Over 43% (230 out of 524) of the currently available bacterial genomes have
more than one type of recognizable ferritin encoded in the DNA, indicating the
importance of these proteins. This percentage is lower in archaea; only 2% of the
sequenced genomes encode more than one type of ferritin, which was exclusively a
combination of bacterioferritin and maxi-ferritin (Fig. 2.2b). While multiple gene
copies of the same type of ferritin occur rarely in archaea (Table 2.2 and Fig. 2.2),
many bacteria have multiple genes encoding functional ferritin proteins.
2.2.2 Ferritin in Eukaryotes
The distribution of ferritin genes in eukaryotes is fairly different; mini-ferritins
and heme-ferritins (BFR) have not been found to date. Studying the distribution
of ferritin genes in eukaryotic genomes using BLAST alignments is less straightforward than in prokaryotic genomes for a number of reasons: (1) the presence
of introns; (2) large numbers of truncated or pseudogenes; (3) multiple gene
annotations. Eukaryotic maxi-ferritins are mixtures of subunits encoded in different
genes. In plants, all the genes encode H-type subunits that are catalytically active. The
multiple names for the catalytic sites include oxidoreductase, ferroxidase, F ox , FC.
Ferritin cages in animals are co-assembled from active H-type subunits or
inactive subunits called L-subunits. Historically, H and L stands for heart, heavy,
or higher; L stands for liver, light or lower, referring to the organ where the subunit
is dominant or the mass, or migration in an SDS-PAGE gel. However, the only
consistent parameter is catalytic activity since H and L combination are found in all
tissues, and SDS gel mobility does not always relate to mass. Multiple genes
encoding H subunits occur in fish, frogs, mice, humans, maize, soy, etc. where
the H subunits are called H and M (frog, fish) (Dickey et al. 1987; Yamashita et al.
1996) or H and mitochondrial (human, mouse) (Corsi et al. 2002; Arosio et al.
2009) or H-1-4/ATF 1–4 (plant ferritins) (Dong et al. 2008; Briat et al. 2009). Some
ferritins have been identified from intronless genes that possibly result from the
high stability of the animal ferritin mRNA enabling copying by viral reverse
transcriptases during infection. In prokaryotes, multiple ferritin genes are expressed
at different times in the culture cycle or with different stimuli so that the bacterial
ferritin nanocages assemble usually only one type of subunit.
To determine the distribution of eukaryotic ferritin sequences in Nature, human
H-ferritin was used in a BLAST search against all eukaryotic curated protein
sequences (the UniProtKB/Swiss-Prot database, Expasy) in order to generate
alignments of animal and plant ferritins to get insight into the conservation of
active site residues among the A and the B site. A total of 55 homologues sequences
were identified (sequence identity ranging from 22% to 99.45% and similarity from
40% to 100%, P-value < 67 Â 10
À11 ) and the conserved active sites and variants
are shown in Table 2.3. Inherent in the analysis is the constraint imposed by the
absence of structural information, which will exclude therefore many ferritins. Such
limits of the search tools can explain the apparent absence of ferritin in organisms
34
L.E. Bevers and E.C. Theil
more than one type of recognizable ferritin encoded in the DNA, indicating the
importance of these proteins. This percentage is lower in archaea; only 2% of the
sequenced genomes encode more than one type of ferritin, which was exclusively a
combination of bacterioferritin and maxi-ferritin (Fig. 2.2b). While multiple gene
copies of the same type of ferritin occur rarely in archaea (Table 2.2 and Fig. 2.2),
many bacteria have multiple genes encoding functional ferritin proteins.
2.2.2 Ferritin in Eukaryotes
The distribution of ferritin genes in eukaryotes is fairly different; mini-ferritins
and heme-ferritins (BFR) have not been found to date. Studying the distribution
of ferritin genes in eukaryotic genomes using BLAST alignments is less straightforward than in prokaryotic genomes for a number of reasons: (1) the presence
of introns; (2) large numbers of truncated or pseudogenes; (3) multiple gene
annotations. Eukaryotic maxi-ferritins are mixtures of subunits encoded in different
genes. In plants, all the genes encode H-type subunits that are catalytically active. The
multiple names for the catalytic sites include oxidoreductase, ferroxidase, F ox , FC.
Ferritin cages in animals are co-assembled from active H-type subunits or
inactive subunits called L-subunits. Historically, H and L stands for heart, heavy,
or higher; L stands for liver, light or lower, referring to the organ where the subunit
is dominant or the mass, or migration in an SDS-PAGE gel. However, the only
consistent parameter is catalytic activity since H and L combination are found in all
tissues, and SDS gel mobility does not always relate to mass. Multiple genes
encoding H subunits occur in fish, frogs, mice, humans, maize, soy, etc. where
the H subunits are called H and M (frog, fish) (Dickey et al. 1987; Yamashita et al.
1996) or H and mitochondrial (human, mouse) (Corsi et al. 2002; Arosio et al.
2009) or H-1-4/ATF 1–4 (plant ferritins) (Dong et al. 2008; Briat et al. 2009). Some
ferritins have been identified from intronless genes that possibly result from the
high stability of the animal ferritin mRNA enabling copying by viral reverse
transcriptases during infection. In prokaryotes, multiple ferritin genes are expressed
at different times in the culture cycle or with different stimuli so that the bacterial
ferritin nanocages assemble usually only one type of subunit.
To determine the distribution of eukaryotic ferritin sequences in Nature, human
H-ferritin was used in a BLAST search against all eukaryotic curated protein
sequences (the UniProtKB/Swiss-Prot database, Expasy) in order to generate
alignments of animal and plant ferritins to get insight into the conservation of
active site residues among the A and the B site. A total of 55 homologues sequences
were identified (sequence identity ranging from 22% to 99.45% and similarity from
40% to 100%, P-value < 67 Â 10
À11 ) and the conserved active sites and variants
are shown in Table 2.3. Inherent in the analysis is the constraint imposed by the
absence of structural information, which will exclude therefore many ferritins. Such
limits of the search tools can explain the apparent absence of ferritin in organisms
34
L.E. Bevers and E.C. Theil