Biology Reference
In-Depth Information
exosite interaction with a membrane component and that it is in this compartmen-
talized state that the proteinases act on their target substrates. The MT-MMPs have
a built-in means to be compartmentalized at the cell surface. If this anchored-to-the-
cell concept is indeed true, then the substrates must also be nearby. Indeed, many of
the confirmed, physiologic substrates of MMPs are membrane proteins themselves.
1.3.2
Identifying Substrates
To define the function of a given MMP produced by a specific cell type, it is critical
to both identify its protein substrate and determine the consequence of that substrate
being proteolyzed, be it gain- or loss-of-function. Because MMPs do not act on
consensus cleavage sites, candidate substrates cannot be selected by in silico
analysis. Identifying MMP substrates has been accomplished using various strate-
gies. Possibly the most common approach has been to incubate a purified active
MMP with a purified candidate substrate under optimal in vitro conditions and
assess if the target protein is cleaved or degraded. However, this approach tells us
only what an MMP can do, not what it does do. The shortcoming of relying on this
approach is that in a test tube most MMPs are nonspecific and can cleave proteins
they may not encounter in vivo. (Papain is an illustrative example of this concern.
This enzyme is used to cleave animal proteins in a lab setting, although it is unlikely
that papain evolved in papaya to perform such functions in the living plant.)
Although in vitro proteolysis assays are easy and an essential tool for verification,
most of the proteins identified as MMP substrates
only by this approach
probably
are not physiologic targets of the enzymes.
1.3.3
In Vivo Observations
A striking observation with many MMP knockout mice is that, with very few
exceptions, they do not reveal profound developmental phenotypes. In contrast,
mice that lack key structural ECM components and related proteins, such as type I
or IV collagens, elastin, fibronectin, various laminin chains, and lysyl oxidase,
among others, have profound tissue defects or, for many, embryonic lethality,
mirroring the severity of human conditions with mutations in these genes (Byers
1990
; Curren et al.
1993
; Gubler
2008
; Li et al.
1998
; Nicolae and Olsen
2010
). In
contrast, most MMP null mice have no to only minor phenotypes in the unchal-
lenged state and to date, no human condition has been linked to mutation in an
MMP gene (Puente et al.
2003
). If MMPs are indeed critical for remodeling in
development (as MMP14 is, see below) or for ECM turnover in homeostasis, then
more profound phenotypes would be expected if they are mutated.
The generalized lack of developmental phenotypes among most MMP knock-
out mice is not surprising. Typically, MMPs are not expressed in normal, healthy
