Biology Reference
In-Depth Information
quench better than do brominated lipids in the more distant bulk lipid bilayer. However,
when similar experiments were repeated using
2
H-NMR instead of a spin- or brominated
probe, no difference between annular and bulk bilayer lipids could be detected. This
suggests that a distinct ring of annular lipids does not exist. Discrepancies between the
experiments are attributed to the inherently slower times that are detectable by NMR
(~10
6
sec) compared to ESR and fluorescence. Therefore, it was concluded that annular
lipids, distinct from bulk bilayer lipids, may exist for short periods of time (
10
6
sec). If
annular lipids do exist, their association with integral proteins must be transient and hence
very weak. A transient existence suggests that lipid composition of annular lipid closely
track composition of the bulk bilayer.
For decades searches have been conducted to find binding sites for specific lipids on the
surface of membrane integral proteins. Of particular interest would be relatively long-lived
lipid protein complexes that are stable on the order of the time required for the turnover of
a typical enzyme (10
3
sec). The large numbers of different lipid species that comprise
a typical membrane (see 'lipid diversity' discussed in Chapters 4 and 5) strongly argue for
the existence of these complexes. A large number of lipid molecular species would be
required to fit the many different binding sites that might exist on the surfaces of membrane
proteins. Such binding sites would help to explain the perplexing problem of 'lipid diversity'.
However, the existence of specific, tight lipid-binding sites and weak binding of annular
lipids seem contradictory.
Another related question that had to be addressed concerned the number of phospho-
lipids required to completely solvate a specific integral protein and the lipid specificity, if
any, of the solvation. In his 1988 topic Biomembranes: Molecular Structure and Function
[42]
,
Robert Gennis presented a crude calculation to estimate the number of phospholipids that
might surround a 'typical' membrane integral protein. Some assumptions and estimates
had to be employed. The 'typical' protein was assigned a molecular weight of 50,000 and
comprised 50% of the membrane by weight. This is well within the range of 20
<
75% protein
for membranes (Chapter 1). Assuming an average molecular weight of ~830 for a phospho-
lipid, the phospholipid-protein ratio was 60:1 by number, again a realistic assumption. The
protein was envisioned to be a cylinder that extended 10
˚
above and below the bilayer
surface, and the radius of the cylinder was 18
˚
. Since phospholipids have a radius of
~4.4
˚
, it would require ~16 phospholipids on each leaflet of the bilayer to completely solvate
the protein. Therefore a total of ~32 lipids would be required to form the annulus. That would
leave only 28 phospholipids left to form the next lipid layer around the protein. Since this
second layer must be larger than the annular layer, there are not enough lipids to complete
a second lipid shell.
Actual experiments agree with Gennis's rough estimate. For example, the integral
membrane enzyme, Ca
2
þ
ATPase was isolated free of phospholipids. Phospholipids were
then slowly added back and the protein's activity monitored
[43]
. Activity did not return until
the lipid-protein ratio reached ~30:1 (
Figure 10.18
). The assumption was that, at this ratio,
activity returned because the proteinwas completely solvated by lipid. Additional lipid added
to the growing bilayer did not further increase activity. Confirming experiments were done
adding a spin-labeled phospholipid to a de-lipidated protein
[44]
. The process was followed
using ESR. Initially the ESR spectra were broad, indicating close association of the lipid and
protein. At some lipid-protein ratio all of the annular lipid sites became filled. Subsequent lipid
e
