trans

ANTIGENIC VARIATION AT THE STRUCTURAL AND FUNCTIONAL LEVEL 93
trypanosomes to humans. The core structure –
ethanolamine–phosphate–6Man1–2Man1–
6Man1–4GlcNH 2 1–6myo-inositol – does not
vary, but the lipid constituents can vary greatly
from cell to cell. The core glycan structure
can be extensively decorated with additional
residues of ethanolamine, sugars and sialic
acids, which can vary among different proteins
synthesized by the same cell. Perhaps the most
unique feature of the VSG GPI anchor is its lipid
moiety: dimyristyl glycerol. GPI is synthesized
by sequential addition of components to cellular
phosphatidyl-inositol (PI), to generate a
GPI precursor that contains no myristic acid.
This precursor is remodeled by the sequential
replacement of each acyl chain by myristic acid.
Another myristate remodeling activity works
on GPI anchors already attached to VSG. Thus,
for reasons that are not understood, T. brucei
goes to quite remarkable lengths to ensure that
its VSG is anchored in the membrane by this C 14
fatty acid, which comprises only 1% of serum
lipids (see Chapter 10).
The release of GPI-anchored VSG, when
the trypanosome membrane is disrupted, is
due to the presence of a potent phospholipase
C (PIPLC) that cleaves the diacylglycerol–
phosphoinositol linkage, releasing soluble VSG
that contains the hydrophilic components of
the GPI anchor. One consequence of this cleavage
is the formation of a terminal inositol cyclic
phosphodiester, which is a major contributor
to the cross-reactivity of rabbit anti-VSG antibodies.
Despite much speculation and experimentation,
the biological role of this PIPLC is
unknown, as is the manner in which it comes
into contact with VSG following membrane
disruption. Its cellular location is unresolved,
although it appears to be acylated and attached
to the membrane of otherwise anonymous
intracellular organelles. Only VSG already at
the surface is susceptible; VSG in the secretory
pathway appears to be inaccessible. Although
it first appeared to be specific for GPI, rather
than simply for PI, this may have been a misleading
consequence of detergent concentrations
and other components of the assay.
When trypanosomes pass from the mammalian
bloodstream into the tsetse midgut,
the VSG coat is shed and replaced by a similarly
dense coat formed from a small family
of midgut-stage proteins called procyclins.
These proteins, also known by the acronym
PARP, for procyclic acidic repetitive proteins,
consist largely of Glu-Pro or Gly-Pro-Glu-Glu-
Thr repeats, and may serve to protect the procyclic
stage from proteases in the insect midgut.
Intriguingly, the procyclins are also GPIanchored,
but the GPI structure is distinct
from that on the VSG. It is not myristylated, it
contains monoacylglycerol, and the inositol is
palmitylated, rendering it resistant to PIPLC.
The idea that the PIPLC was involved in replacement
of the VSG coat during this developmental
transition was intriguing, but apparently
misleading. When the PIPLC gene is deleted,
trypanosomes do not appear to be impaired
in their ability to complete their life cycle in
the tsetse. In contrast, developmental shedding
of the VSG coat appears to be mediated
by proteolytic cleavage. Interestingly, disruption
of an essential gene in the pathway for
GPI synthesis is lethal to bloodstream-form
T. brucei but not to the procyclic forms, which
can apparently survive when deprived of all
GPI-linked proteins. This observation, coupled
with the many differences between GPI
synthesis in trypanosomes and human cells,
validates this pathway as a potential target for
drug development.
Calculations and physical studies suggest
that the trypanosome coat is composed of a
monolayer of VSG molecules packed as tightly
as possible on the surface membrane. The evidence
for this view of the coat comes from
various experimental observations. Where VSG
MOLECULAR BIOLOGY