Strategize a Smooth Tenant-to-tenant Migration and Copilot Takeoff
Gutell 083.jmb.2002.321.0215
1. Modeling a Minimal Ribosome Based on Comparative
Sequence Analysis
Jason A. Mears1
, Jamie J. Cannone2
, Scott M. Stagg1
, Robin R. Gutell2
Rajendra K. Agrawal3,4
and Stephen C. Harvey1
*
1
Department of Biochemistry
and Molecular Genetics
University of Alabama at
Birmingham, Birmingham, AL
35295-0005, USA
2
Institute for Cellular and
Molecular Biology and Section
of Integrative Biology
University of Texas at Austin
2500 Speedway, Austin, TX
78712-1095, USA
3
Wadsworth Center, New York
State Department of Health
State University of New York at
Albany, Empire State Plaza
P.O. Box 509, Albany, NY
12201-0509, USA
4
Department of Biomedical
Sciences, State University of
New York at Albany, Empire
State Plaza, P.O. Box 509
Albany, NY 12201-0509, USA
We have determined the three-dimensional organization of ribosomal
RNAs and proteins essential for minimal ribosome function. Comparative
sequence analysis identifies regions of the ribosome that have been evolu-
tionarily conserved, and the spatial organization of conserved domains is
determined by mapping these onto structures of the 30 S and 50 S sub-
units determined by X-ray crystallography. Several functional domains of
the ribosome are conserved in their three-dimensional organization in
the Archaea, Bacteria, Eucaryotic nuclear, mitochondria and chloroplast
ribosomes. In contrast, other regions from both subunits have shifted
their position in three-dimensional space during evolution, including the
L11 binding domain and the a-sarcin–ricin loop (SRL). We examined con-
served bridge interactions between the two ribosomal subunits, giving an
indication of which contacts are more significant. The tRNA contacts that
are conserved were also determined, highlighting functional interactions
as the tRNA moves through the ribosome during protein synthesis. To
augment these studies of a large collection of comparative structural
models sampled from all major branches on the phylogenetic tree, Caenor-
habditis elegans mitochondrial rRNA is considered individually because it
is among the smallest rRNA sequences known. The C. elegans model sup-
ports the large collection of comparative structure models while providing
insight into the evolution of mitochondrial ribosomes.
q 2002 Elsevier Science Ltd. All rights reserved
Keywords: ribosome; conservation; rRNA; evolution; phylogenetics*Corresponding author
Introduction
The ribosome is the site of protein synthesis in
all living cells. It is composed of two subunits that
combine to form a functional ribosome, which is a
70 S particle in bacterial cells. In Escherichia coli,
the smaller 30 S subunit is composed of the 16 S
rRNA and 21 proteins, while the larger 50 S sub-
unit contains 23 S and 5 S rRNAs along with 31
proteins. It was long believed that ribosomal pro-
teins were the principal catalysts for translation,
but the discovery of catalytic RNAs1,2
suggested
that RNA might provide the catalytic sites. This
idea was confirmed when the crystal structure of
the 50 S subunit revealed an all-RNA peptidyl
transferase site, indicating that the ribosome is a
ribozyme.3
Proofreading of mRNA takes place at
the mRNA binding site on the 30 S subunit, and
RNA is implicated in this process as well.4 – 6
The
large subunit is monolithic and is believed to act
as a relatively static platform, while the small sub-
unit has been shown to be a dynamic structure
that moves during translation in a ratchet-like
manner.7,8
While proofreading of the codon/anticodon
interaction and peptidyl transfer are the funda-
mental processes in translation, several other
regions of the ribosome also have important func-
tional roles in protein synthesis. Within the large
0022-2836/02/$ - see front matter q 2002 Elsevier Science Ltd. All rights reserved
E-mail address of the corresponding author:
harvey@neptune.cmc.uab.edu
Abbreviations used: 3P, three phylogenetic domains;
3P2O, three phylogenetic domains plus two organelles;
MRP, mitochondrial ribosomal protein; mt-rRNA,
mitochondrial ribosomal RNA; SRL, a-sarcin–ricin loop;
CRW, comparative RNA web; cryo-EM, cryo-electron
microscopy.
doi:10.1016/S0022-2836(02)00568-5 available online at http://www.idealibrary.com on
Bw
J. Mol. Biol. (2002) 321, 215–234
2. subunit, the L11 protein and binding site of the
toxin a-sarcin/ricin, the a-sarcin–ricin loop (SRL;
residues 2663–2777, E. coli numbering) are
required for functional interactions with trans-
lation cofactors, including elongation factors
(EF-Tu and EF-G).9 – 11
Within the 16 S rRNA the
530 stem/loop and the 1490 region are involved in
proofreading,12
helix 27 (residues 885–912) has
been indicated as a conformational switch,13
and
several other areas are functionally significant for
translation. The two ribosomal subunits must com-
municate during proofreading, peptidyl transfer
and translocation, and bridge contacts between
the two subunits are important in this communi-
cation, as are specific interactions between the
tRNAs and the ribosome.
