C with acetonestandard (2.225 ppm for 1H and 31.5 ppm for 13C) using standard pulsesequence COSY, TOCSY (mixing time 120 ms), NOESY (mixing time 200 ms),and Heteronuclear Single Quantum Coherence and Heteronuclear MultipleBond Correlation (100-ms long range transfer delay). Capillary electrophoresis(CE)–MS was obtained on a 4000 QTRAP mass spectrometer (Applied Biosys-tems/MDS Sciex) with a Prince CE system (Prince Technologies) with a 90-cmlength bare fused-silica capillary using 15 mM ammonium acetate in deion-ized water, pH 9.0, as an injection module. A sheath solution (isopropanol/

methanol, 2:1) was delivered at a flow rate of 1.0 �L/min. A 5-kV electrosprayionization voltage was used for negative ion detection modes. For someexperiments samples were dephosphorylated at 1 mg/mL of 40% hydrofluoricacid (HF) for 24 h at 4 °C. MALDI-TOF mass spectra of the derivatized proteinsand the conjugates were obtained with an OmniFlex MALDI-TOF instrument(Bruker Daltonics) operated in the linear mode. Samples for analysis weredesalted, and 1 �L was mixed with 20 �L of sinnapinic acid matrix made in 30%CH3CN and 0.1% trifluroacetic acid. Next, 1 �L of mixture was dried on thesample stage and placed in the mass spectrometer.

Preparation of Conjugates. To 15 mg of BSA (Sigma) or rDT [CRM H21G (19)] in2.2mLofbufferA(PBS,0.1%glycerol,5mMEDTA,pH7.2),4mgofN-succinimidyl3-(bromoacetamido) propionate (SBAP; Pierce) in 40 �L of DMSO was added andreacted at pH 7.2 in room temperature with mixing for 1.5 h. Next, the solutionwas applied to a Sephadex G-50 column (1 � 50 cm) in PBS, and the void volumefraction (Pr-Br) was concentrated by using an Amicon Ultra-15 centrifuge filterdevice (Millipore) to 2.6 mL (13 mg recovered), and 0.1 mL was removed foranalysis. To 12 mg of Pr-Br in 2.4 mL of buffer A, 10 mg of O-(3-thiopropyl)hy-droxylamine was added in 300 �L of 1 M KCl and reacted at pH 7.2 in roomtemperature with mixing for 3 h. Next, the solution was passed through theSephadex G-50, and the void volume fraction (Pr-ONH2) was concentrated to 2.6mL as above, and 0.2 mL was removed for analysis. Ten milligrams of Pr-ONH2 wasreacted with 25 mg (7.8 �mol) O-SPC in 3 mL of buffer A overnight, at pH 7.2, inroom temperature with mixing. The solution was then passed through theSepharose G-75 (1 � 100 cm) in PBS, and the void volume fraction was collectedand analyzed for sugar and protein contents and molecular mass by MALDI-TOFand SDS/PAGE. Three conjugates were obtained this way: 1, BSA/O-SPC-F2; 2,rDT/O-SPC-F2; and 4, BSA/O-SPC-F3. For preparation of conjugate 3, binding ofrDT-ONH2 and O-SPC-F2 was done in 1.5 mL of buffer A, and the product(rDT/O-SPC-F2) contained twice the number of O-SPC-F2 chains per rDT moleculeas conjugate 2 (12 vs. 6). Conjugates of full-length O-SP bound by multipleattachments to either recombinant Bacillus anthracis protective antigen or rEPAwere prepared as described (7).

Immunization. Five- to 6-week-old female National Institutes of Health Swiss–Webstermice were injected s.c. 3 times at 2-week intervals with 2.5 �g ofsaccharide as a conjugate in 0.1 mL of PBS. Groups of 10 mice were exsangui-nated 7 days after the second or third injections (19). Controls received PBS.

Antibodies. Serum IgG antibodies were measured by ELISA using S. sonnei or P.shigelloides LPS as a coating antigen (20) The results were computed with anELISA data processing program provided by the Biostatistics and InformationManagement Branch, Centers of Disease Control, Atlanta (18). A polyclonalrabbit antiserum obtained by immunizing rabbits with multiple i.v. injections ofheat-killed S. sonnei bacteria was used in the immunodiffusion assays.

Competitive Inhibition ELISA. Inhibition was assayed by incubating conjugate-induced mice sera, diluted to the concentration that gave an absorbance of �1.0at A405, with 2, 10, or 50 �g of inhibitor (Table 2) per well, for 1 h at 37 °C andovernight at 4 °C. The ELISA was then continued as usual. Sera at the samedilution with and without inhibitor were compared. Percentage inhibition wasdefined as (A405 adsorbed serum/A405 nonadsorbed serum) � 100%. Core-free P.

