Gene Cloning

Bioinformatics 237
These two regular expressions, describing the motifs for both the serine
and histidine active sites are diagnostic of trypsin-like serine proteases. If
sequences matching both are present then the protein can be unambiguously
assigned to the family. In situations like this where several motifs are
required to define a protein family these can be grouped together into what
are known as fingerprints. The fingerprint provides a signature for the protein
family. The PRINTS protein fingerprint database is a collection of these
fingerprints which can be searched in a variety of different ways. The prints
entry for the trypsin-like serine proteases consists of three elements
including both of the Prosite patterns that we have already discussed and a
third motif representing the active site aspartate.
Both of the above methods of representing the characteristic patterns of
amino acids typical of specific protein families are a compromise between
a very restricted expression, which is likely to miss some of the more divergent
examples, and a very fuzzy expression which may result in false identifications
of family members. These expressions are refined by an iterative
process of scanning the primary protein databases and evaluating the families
identified. An alternative solution is to preserve the full alignment
used to identify the family in the form of a matrix in which the frequency of
each amino acid at each position is recorded. These are known as profiles
and are found in addition to the patterns in the Prosite database.
The real power of these secondary or pattern databases is that they offer
a fast track to identify important structural and functional regions of proteins
and in doing so provide a link to a wealth of information about the
biological function of proteins. These patterns which are derived from
extensive alignments of many sequences may in some cases be able to
identify more distant relationships, and more divergent members of families,
than similarity searching. As the primary databases expand at ever
increasing rates, secondary databases can also help to overcome some of
the problems of “noise” created by the occurrence of many very similar
sequences. We have already come across an example of this when looking
at BLAST searches (Section 8.10). A search of the conserved domain database
(CDD) (Figure 8.13d) is run by default at the same time as a protein
BLAST at NCBI. This makes it possible to study the domain architecture of
the query protein even before the BLAST search is complete.
8.16 Investigating the Three-dimensional Structures of
Biological Molecules
The logical extension of studying the sequence of DNA and protein molecules
is to understand the three-dimensional structures that they adopt.
These three-dimensional structures must be determined experimentally
by techniques such as X-ray crystallography and nuclear magnetic resonance
(NMR) spectroscopy. This is a much more involved process than
determining sequence and is beyond the scope of this book. However,