N-Glycosylation - The Sugar Code - page 108

issues that glycobiologists investigate. It is based on a particular evolutionary
history and intermingled with epigenetic pressure(s). Gene duplication events
generated a set of homologous glycosyltransferases that provided our predecessors
with redundant enzyme pools. Mutational changes switched either donor (blood
group AB) or acceptor substrate specifi cities of glycosyltransferases or, when
affecting gene regulation regions, their cell- and tissue - specifi c expression pat-
terns. A single point mutation in the respective glycosyltransferase [BGAT (P16442)]
induced blood group A-B switch. Such mutational events equipped prevertebrates
with a splendid glycosylation machinery serving all upcoming needs for successful
development. Note that biological functionalism for N - glycan structural versatility
often evolved later on. As a case in point, the immune system and brain profi ted
tremendously from this preexisting glycosylation machinery to synthesize func-
tional (poly-)sialylated, fucosylated, sulfated and/or O - acetylated (not covered in
this chapter) N -glycans to realize their tasks according to their functional
requirements.
Differences in glycan structures can be explained primarily through cell type-
specifi c expression patterns of glycosylation enzymes and of glycoproteins them-
selves. Human beings, in contrast to all other mammals including chimpanzees,
are not able to synthesize N -glycolylneuraminic acid and therefore N - acetylneur-
aminic acids (Neu5Ac) are the only sialic acids synthesized in our tissues. Glycans
on human serum α 1 -antitrypsin, secreted from human liver, are all terminated by

α 2,6-linked sialic acids and, therefore, lack α 2,3 - linked Neu5Ac in sLe x antigens.

However, in α 1 -antitrypsin synthesized by leukocytes, one out of three glycans

possesses 2,3 Neu5Ac in sLe x antigens. To refer to NCAM1, I am carrying poly-

Neu5Ac - modifi ed N -glycans in outgrowing neurons, but mostly sulfated ones in
quiescent cells. It is also known that particular motifs on a subset of proteins
render them a target for particular glycosylation steps. Hence, only lysosomal
proteins, but not for instance NCAM1, are modifi ed with a GlcNAc-P on defi ned
mannoses to guarantee proper targeting to lysosomes. On the other hand, I am
the main carrier for polyNeu5Ac, which is not found on lysosomal proteins at all.
Last, but not least, N -glycan structure can be controlled by earlier glycosyltransfer-
ase reactions. The bisecting GlcNAc transferase-III (MGAT3) will prevent GlcNAc
transferase-V (MGAT5) from acting. In conclusion, glycosylation patterns are cell-
type defi ned, may be dependent on developmental and growth stage, as well as
glycoprotein specifi c. As a consequence, transgenic glycoproteins may carry abnor-
mal, pathology - causing glycans [22] .
Databank entries about expressed sequence tags provide us with mRNA expres-
sion information for most enzymes in human tissues (unigene database at www.
ncbi.nlm.nih.gov). Data about glycan structures on individual proteins synthesized
in particular tissues are still rare, but if available, document a remarkable complex-
ity. For instance, 48 different complex- type N -glycans were characterized on a
subset of NCAM1 obtained from calf brain (Figure 6.1 for some details). An even
greater diversity was found in isolated human neutrophils. Here, 78 different
complex N -glycan structures have been characterized among others (www.
functionalglycomics.org). Such diversity results from glycosyltransferases compet-

α