Table 2. Composition and GM of serum IgG anti-S. sonnei LPS induced by O-SPC conjugates bound to BSA or rDT and by full-lengthO-SP bound to rEPA

No. Conjugate

Molecular massof conjugate,

kDa

Molecular massof sugar part,

kDa

No. O-SPCchains per

proteinmolecule

IgG, EU

Anti-LPS Anti-rDT

Secondinjection

Thirdinjection

Secondinjection

Thirdinjection

1 BSA/O-SPC-F2* 93.1 22.2 7 79 366 ND ND2 rDT/O-SPC-F2 80.6 18.5 6 5 392 2 913 rDT/O-SPC-F2 99.5 37.4 12 11 150 0.1 24 BSA/O-SPC-F3† 95.2 24.3 11 ND ND ND ND5 rEPA/O-SP‡ ND ND ND 11 67 ND ND

Mice (10 per group) were injected with 2.5 �g per mouse of saccharide as a conjugate, 3 times, 2 weeks apart and bled 1 week after the last two injections.ND, not determined.*S. sonnei O-SPC-F2 � core � average 3.5 RU of O-SP; average molecular mass � 3,206 Da.†S. sonnei O-SPC-F3 � core � average. 1.3 RU of O-SP; average molecular mass � 2,311 Da.‡S. sonnei O-SP � core � average 29 RU of O-SP bound by multipoint attachment.

Table 3. Competitive inhibition of antisera induced byBSA/O-SPC-F2 and by rPA/O-SP conjugates to the S. sonnei LPS,by different inhibitors

InhibitorAmount,

�g

Inhibition, %

Anti-S. sonneiBSA/O-SPC

Anti-S. sonneirPA/O-SP

S. sonnei O-SP 50 100 100S. sonnei O-SP 10 100 100S. sonnei O-SP 2 95 95P. shigelloides O-SP 50 100 100P. shigelloides O-SP 10 100 100P. shigelloides O-SP 2 96 94P. shigelloides O-SPno core* 50 100 100P. shigelloides O-SPno core 10 100 100P. shigelloides O-SPno core 2 92 90S. sonnei O-SPC-F2 50 100 100S. sonnei O-SPC-F2 10 81 84S. sonnei O-SPC-F2 2 10 19Hia CPS† 50 0 0Hia CPS 10 0 0Hia CPS 2 0 0

O-SP indicates high molecular mass saccharide (core � average 29 O-SP RU).O-SPC-F2 indicates low molecular mass saccharide (core � average. 3.5 O-SP RU.*Core was hydrolyzed by anhydrous HF (see Materials and Methods).†Hia CPS indicates H. influenzae type a capsular polysaccharide as control

Robbins et al. PNAS � May 12, 2009 � vol. 106 � no. 19 � 7977

MED

ICA

LSC

IEN

CES

Dow

nloa

ded

by g

uest

on

Mar

ch 1

7, 2

020

Page 5: Synthesis, characterization, and immunogenicity in mice of ...Synthesis, characterization, and immunogenicity in mice of Shigella sonneiO-specific oligosaccharide-core-protein conjugates

shigelloides O-SP was obtain by treating the O-SP with anhydrous HF for 1 h at25 °C as described (21). Haemophilus influenzae type a capsular polysaccharidewas used as a control.

Statistics. GM values of the groups of 10 mice were calculated. Unpaired t testwas used to compare GMs between different groups of mice.

ACKNOWLEDGMENTS. We thank Dr. Vince Pozsgay (National Institutes ofHealth, Bethesda, MD) for providing the O-(3-thiopropyl)hydroxylamine linker,Dr. Joseph Shiloach (National Institutes of Health, Bethesda, MD) for providingrDT; Dr. Bruce Coxon for assessing sample purities by NMR; Chunyan Guo fortechnical assistance; and Dr. Arthur Karpas for statistical analyses and manuscriptreview.ThisworkwassupportedbytheIntramuralResearchProgramoftheNationalInstitutes of Health, National Institute of Child Health and Human Development.

1. Szu SC, Robbins JB, Schneerson R, Pozsgay V, Chu C (2004) in New Generation Vaccines, edsLevine MM, Kaper JB, Rappuoli R, Liu MA, Good M (Dekker, New York), pp 471–478.

2. Ferreccio C, et al. (1991) Epidemiologic patterns of acute diarrhea and endemic Shigellainfections in children in a poor periurban setting in Santiago, Chile. Am J Epidemiol134:614–627.

