DOI: 10.1002/cbic.200900388

Prokaryotic Protein Glycosylation Is Rapidly Expandingfrom “Curiosity” to “Ubiquity”Paul Messner*[a]

Since Bahl’s statement in 1992[1] that,“bacteria, in general, do not contain gly-coproteins, except the obligately halo-philic bacteria of the genus Halobacte-ria[2] and Bacillus stearothermophilus NRS2004/3a,[3] in which their presence hasbeen demonstrated with some degree ofcertainty”, a complete chance of percep-tion has taken place. Nowadays, the oc-currence of prokaryotic glycoproteins isfully accepted,[4, 5] particularly in patho-gens.[6, 7] Recent work in several laborato-ries has not only shown a great structur-al diversity of these glycoconjugates buthas also provided key advances in un-derstanding the molecular mechanismsof prokaryotic glycosylation, and the ap-plication potential and relevant immuno-logical aspects of selected glycopro-teins.[8–13] Among these prokaryotic gly-coconjugates are archaeal and bacterialsurface layer (S-layer) glycoproteins,[2, 14]

flagella and pili,[10, 15, 16] and several intra-and extracellular glycosylated enzymesand haptens.[4]

In the following, the different prokary-otic glycoproteins will be discussed inthe order of their discovery. GlycosylatedS-layer proteins were the first prokaryoticglycoproteins ever described in detail.Much of the current knowledge hasbeen derived from initial analyses of theN- and O-glycosylated S-layer protein ofHalobacterium salinarum by Strominger’sgroup. That work was then extended bySumper, Wieland, and co-workers in the1980s.[2] Another decade later, Eichlerand co-workers set out to investigatethe molecular mechanisms behind arch-

aeal S-layer protein glycosylation,[17] inparticular of the haloarchaeon Haloferaxvolcanii.[18] Recent studies on the metha-noarchaeon Methanococcus voltae re-vealed the structure of an N-glycanentity decorating several flagellin pro-teins as well as the S-layer protein.[19]

The similar modification of both flagel-lins and the S-layer glycoprotein pointsto a common N-glycosylation mecha-nism in these cells. In Hf. volcanii and Mc.voltae, several agl (archaeal glycosyla-tion) genes have been identified, andtheir products have been functionallycharacterized through combined gene-deletion and mass spectrometry ap-proaches.[20] In both organisms, aglB en-codes the sole component of the arACHTUNGTRENNUNGchae-ACHTUNGTRENNUNGal oligosaccharyltransferase (OTase) thatis responsible for the transfer of thecompleted N-glycan chain to the corre-sponding acceptor protein. Whereas N-glycosylation of archaeal proteins seemsto be more widespread, virtually nothingis known of the archaeal version of O-glycosylation, except the existence ofdisaccharides linked to threonine resi-dues of the S-layer proteins from Hb. sali-narum and Hf. volcanii.[2]

A completely different situation isknown from bacterial S-layer glycopro-teins. Since the first description of glyco-sylated bacterial S-layer proteins bySleytr and Thorne about three decadesago,[14] approximately 15 different, exclu-sively O-linked, S-layer glycan structureshave been fully or at least partially char-acterized in Gram-positive organisms.[12]

Inspection of these glycan structuressupports the assumption that the gly-cans are comparable to lipopolysacchar-ide (LPS) O-antigens of Gram-negativebacteria.[21] In some S-layer glycopro-teins, modifications of the terminal re-peating unit at the nonreducing end arepresent.[3, 12] These capping elements areassumed to be involved in determining

the chain length of the glycans, compa-rable to the ABC-transporter-dependentbiosynthesis pathway of LPS O-anti-gens.[21] Very recently, the existence ofglycosylated S-layer proteins has beendescribed in a Gram-negative bacterium,namely in the oral pathogen Tannerellaforsythia.[22]