Structural studies have begun to provide
insights into the details of various steps in the
translation cycle. Crystal structures of individual
ribosomal subunits (50 S from an archeon,
Haloarcula marismortui;14
30 S from a thermophilic
bacterium, Thermus thermophilus15,16
) have been
solved at 2.4 to 3.3 A˚ resolution. The crystal
structure of the whole 70 S ribosome from
T. thermophilus was solved at 5.5 A˚ resolution.17
Results from crystallography have been sup-
plemented by cryo-electron microscopy (cryo-EM)
images of the ribosome at various stages of the
translation cycle.10,18– 20
These advances allow
exploration of structural relationships that are
important for function. Here we examine which
regions of the ribosome have been most highly
conserved throughout evolution and define a
“minimal ribosome” where structure–function
relationships can be elicited.
We have used comparative sequence analyses21,22
to examine conservation of the various regions of
ribosomal RNA through the Archaea, Bacteria and
Eucarya (herein named 3P, indicating three phylo-
genetic domains) phylogenetic domains. An even
more inclusive model was obtained by examining
conservation through these three phylogenetic
domains plus mitochondria and chloroplasts
(herein named 3P2O, for three phylogenetic
domains plus two organelles). To investigate the
most minimal, truncated ribosome, we have
examined the rRNA sequence of C. elegans
mitochondria.
The work presented here suggests the minimal
structures required for translation, and it also
addresses how evolution affects interactions within
the ribosome, and with translation cofactors. Pre-
viously it was predicted that viewing ribosomal
structures from a phylogenetic perspective would
address how rRNA structure evolves.23
We find
that several functional domains are maintained as
an RNA core in the reduced ribosomal models,
but that sequence deletions and insertions require
that the positions of some regions must have
moved in three-dimensional space during
evolution. The largest differences between widely
diverged species are at the periphery, away from
the central core.
Results
Conservation through three phylogenetic
domains (3P)
The first step in building a minimal ribosome is
to compare conservation through the three phylo-
genetic domains: Archaea, Bacteria and Eucarya.
There are five levels of conservation as defined at
the Comparative RNA Web (CRW) site†.24
Positions that are present in more than 95% of the
sequences are divided into four categories where
specific nucleotides are conserved at that position
in (1) 99–100%, (2) 90–98%, (3) 80–89%, and (4)
below 80% of the sequences. The fifth category is
for those positions that are not present in more
than 95% of the sequences. The results of these
comparative studies can be viewed on the second-
ary structure diagrams for both the 16 S rRNA
and 23 S rRNA at the CRW site. The results were
used to generate the ribosome models shown in
Figure 1(a) and (b). Nucleotides that are 99–100%
conserved are called “universally conserved resi-
dues”, and these are represented as red spheres in
Figure 1. While 23% of the 16 S rRNA positions
are universally conserved (Table 1), the region of
greatest conservation in the small subunit sur-
rounds the area where the mRNA and tRNAs
interact and move through the ribosome (Figure
1(a)). Other universally conserved areas are mostly
at the interface where the small subunit interacts
with the large subunit, but the regions of domain I
involved in the formation of intersubunit bridges
have the lowest percentage of universally con-
served residues. The large subunit has a cluster of
universally conserved residues around the cleft
where the acceptor stems of A and P-site tRNAs
bind and peptidyl transfer occurs (Figure 1(b)).
These regions are constructed from domains IV
and V of the 23 S rRNA, and they exhibit the
highest percentage of universally conserved resi-
dues (29 and 28%, respectively). Conservation also
extends from the peptidyl transfer site to many
residues associated with the polypeptide-conduct-
ing tunnel.3,25,26
The SRL is universally conserved,
and the regions of rRNA that bind L1, L11 and
other proteins contain many residues that are
universally conserved. As is the case for the small
subunit, many of the residues at the interface of
the large subunit are universally conserved.