3. Robbins JB, Chu C, Schneerson R (1992) Hypothesis for vaccine development: Protectiveimmunity to enteric diseases caused by nontyphoidal salmonellae and shigellae may beconferred by serum IgG antibodies to the O-specific polysaccharide of their lipopoly-saccharides. Clin Infect Dis 2:346–361.

4. Claesson B, et al. (2005) Antibodies against Haemophilus influenzae type b capsularpolysaccharide and tetanus toxoid before and after a booster dose of the carrierprotein 9 years after primary vaccination with a protein conjugate vaccine. PediatrInfect Dis J 24:463–464.

5. Taylor DN, et al. (1993) Synthesis, characterization, and clinical evaluation of conjugatevaccines composed of the O-specific polysaccharides of Shigella dysenteriae type 1,Shigella flexneri type 2a, and Shigella sonnei (Plesiomonas shigelloides) bound tobacterial toxoids. Infect Immun 61:3678–3687.

6. Cohen D, et al. (1997) Double-blinded vaccine-controlled randomized efficacy trial of aninvestigational Shigella sonnei conjugate vaccine in young adults. Lancet 349:155–159.

7. Passwell JH, et al. (2003) Safety and immunogenicity of Shigella sonnei-CRM9 andShigella flexneri type 2asucc conjugate vaccines in 1- to 4-year-old children. Pediatr InfDis J 22:701–706.

8. Pozsgay V, et al. (1999) Protein conjugates of synthetic saccharides elicit higher levelsof serum IgG lipopolysaccharide antibodies in mice than do those of the O-specificpolysaccharide from Shigella dysenteriae type 1. Proc Natl Acad Sci USA 96:5194–5197.

9. Kenne L, Lindberg B, Petersson K, Katzenellenbogen E, Romanowska E (1980) Struc-tural studies of the O-specific side-chains of the Shigella sonnei phase I lipopolysac-charide. Carbohydr Res 78:119–126.

10. Gamian A, Romanowaska E (1982) The core structure of Shigella sonnei lipopolysac-charide and the linkage between O-specific polysaccharide and the core region. EurJ Biochem 129:105–109.

11. Kubler-Kielb J, et al. (2008) Saccharide/protein conjugate vaccines for Bordetellaspecies: Preparation of saccharide, development of new conjugation procedures, andphysico-chemical and immunological characterization of the conjugates. Vaccine26:3587–3593.

12. Kubler-Kielb J, Schneerson R, Mocca C, Vinogradov E (2008) The elucidation of thestructure of the core part of the LPS from Plesiomonas shigelloides serotype O17expressing O-polysaccharide chain identical to the Shigella sonnei O-chain. CarbohydrRes 343:2701–2705.

13. Toth A, et al. (2000) Synthesis of the repeating unit of the O-specific polysaccharide ofShigella sonnei and quantitation of its serologic activity. Bioorg Med Chem Lett10:19–21.

14. Pozsgay V, Kubler-Kielb J, Schneerson R, Robbins JB (2007) Effect of the nonreducingend of Shigella dysenteriae type 1 O-specific oligosaccharides on their immunogenicityas conjugates in mice. Proc Natl Acad Sci USA 104:14478–14482.

15. Westphal O, Jann K, Himmelspach K (1983) Chemistry and immunochemistry of bac-terial lipopolysaccharides as cell wall antigens and endotoxins. Prog Allergy 33:9–39.

16. Lowry JO, Rosenbrough NJ, Farr AL, Randall R (1951) Protein measurement with theFolin phenol reagent. J Biol Chem 193:265–275.

17. Scott TA, Melvin EH (1953) Determination of dextran with anthrone. Anal Chem25:1656–1661.

18. Johnson VG, Nicholls PJ (1994) Histidine 21 does not play a major role in diphtheriatoxin catalysis. J Biol Chem 269:4349–4354.

19. Schneerson R, Barrera O, Sutton A, Robbins JB (1980) Preparation, characterization,and immunogenicity of Haemophilus influenzae type b polysaccharide-protein con-jugates. J Exp Med 152:361–376.

20. Plikaytis BD, Carlone GM (2005) Program ELISA for Windows User’s Manual (Centersfor Disease Control, Atlanta), Version 2.

21. Preston A, et al. (2006) Complete structure of Bordetella bronchiseptica and Bordetellaparapertussis lipopolysaccharides. J Biol Chem 281:18135–18144.

7978 � www.pnas.org�cgi�doi�10.1073�pnas.0900891106 Robbins et al.

Dow

nloa

ded

by g

uest

on

Mar

ch 1

7, 2

020