In 1999 Guerry’s group, for the firsttime, described in Campylobacter jejuni81–176 genes that are part of a generalprotein glycosylation system, by whichup to 40 soluble and membrane proteinsare N-glycosylated. Since that discovery,the entire field has moved at an extraor-dinary pace.[7, 23] The genes required forthe general N-linked protein glycosyla-tion pathway (pgl) are remarkably con-served and do not seem to have theACHTUNGTRENNUNGpotential of phase variation. Because pglgene mutation affects the glycosylationof numerous proteins, there are pleio-tropic effects including decreased bacte-rial attachment and invasion of humanepithelial cells, loss of mouse and chickcolonization, and reduced protein reac-tivity to antisera.[24] Of particular interestwas the early observation that this N-gly-cosylation system could be functionallytransferred from C. jejuni into Escherichiacoli for the production of recombinantglycoproteins in this bacterium.[25] Thisachievement will open up numerouspossibilities for engineering novel glyco-conjugates for biotechnological applica-tions.

The biosynthesis, assembly and regu-lation of the flagellar apparatus of Bacte-ria and Archaea has been the subject ofextensive studies over many decades.The characterization of flagellar systemsfrom different bacterial species has re-vealed subtle yet distinct differences incomposition, regulation, and mode of as-sembly. Glycosylation of the major struc-tural protein, the flagellin, has recentlybeen shown to be an important feature

[a] Prof. Dr. P. MessnerDepartment f�r NanoBiotechnologieUniversit�t f�r Bodenkultur WienGregor-Mendel-Strasse 331180 Vienna (Austria)Fax: (+ 43) 1-4789112E-mail : paul.messner@boku.ac.at

ChemBioChem 2009, 10, 2151 – 2154 � 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim 2151

of numerous flagellar systems in bothBacteria and Archaea.[10]

The flagellin of Campylobacter coliVC167 was the first identified flagellarglycoprotein. Early studies indicated con-siderable heterogeneity in flagellin com-position, which was the result of asmany as 19 sites of O-linked glycosyla-tion on that protein.[26] The flagellinmodifications include pseudaminic acid(Pse5Ac7Ac), a nine-carbon sugar that isstructurally similar to sialic acid(Neu5Ac),[27] as well as a second nine-carbon sugar, legionaminic acid and itsderivatives.[28] Recent studies have indi-cated that glycosylation of flagellin iscrucial for filament assembly. The flagel-lin glycans also contribute to the viru-lence of C. jejuni. The loss of Pse5Am7Acfrom the C. jejuni 81–176 flagellin resultsin reduced adherence and invasion ofACHTUNGTRENNUNGintestinal epithelial cells and attenuationin ferret diarrheal disease.[10]

Glycosylation of flagellin is extensivethroughout the archaeal domain. De-tailed studies are scarce, but examplesare the flagellins of Hb. salinarum,Mc. voltae, and Mc. maripaludis.[10, 29] Thefirst structural characterization of a fla-ACHTUNGTRENNUNGgellar glycan was provided by Wielandand co-workers[2] for the flagellin ofHb. salinarum. The flagellins possess thesame N-linked glycans as the S-layer pro-teins. Similar observations were madewith the flagellins of the previously men-tioned methanococci, thus implying acommon N-linked glycosylation pathwayfor these two major surface proteinstructures in Archaea.[10, 29]

Pilin of pathogenic Neisseria was oneof the first examples of an O-glycosylat-ed glycoprotein in a bacterial pathogen.A series of investigations has identifiedthe genes that encode the glycosyltrans-ferases required for the biosynthesis ofthe pilin glycan.[30–34] Studies in differentNeisseria species have proposed that theaddition of O-linked glycan to pilin isperformed by an OTase that is homolo-gous to the O-antigen ligases that cata-lyze the addition of O-antigen toLPS.[34, 35]