Nucleotides that are 90–98% conserved are
referred to as “highly conserved residues”, and
they are represented by green spheres in Figure 1.
In the small subunit, the highly conserved areas
are generally located in the same regions as the
universally conserved residues. They fill many of
the gaps left in the sequence between universally
conserved nucleotides in the mRNA cleft and at
the interface side of the subunit. The platform of
the small subunit is highly conserved, with a few
† http://www.rna.icmb.utexas.edu
216 Modeling a Minimal Ribosome
3. universally conserved residues interspersed. This
region is involved in limited, specific interactions
with the E-site tRNA and mRNA, and the very
high level of conservation of these residues
suggests that they play additional important struc-
tural and functional roles. Highly conserved resi-
dues in the large subunit are concentrated near
the universally conserved residues, but the
percentages for these residues are more equally
distributed throughout the six structural domains
(Table 1).
The third level of conservation for the three
phylogenetic domains is the category we call “con-
served positions”. These nucleotides are less than
90% conserved at the nucleotide level but are
maintained as a position in at least 95% of the
sequences in the three phylogenetic domain data-
set (combining categories 3 and 4 from the CRW
site). This category represents the preservation of
specific structural elements (e.g. base-pairs; stems;
loops) at particular positions in the secondary
structure, even though specific nucleotides are
not required. These structural elements presum-
ably act as scaffolds to give the more highly con-
served residues critical orientations in three-
dimensional space. These conserved positions are
represented as dark gray helices in Figure 1. These
residues are most predominant in domains II and
VI of the large subunit (60% of the sequence in
each).
“Non-conserved regions” are those positions
that are not present in 95% of the sequences in the
3P dataset; these are represented as transparent,
gray ribbons in Figure 1. In many cases, these
areas occur in the rRNA from some species but
are completely absent from ribosomes in other
Figure 1. Conserved regions of the ribosome through three phylogenetic domains ((a) and (b)) and through three
phylogenetic domains plus mitochondria and chloroplasts ((c) and (d)), mapped onto the small subunit ((a) and (c))
and on the large subunit ((b) and (d)). Positions present in more than 95% of the sequences are shown in red, green
and gray. Red spheres represent those nucleotides that are universally conserved (99% conserved or greater), green
spheres designate highly conserved residues (90–98% conserved). Dark gray helices represent conserved positions
(less than 90% conserved at the nucleotide level). The light, transparent gray ribbons represent regions that are not
present in at least 95% of the sequences. Conserved proteins are shown in red and non-conserved are gray
(summarized in Table 1). Main structural features for both subunits are indicated with the following
abbreviations: bk, beak; sh, shoulder; sp, spur; P, platform; L1, L1 arm; CP, central protuberance; L11, L11 arm; SRL,
a-sarcin–ricin loop. The four models can be viewed from a 3608 perspective in movies at the Harvey laboratory
website (http://uracil.cmc.uab.edu/Publications/).
Modeling a Minimal Ribosome 217
4. species. Also, the majority of the insertions to the
E. coli, T. thermophilus and H. marismortui rRNA
sequences occur in these regions.27
Non-conserved
regions in the small subunit include helices 16 and
17 (the shoulder, residues 404–499; Figure 1(a)),
helices 33 and 33a (the beak, residues 1000–1040),
and helix 6 (the spur, residues 79–100). The highest
percentages for non-conserved residues in the 23 S
rRNA are found for domains I and III in the
secondary structure (Table 1), which are apparently
not absolutely required. In general, non-conserved
regions are located at the periphery of the small
and large subunits. The 5 S rRNA of the large
subunit is not included in this model. While the
general structure of the 5 S rRNA is conserved
over three phylogenetic domains, the sequence
conservation is low.
The number of sequences in the ribosomal
protein database is much smaller than for rRNA,
so conclusions about their conservation are tenta-
tive. Here we do not concern ourselves with the
specific conservation of amino acid residues or the
presence or absence of regions of the ribosomal
proteins. Rather, we only consider whether the
homologous protein is present in the organisms in
the 3P dataset (see Materials and Methods). Within
these limits, protein conservation is shown in Table
2. Conserved proteins are shown in red, and non-
conserved proteins are shown with transparent
gray ribbons in Figure 1. As with rRNA, conserved
proteins are generally located near the core of both
subunits, and proteins that are not conserved tend
to be located on the periphery of the two subunits.