In analogy to the well-accepted gener-al N-glycosylation system of Campylo-bacter jejuni,[7, 24] within one month in theautumn of 2008, three independent re-search groups submitted a description of

their specific model system as an exam-ple for a general prokaryotic protein O-glycosylation system. Two of thesepapers deal with the glycosylation sys-tems of pathogenic Neisseria strains[36, 37]

and will be discussed later.The third one is a seminal paper by

Comstock and colleagues who describea general O-glycosylation system that isimportant to the physiology of themajor human intestinal symbiont Bacter-oides fragilis.[38] The Bacteroides are spe-cifically adapted for survival in this eco-system and collectively comprise one ofthe most abundant bacterial genera inthe human colon. Members of thisgenus have an unprecedented repertoireof genetic systems devoted to acquiringand metabolizing carbohydrates. Thisallows the organisms to respond rapidlyto food supplies that are shifting andoften times scarce. The organisms pro-duce enzymes to harvest l-fucose fromhost mucosal glycans and harbor a rarebacterial pathway for the incorporationof exogenous fucose into capsular poly-saccharides and glycoproteins.[39] It wasshown that B. fragilis produces many gly-coproteins that bind the fucose-specificAleuria aurantia lectin. l-Fucose is anabundant surface molecule of host intes-tinal epithelial cells, and expression ofthis surface molecule is coordinated byhost and symbiont. The B. fragilis mutantDgmd-fclDfkp, defective in the pathwaysfor the biosynthesis of GDP-fucose andconsequently unable to incorporate l-fucose into glycoproteins, is also defec-tive in colonizing the mammalian intes-tine under competitive conditions. Thus,protein fucosylation is central to thephysiology of this organism and is nec-essary for it to competitively colonizethe mammalian intestine.[39] The proteinstargeted for glycosylation include thosepredicted to be involved in protein fold-ing, protein–protein interactions, peptidedegradation, and surface lipoproteins. Toelucidate their localizations, whole bacte-ria were treated with proteinase K so asto digest surface proteins. As a result, allof the eight selected glycoproteins aresecreted from the cytoplasm and target-ed either to the periplasm or the inneror outer leaflet of the outer membrane(Figure 1). Glycosylation of periplasmicproteins is unusual. So far, it has only

been observed for glycoproteins of Cam-pylobacter jejuni,[7] Desulfovibrio gigas,[40]

and Neisseria gonorrhoeae.[37] By site-di-rected mutagenesis it was further dem-onstrated that on a selected glycopro-tein (BF2494), which is representative ofall other identified glycoproteins, theamino acids Thr87, Thr178, and Thr231are the only glycosylation sites and,therefore, all the glycans of this proteinare O-linked.[38] Although O-glycosylationsystems in other organisms usually donot require a motif other than Ser, Thr,[8]

or Tyr,[4, 12] inspection of the protein se-quence around the three glycosylationsites of BF2494 revealed that in B. fragilis,a three-residue motif D–S/T–A/I/L/V/M/Tfor O-glycosylation is present.

Glycosylation of bacterial proteinsgenerally requires the transport of thetarget proteins from the cytoplasm intothe bacterial periplasm.[3, 7] This criterionwas also tested with B. fragilis ; mutationof the signal peptide of BF2494 in B. fra-gilis showed that the mutant protein isretained in the cytoplasm and that it isnot glycosylated.[38] To test whether ornot a genomic region of B. fragilis isACHTUNGTRENNUNGinvolved in protein glycosylation, thewhole genome was inspected. Theregion spanning genes BF4298 toBF4306 encodes a putative flippase, fiveputative glycosyltransferases, and othergenes likely to be involved in oligosac-charide synthesis, but no polymerase.This demonstrates that the BF4298–4306region is involved in protein glycosyla-tion. Furthermore, because this region isrequired for the synthesis of all the fuco-sylated glycoproteins of B. fragilis, it ap-pears to be part of a general glycosyla-tion system. Based on these results, theregion was named lfg for locus of fragilisglycosylation.[38] Interestingly, the lfgregion is adjacent to metG, which en-codes the essential enzyme methionyl-tRNA synthetase. The transcriptional link-age of the protein translation machinery(metG) with the protein glycosylationmachinery (lfg) suggests a high level ofimportance for protein glycosylation inB. fragilis. The analysis of five other intes-tinal Bacteroides species demonstratedthat they all have regions similar to lfg,with a flippase gene downstream frommetG and several glycosyltransferasegenes. This finding, coupled with the ob-

2152 www.chembiochem.org � 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemBioChem 2009, 10, 2151 – 2154

P. Messner

servation that other Bacteroides speciesproduce fucosylated glycoproteins, wasa strong hint for the presence of a con-served general O-glycosylation system inB. fragilis and in other intestinal Bacter-oides species. Alignment of the sequenceof BF2494 with the sequences from theother five orthologues confirmed thatthey perfectly match with the glycosyla-tion motif D–S/T–A/I/L/V/M/T. These datasuggest that the intestinal Bacteroidesspecies possess similar general O-glyco-sylation systems, one of the few generalglycosylation systems known in bacte-ria.[7, 36–38]

The mechanism of O-glycosylation inB. fragilis appears to have some similari-ties to protein glycosylation in other bac-teria (Figure 1). It appears to be similarto pilin O-glycosylation in Neisseria men-ingitidis and N. gonorrhoeae, and in somestrains of Pseudomonas aeruginosa,[41]

and to a recently described S-layer O-gly-cosylation system of Geobacillus stearo-thermophilus,[3] but different from O-gly-cosylation in eukaryotes.[7, 38, 42]

As mentioned before, similar to theproposed general O-glycosylation ofB. fragilis, in pathogenic Neisseria menin-gitidis, an O-glycosylation pathway modi-fies both a single abundant protein,pilin, the subunit protein that forms pili,and an outer membrane glycoprotein,

the nitrite reductase AniA.[36] Since pilinis under intense immune selection and isthe archetypal example for antigenic var-iation, the decoration of the surface-exposed, flexible, C-terminal domain ofAniA with the same phase-variable O-linked glycan modification might also bean immune evasion strategy. Immune se-lection acting on the surface proteins ofthis host-adapted pathogen might havebeen the driving force for the evolutionof this proposed general O-glycosylationpathway.[36]

In the related Gram-negative bacteri-um Neisseria gonorrhoeae, a general O-linked glycosylation system was pro-posed that targets structurally and func-tionally diverse groups of membrane-as-sociated proteins.[37] The analyzed glyco-proteins are implicated in activities thatvary from protein folding, disulfide bondformation, and solute uptake to bothaerobic and anaerobic respiration. As ineukaryotes, the broad scope of thissystem is dictated by the relaxed specif-icity of the OTase as well as the bulkproperties and context of the protein-targeting signal rather than by a strictamino acid consensus sequence. Togeth-er, these findings reveal previously un-recognized commonalities connecting O-linked protein glycosylation in distantlyrelated life forms.[37] They also extend

the ways in which the N. gonorrhoeae O-linked system emulates the C. jejuni N-linked systems to include the targetingof multiple periplasmic substrates.[7, 42]

As in the C. jejuni system, these find-ings raise obvious questions as to whatbiologic significance global protein gly-cosylation might have and what forcesand processes have shaped glycopro-teome content.[37] The great importanceand potential for exploitation in this areaof prokaryotic protein glycosylation arecertainly very clear.

Acknowledgements

I thank Christina Sch�ffer for her input theduring preparation of this contribution,my colleagues in the NanoGlycobiologygroup for sharing ideas, and Uwe B. Sleytrfor his interest. Work in the laboratory issupported by the Austrian Science Fundand by the Federal Ministry of Science andEducation.

Keywords: Archaea · Bacteria · N-glycosylation · O-glycosylation ·prokaryotic glycoproteins

[1] O. P. Bahl in Glycoconjugates–Composition,Structure, and Function (Eds. : H. J. Allen, E. C.Kisailus), Marcel Dekker, New York, 1992,pp. 1–12.

[2] M. Sumper, F. T. Wieland in Glycoproteins(Eds. : J. Montreuil, J. F. G. Vliegenthart, H.Schachter), Elsevier, Amsterdam, 1995,pp. 455–473.