This finding indicates that several rRNA elements
that interact with proteins are conserved in both
subunits. A strong correlation exists between pro-
tein conservation and RNA conservation within
the structure.
In summary, most of the 30 S and 50 S subunits
are conserved across the three phylogenetic
domains. Furthermore, the degree of conservation
is highest near the functional centers of the two
subunits, as would be expected, and the com-
ponents of the bridges between the two subunits
tend to be conserved, particularly in the neighbor-
hood of the active sites.
Table 1. Comparative analysis
Category 3P conservation analysis (small subunit)
All (1542)a
I (568) II (352) III (476) IV (146)
1 23 (351)b
19 (106) 24 (84) 26 (126) 24 (35)
2 10 (156) 9 (49) 11 (38) 11 (53) 11 (16)
3 5 (80) 5 (27) 5 (17) 6 (29) 5 (7)
4 35 (545) 37 (209) 39 (139) 28 (131) 45 (66)
5 27 (410) 31 (177) 21 (74) 29 (137) 15 (22)
Large subunit
All (2904) I (561) II (733) III (352) IV (396) V (587) VI (275)
1 20 (595) 17 (95) 20 (147) 10 (35) 29 (113) 28 (167) 14 (38)
2 12 (342) 12 (69) 13 (94) 12 (42) 9 (36) 12 (72) 11 (29)
3 10 (282) 11 (63) 10 (73) 4 (15) 9 (36) 9 (54) 15 (41)
4 41 (1201) 38 (214) 50 (363) 33 (116) 39 (155) 39 (229) 45 (124)
5 17 (484) 21 (120) 8 (56) 41 (144) 14 (56) 11 (65) 16 (43)
3P2O conservation analysis (small subunit)
All (1542)a
I (568) II (352) III (476) IV (146)
1 12 (178) 9 (51) 12 (42) 13 (60) 17 (25)
2 11 (175) 8 (44) 14 (51) 15 (70) 7 (10)
3 7 (115) 5 (29) 9 (33) 9 (44) 6 (9)
4 26 (395) 24 (137) 31 (109) 26 (122) 18 (27)
5 44 (679) 54 (307) 33 (117) 38 (180) 51 (75)
Large subunit
All (2904) I (561) II (733) III (352) IV (396) V (587) VI (275)
1 5 (151) 1 (5) 3 (25) 0 (0) 12 (48) 11 (64) 3 (9)
2 7 (202) 6 (32) 8 (48) 0 (0) 11 (45) 9 (50) 6 (17)
3 6 (168) 2 (9) 6 (47) 0 (0) 8 (33) 9 (52) 10 (27)
4 22 (646) 11 (61) 31 (224) 0 (0) 29 (113) 28 (163) 31 (85)
5 60 (1737) 81 (454) 52 (379) 100 (352) 40 (157) 44 (258) 50 (137)
There are five categories listed; the first four include positions present more than 95% of the time: (1) 99–100% nucleotide conserva-
tion, (2) 90–98% nucleotide conservation, (3) 80–89% nucleotide conservation, (4) less than 80% nucleotide conservation, (5) positions
present less than 95%.
a
Comparative results are shown by domain (All ¼ all domains, I ¼ domain I, etc.) with the total number of residues in that domain
listed in parentheses.
b
Percentages are listed (number/total) with the observed number of residues shown in parentheses. Additional comparative infor-
mation is available at the CRW site (http://www.rna.icmb.utexas.edu/ANALYSIS/MINRIB/).
218 Modeling a Minimal Ribosome
5. Conservation through three phylogenetic
domains plus two organelles (3P2O)
Due to the additional sequence and structure
variation in the mitochondria and to a lesser extent
in chloroplast ribosomes, there is less conservation
when these sequences are added to the 3P dataset.
Figure 1(c) and (d) shows conservation across the
three phylogenetic domains plus the two
organelles, using the same coloring scheme as in
Figure 1(a) and (b) (3P2O secondary structures are
available for both the 16 S and 23 S rRNA at the
CRW site†).
The small subunit model for 3P2O conservation
(Figure 1(c)) is similar to the model based on 3P
conservation (Figure 1(a)), although overall con-
servation is significantly lower. Many of the 3P
universally conserved residues are only highly
conserved in the 3P2O model. Roughly half of the
universally conserved residues fall to a lower
degree of conservation for each domain (Table 1).