[3] K. Steiner, R. Novotny, D. B. Werz, K. Zarsch-ler, P. H. Seeberger, A. Hofinger, P. Kosma, C.Sch�ffer, P. Messner, J. Biol. Chem. 2008, 283,21120–21133.

[4] P. Messner, C. Sch�ffer in Progress in theChemistry of Organic Natural Products, Vol. 85(Eds. : W. Herz, H. Falk, G. W. Kirby), Springer,Wien, 2003, pp. 51–124.

[5] P. Messner, J. Bacteriol. 2004, 186, 2517–2519.

[6] I. Benz, M. A. Schmidt, Mol. Microbiol. 2002,45, 267–276.

[7] C. M. Szymanski, B. W. Wren, Nat. Rev. Micro-biol. 2005, 3, 225–237.

[8] J. Eichler, M. W. W. Adams, Microbiol. Mol.Biol. Rev. 2005, 69, 393–425.

[9] M. F. Feldman, M. Wacker, M. Hernandez,P. G. Hitchen, C. L. Marolda, M. Kowarik, H. R.Morris, A. Dell, M. A. Valvano, M. Aebi, Proc.Natl. Acad. Sci. USA 2005, 102, 3016–3021.

[10] S. M. Logan, Microbiology 2006, 152, 1249–1262.

[11] C. M. Fletcher, M. J. Coyne, D. L. Bentley, O. F.Villa, L. E. Comstock, Proc. Natl. Acad. Sci. USA2007, 104, 2413–2418.

Figure 1. Proposed model for the general O-glycosylation system of Bacteroides fragilis. The glycanchain is synthesized on a lipid carrier on the inner side of the cytoplasmic membrane by the sequentialaction of five putative glycosyltransferases and a fucosyltransferase not encoded by the lfg region. Theglycan is flipped to the periplasmic face of the inner membrane by Wzx (blue). Up to this point, thepathway is common to the synthesis of many O-antigens, capsular polysaccharides and bacterial glyco-proteins. The glycosylation of proteins in B. fragilis occurs in the periplasm, by the activity of an unde-fined OTase (gray), which is likely an integral membrane protein as in other bacteria. This schematicshows the glycosylated proteins (orange) in the three cellular locations that have been experimentallydemonstrated. Modified from ref. [38] .

ChemBioChem 2009, 10, 2151 – 2154 � 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.chembiochem.org 2153

Prokaryotic Protein Glycosylation

[12] P. Messner, K. Steiner, K. Zarschler, C. Sch�ffer,Carbohydr. Res. 2008, 343, 1934–1951.

[13] S. Yurist-Doutsch, B. Chaban, D. J. VanDyke,K. F. Jarrell, J. Eichler, Mol. Microbiol. 2008,68, 1079–1084.

[14] U. B. Sleytr, K. J. I. Thorne, J. Bacteriol. 1976,126, 377–383.

[15] P. Castric, Microbiology 1995, 141, 1247–1254.

[16] E. Stimson, M. Virji, K. Makepeace, A. Dell,H. R. Morris, G. Payne, J. R. Saunders, M. P.Jennings, S. Barker, M. Panico, I. Blench, E. R.Moxon, Mol. Microbiol. 1995, 17, 1201–1214.

[17] M. Abu-Qarn, J. Eichler, Mol. Microbiol. 2006,61, 511–525.

[18] M. Abu-Qarn, F. Giordano, A. Battaglia, A.Trauner, P. G. Hitchen, H. R. Morris, A. Dell, J.Eichler, J. Bacteriol. 2008, 190, 3140–3146.

[19] S. Voisin, R. S. Houliston, J. Kelly, J.-R. Brisson,D. Watson, S. L. Bardy, K. F. Jarrell, S. M.Logan, J. Biol. Chem. 2005, 280, 16586–16593.

[20] B. Chaban, S. Voisin, J. Kelly, S. M. Logan, K. F.Jarrell, Mol. Microbiol. 2006, 61, 259–268.

[21] C. R. H. Raetz, C. Whitfield, Annu. Rev. Bio-chem. 2002, 71, 635–700.

[22] S.-W. Lee, M. Sabet, H.-S. Um, J. Yang, H. C.Kim, W. Zhu, Gene 2006, 371, 102–111.

[23] P. Guerry, C. M. Szymanski, Trends Microbiol.2008, 16, 428–435.

[24] H. Nothaft, S. Amber, M. Aebi, C. M. Szyman-ski in Campylobacter, 3rd ed. (Eds. : I. Na-chamkin, C. M. Szymanski, M. J. Blaser), ASMPress, Washington, DC, 2008, pp. 447–469.

[25] M. Wacker, D. Linton, P. G. Hitchen, M. Nita-Lazar, S. M. Haslam, S. J. North, M. Panico,H. R. Morris, A. Dell, B. W. Wren, M. Aebi, Sci-

ence 2002, 298, 1790–1793.[26] S. M. Logan, J. F. Kelly, P. Thibault, C. P. Ewing,

P. Guerry, Mol. Microbiol. 2002, 46, 587–597.[27] P. Thibault, S. M. Logan, J. F. Kelly, J.-R. Bris-

son, C. P. Ewing, T. J. Trust, P. Guerry, J. Biol.Chem. 2001, 276, 34862–34870.

[28] D. J. McNally, A. J. Aubry, J. P. M. Hui, N. H.Khieu, D. Whitfield, C. P. Ewing, P. Guerry, J.-

R. Brisson, S. M. Logan, E. C. Soo, J. Biol.Chem. 2007, 282, 14463–14475.

[29] D. J. VanDyke, J. Wu, S. M. Logan, J. F. Kelly, S.Mizuno, S.-I. Aizawa, K. F. Jarrell, Mol. Micro-

biol. 2009, 72, 633–644.[30] E. Stimson, M. Virji, K. Makepeace, A. Dell,

H. R. Morris, G. Payne, J. R. Saunders, M. P.Jennings, S. Barker, M. Panico, I. Blench, E. R.

Moxon, Mol. Microbiol. 1995, 17, 1201–1214.[31] M. Virji, Gene 1997, 192, 141–147.[32] M. Marceau, K. Forest, J.-L. Beretti, J. Tainer,

X. Nassif, Mol. Microbiol. 1998, 27, 705–715.

[33] A. Banerjee, S. K. Ghosh, Mol. Cell. Biochem.2003, 253, 179–190.

[34] P. M. Power, K. L. Seib, M. P. Jennings, Bio-chem. Biophys. Res. Commun. 2006, 347,904–908.

[35] F. E. Aas, �. Vik, J. Vedde, M. Koomey, W.Egge-Jacobsen, Mol. Microbiol. 2007, 65,607–624.

[36] S. C. Ku, B. L. Schulz, P. M. Power, M. P. Jen-nings, Biochem. Biophys. Res. Commun. 2009,378, 84–89.

[37] �. Vik, F. E. Aas, J. H. Anonsen, S. Bilsborough,A. Schneider, W. Egge-Jacobsen, M. Koomey,Proc. Natl. Acad. Sci. USA 2009, 106, 4447–4452.

[38] C. M. Fletcher, M. J. Coyne, O. F. Villa, M.Chatzidaki-Livanis, L. E. Comstock, Cell 2009,137, 321–331.

[39] M. J. Coyne, B. Reinap, M. M. Lee, L. E. Com-stock, Science 2005, 307, 1778–1781.

[40] T. Santos-Silva, J. M. Dias, A. Dolla, M. C.Durand, L. L. Goncalves, J. Lampreia, I.Moura, M. J. Romao, J. Mol. Biol. 2007, 370,659–673.

[41] A. Faridmoayer, M. A. Fentabil, D. C. Mills,J. S. Klassen, M. F. Feldman, J. Bacteriol. 2007,189, 8088–8098.

[42] A. Helenius, M. Aebi, Annu. Rev. Biochem.2004, 73, 1019–1049.

Received: June 25, 2009Published online on July 20, 2009

2154 www.chembiochem.org � 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemBioChem 2009, 10, 2151 – 2154

P. Messner