The head and platform portions of the subunit in
the 3P2O model are very highly conserved, and
most of the universally conserved residues are
found at or near the mRNA binding site. Fewer
universally and highly conserved residues are
found in the lower body regions, except for some
at the interface. Although overall conservation in
the body is not as high as in the 3P model, 29% of
the residues in domain I are conserved as posi-
tions, indicating that the structural elements are
still required at these positions.
The number of universally conserved residues is
dramatically reduced in the 3P2O model for the
large subunit (constituting only 5% of the
sequence, Table 1), and almost all of the peripheral
rRNA is non-conserved (60% of the positions are
not maintained in 95% of the sequences in the
Table 2. Protein conservation
SSU protein 3P 3P2O In crystal?a
LSU protein 3P 3P2O In crystal?a
S1 2 2 Nob
L1P þ þ Noc
S2 þ þ Yes L2P þ þ Yes
S3 þ þ Yes L3P þ 2 Yes
S4 þ þ Yes L4P þ 2 Yes
S5 þ þ Yes L5P þ þ Yes
S6 2 2 Yes L6P þ 2 Yes
S7 þ þ Yes L7AE þ 2 Yes
S8 þ þ Yes LPO (L10P) þ 2 No
S9 þ þ Yes L10E (L10AE) þ þ Yes
S10 þ þ Yes L11P þ þ Noc
S11 þ þ Yes L12A (L7/L12) þ 2 No
S12 þ þ Yes L13Pd
þ 2 Yes
S13 þ þ Yes L14Ed
2 2 No
S14 þ 2 Yes L14P þ 2 Yes
S15 þ 2 Yes L15E 2 2 Yes
S16 2 2 Yes L15P þ 2 Yes
S17 þ þ Yes L18E 2 2 Yes
S18 2 2 Yes L18P þ þ Yes
S19 þ þ Yes L19E 2 2 Yes
S20P 2 2 Yes L21E 2 2 Yes
S21 2 2 Noe
L22P þ þ Yes
S22 2 2 No L23P þ þ Yes
LTRA 2 2 No L24E 2 2 Yes
THX 2 2 Yes L24P þ þ Yes
L29P þ 2 Yes
L30E 2 2 No
L30P þ 2 Yes
L31E 2 2 Yes
L32E 2 2 Yes
L34E 2 2 No
L35AE 2 2 No
L37E 2 2 Yes
L37AE 2 2 Yes
L39E 2 2 Yes
L40E 2 2 Yes
L41E 2 2 No
L44E 2 2 Yes
LX 2 2 No
Conservation based on website http://geta.life.uiuc.edu/~nikos/Ribosome/rproteins.html
a
The crystal structures are a bacteria (T. thermophilus, 1FJG) for the SSU and an archea (H. marismortui, 1FFK) for the LSU.
b
S1 was removed from the 30 S subunit before crystallization.
c
Proteins L1P and L11P were modeled onto the structures for spatial representation.
d
P and E represent prokaryotic and eucaryotic ribosomal proteins, respectively.
e
S21 is not found in T. thermophilus.
† http://www.rna.icmb.utexas.edu
Modeling a Minimal Ribosome 219
6. 3P2O dataset). Much of the reduction is attribu-
table to animal mitochondrial ribosomes, and not
to plastid (chloroplast) ribosomes because their
conservation is similar to the bacterial 70 S
ribosomes.28,29
A universally conserved core still
remains at the cleft near the peptidyl transferase
site (domains IV and V), but most of the polypep-
tide-conducting tunnel is absent (Figure 2). Highly
conserved residues or conserved positions make
up most of the remaining core. The SRL is still uni-
versally conserved, indicating its essential role.
Some universally conserved residues are found at
the interface, but many of the subunit–subunit
bridge contacts are apparently lost in organellar
ribosomes, as is discussed in more detail below.
The 5 S rRNA is not included in the 3P2O model,
since the majority of the mitochondria do not have
a 5 S rRNA. All organisms that have the 5 S rRNA
also contain the 2282–2394 region of the 23 S
rRNA that is protected by 5 S rRNA.
While the majority of the deletions relative to the
E. coli, T. thermophilus and H. marismortui secondary
structures are on the periphery of the rRNA three-
dimensional structure, a few are internal with
islands of conservation at the outer edges of the
3P2O structural model. One example is the L11
protein binding domain (residues 1055–1105).
This conserved region is located at the periphery
of the subunit and interacts with translation
cofactors, like EF-G.30
The nearest connecting helix
to the conserved L11 binding domain is roughly
80 A˚ away in the 3P2O model. This raises the
possibility that the position of the rRNA has
moved during evolution. The distance from the
L11 binding domain to the central core of the par-
ticle has to be smaller, as the adjoining helix (helix
41 and/or 42) is absent from the 3P2O model. The
SRL is also on the periphery of the subunit, and
its position must have also evolved to accommo-
date changes in the size of the ribosome. Both the
SRL and the L11 binding domain are important in
binding elongation factors, so their positions must
have evolved together. Another interesting feature
of the 3P2O model is that protein L1 is conserved,
but the rRNA that it contacts in cytoplasmic ribo-
somes (residues 2109–2180) is not conserved in
organellar ribosomes. Recent studies raise the
possibility of structural compensation by ribo-
somal proteins for loss of parts of the rRNA within
bovine mitochondrial ribosomes.31 –34
However, it is
not clear which protein would be involved in each
compensation.
The most dramatic loss of RNA using 3P2O con-
servation analysis in the 23 S secondary structure
Figure 2. The polypeptide-con-
ducting tunnel is illustrated for
both (a) 3P and (b) 3P2O conserva-
tion by taking a cross-section of the
50 S subunit. The coloring scheme
is the same as in Figure 1, with the
exception that regions that are not
present in at least 95% of the
sequences are not transparent, but
they are still colored light gray. The
P-site tRNA and an elongated poly-
peptide in the tunnel are colored
blue. The tunnel is conserved
through three phylogenetic
domains, but it is severely trun-
cated when mitochondria and
chloroplast sequences are included.
220 Modeling a Minimal Ribosome
7. occurs in domain III (100% non-conserved; Table 1).
It may be that this region is not essential for
translation, but that it has been developed to
make the process more efficient or to assist with
complementary functions (i.e. protein secretion).
Also, most of domain I is absent in 3P2O conserva-
tion (81% non-conserved), and the few RNA
loops that are conserved are close to the conserved
core.
Few proteins are shown in Figure 1(c) and (d),
but it should be noted that mitochondrial ribo-
somes contain more proteins than their bacterial
counterparts. Unfortunately, the limited knowl-
edge about mitochondrial ribosomal proteins
(MRPs) makes it difficult to say exactly which
proteins are conserved and how they are arranged
in three-dimensional space (see Materials and
Methods).
A model for the C. elegans
mitochondrial ribosome
While these comparative phylogenetic studies
have suggested a model of the minimal ribosome,
we also developed a model of an extremely
reduced ribosome found in nature for comparison.
The rRNA from C. elegans mitochondria was
modeled, since this ribosome contains one of the
most reduced small and large subunit sequences
known.35
Figure 3 shows the alignment of these
sequences with the sequences of the species used
in the crystal structures (T. thermophilus for the
small subunit and H. marismortui for the large sub-
unit). This gives the information needed to gener-
ate the corresponding three-dimensional models
(Figure 4). The MRPs found in C. elegans are not
well defined, so we have not taken them into con-
sideration. Nucleotides that align between the
C. elegans mitochondrion and the species from the
crystal structures are shown in red in Figures 3
and 4. Yellow indicates regions where the C. elegans
sequence does not align with that from the species
in the crystal structures. The secondary structure
of the yellow regions has not been determined, in
part because the C. elegans sequence is very A-U
rich. They are mapped to the corresponding region
of the three-dimensional structure even though the
actual structure of the C. elegans rRNA may differ
substantially from that in the crystal structures.
However, since these unassigned sequences flank
conserved regions, a rough estimate of where the
RNA sequence would be located in the model can
be made. It should be pointed out here again that
the 5 S rRNA has not been found in animal mito-
chondrial genomes, including C. elegans.
The models of the C. elegans mitochondrial ribo-
somal RNAs (mt-rRNAs) are very similar to those
obtained using 3P2O phylogenetic conservation.
The small subunit mt-rRNA maps to the region
around the head and mRNA binding site of the
bacterial small subunit. The platform is also con-
served in this model, but most of the rRNA in the
body is lost. Missing regions in the small subunit
include the beak, helices 16 and 17, and most of
the periphery. Certain regions at the interface are
conserved, including the uppermost region of
helix 44 (residues 1401–1415; 1485–1501). The
large subunit also looks very similar to the 3P2O
model. A core region around the peptidyl transfer-
ase site and the interface is conserved. The con-
served regions at the interface of the large subunit
are complementary to those that they interact with
in the small subunit. In particular, helices 69 and
71 (residues 1906–1961 in E. coli) are conserved.
These helices are known to interact with helix 44
of the small mt-rRNA. It should be noted that the
unaligned regions of the C. elegans sequence
(yellow) are located near the periphery of the
conserved core, and both the L11 protein-binding
and SRL domains are present.
As with the 3P2O model, the principal deletions
in sequence occur mainly in the periphery of the
small and large subunits. However, a number of
exceptions are listed in Table 3, along with an
analysis of the effects of losing these sequences on
the three-dimensional structure of C. elegans mt-
rRNA. The distances between consecutive nucleo-
tides on either side of the deletion are reported,
and any unaligned sequence that fits in the gap of
the deletion is listed in the next column. On the
basis of the distance to be spanned and the length
of the unaligned sequence, we have estimated the
probability, P(move), that the regions on either
side of the gap must be moved closer together in
order to permit closure of the RNA chain. Briefly,
for a given gap length of N unaligned nucleotides,
we built a histogram of end-to-end distances for
all N-nucleotide fragments within the crystallo-
graphic database. If the distance between the last
phosphate before the gap and the first phosphate
after the gap in the model is r, then the probability,
P(move), that such a gap cannot be closed by a
fragment with N nucleotides using a structural
motif resembling one found in the crystallographic
database is simply that fraction of the histogram
distances that is less than r (see Materials and
Methods for details). In this way it is possible to
see which conserved regions in the C. elegans sub-
units probably have different positions in three-
dimensional space than the corresponding regions
do in the available crystal structures.
Several loops in the small subunit align to
regions far from their respective helices. An
example of this is the loop at the end of the penul-
timate stem, or helix 44 (residues 1450–1452)
where three nucleotides align. They are roughly
80 A˚ away from the nearest connecting helix, so
the position of this loop in three-dimensional
space must be different from that in bacteria. Of
the deletions that are found in the small subunit
(Table 3), all but two result in some conserved
feature being in a position where it cannot reach
the nearest connecting helix. This suggests that
these regions must have moved during evolution.
One exception is the deletion of helices 16 and 17
where three unaligned nucleotides are available to
Modeling a Minimal Ribosome 221
9. Figure 3. The alignment of C. elegans mitochondrial (a) small and (b) large rRNA to the secondary structures of
T. thermophilus and H. marismortui, respectively. Red residues represent nucleotides that align, while yellow residues
represent nucleotides that are found in C. elegans but could not be aligned with the correlating secondary structure.
Since these unassigned sequences flank conserved regions, a rough estimate of where the RNA sequence would be
located in the three-dimensional model (Figure 4) can be made, but the structure may be different from the secondary
structure, since these sequences do not align. A few functional regions are labeled and helix numbers are shown in
green for those areas that are mentioned in the text.
Modeling a Minimal Ribosome 223
10. Figure 4. Mapping the alignment of C. elegans mitochondrial rRNA to the crystal structures of T. thermophilus (1FFK)
and H. marismortui (1FJG). The red helices represent those nucleotides that do align. The yellow helices are used for
spatial representation of C. elegans sequence that does not align, but is inferred to be in a particular region of the
three-dimensional structure. Mitochondrial proteins are not included. Both models can be viewed from a 3608 perspec-
tive in movies at the Harvey laboratory website (http://uracil.cmc.uab.edu/publications).
Table 3. Deletions in the C. elegans alignment
Small subunit
Helices losta,b
Deletionc
Gap distance (A˚ ) Sequence length (nt)d
P(Move)
e
SH5–15 51–393 Loopf
44 –
SH16–17 408–499 17 3 0.3
SH21–22 (690 region) 588–672, 736–760 65, 43 None 1.0
SH25–26a (3 nt loop) 814–863, 867–879 35, 36 None 1.0
SH33 and SH35 996–1045 16 NA 1.0
SH37 (6 nt loop) 1084–1089, 1096–1102 25, 17 None 1.0
SH38–40 1113–1191 23 11 0.2
SH41 (3 nt loop) 1243–1265, 1269–1294 35, 29 None 1.0
LH44 (3 nt loop) 1416–1449, 1453–1484 83, 97 None 1.0
Anti-Shine-Delgarno 1536–1542 ? None –
Large subunit
LH1–25 1–562 NA 22 (50
end) –
LH28–31 595–662 Loopf
4 –
LH34 (3 nt loop) 698–713, 717–763 26, 26 None 1.0
LH37–42 and LH45 (GTPase region) 818–1050, 1109–1185 91, 100 35, 13 1.0
LH46 (4 nt loop) 1206–1222, 1227–1240 36, 19 None 1.0
LH47–63 1276–1765, 1987–2010 65, 46 44, 19 0.7
LH66 (4 nt loop) 1798–1806, 1811–1821 26, 30 None 1.0
LH68 (3 nt loop) 1844–1868, 1872–1896 51, 60 None 1.0
LH76–79 2092–2227 Loopf
4 –
LH81–88 2259–2421 26 24 0.2
LH94 (connects to SRL) 2627–2645 43 8 1.0
LH96 2679–2728 Loopf
5 –
LH97 2745–2759 Loopf
None –
LH98–101 2770–2904 NA None (30
end) –
Deletions of 5 nt or less are not included.
a
SH and LH are used to represent small subunit helices and large subunit helices, respectively.
b
The identity shown in red is the region that would need to move to compensate for the deletion.
c
The residues that are deleted from the C. elegans sequence are listed with E. coli numbering.
d
Sequence length represents the unassigned sequence from the C. elegans structure that is not represented by a homologous region
in the E. coli secondary structure and is therefore available to close the gap.
e
Probability that movement would be required to compensate for the deletion (see the text for explanation).
f
Gap distance is not important because the deletion is at the end of a helix and can easily be closed by a new loop.
224 Modeling a Minimal Ribosome
11. cover a gap distance of 17 A˚ . The probability of
movement is only 30%, indicating that this
distance can be closed by the unaligned nucleo-
tides. The 690 region, which has been implicated
in interactions with the P and E-site tRNAs in the
small subunit,36,37
must have moved during mito-
chondrial evolution because there is a gap of 65 A˚
on the Watson strand and 43 A˚ on the Crick strand
of the deleted sequence. No unaligned nucleotides
are present to span this distance, so the probability
of movement is 100%.
The large subunit also has several deletions,
resulting in large gaps. Some examples of deletions
do not require movement because of unaligned
sequence. The region in domain V that interacts
with the 5 S rRNA (helices 81–88, residues 2259–
2421) is absent, as was found in 3P2O conservation.
In C. elegans there are 24 unaligned nucleotides
in this region. Such a large fragment could easily
close the remaining 26 A˚ gap, and the probability
that movement was required during evolution is
calculated at only 19.5% (Figure 5).
The large subunit also has regions that must
have moved during evolution to close gaps left by
deletions. These movements also correlate to the
results found in the model of 3P2O phylogenetic
conservation. The L11 binding domain (helices 43
and 44) is found in C. elegans, but the connecting
helices (41 and 42) are shortened. A total of 48
nucleotides of unaligned sequence are in this
region, but the distance of 91 A˚ would be difficult to
span with 35 nucleotides on the Watson strand
of the helix, and 13 nucleotides cannot span 100 A˚
on the Crick strand of the helix. We therefore con-
clude that the position of the L11 binding domain
must have moved during evolution. Similarly, the
SRL (at the end of helix 95) must have moved,
because the deletion of helix 94 and the base of
helix 95 creates a gap of 43 A˚ , and the eight
unaligned nucleotides in this region cannot span
that distance (Figure 5). Since both the SRL and
the L11 binding domain are known to interact
with translation cofactors on the surface of the
ribosome, comparison of C. elegans with
H. marismortui reveals substantial differences in
those interactions. At present, it is unclear whether
the interactions between cofactors and the large
subunit are simply moved to a different location
in the C. elegans large subunit (but are otherwise
similar to those in H. marismortui), or whether
other parts of the subunit (such as proteins not
found in H. marismortui) take on roles otherwise
associated with the SRL, L11 and the L11 binding
domain. The latter appears unlikely because these
specific regions are so highly conserved.
Conservation of intersubunit bridges
The two ribosomal subunits must communicate
with one another to synchronize various events
during translation. This communication takes
place at the interface of the subunits through the
movements of bridges. These bridges have been
shown to be dynamic, with different connections
made and broken at different stages of the trans-
lation cycle.8,38
Table 4 examines the conservation
of intersubunit bridges, including the type of inter-
action (RNA/RNA, RNA/protein, protein/pro-
tein), whether or not the bridge is conserved
based on 3P and 3P2O conservation, and if the
rRNA sequence involved can be inferred to exist
in C. elegans. At the 3P conservation level, all 11
listed bridges are conserved. This points to the
importance of these bridges for both the structural
integrity of the ribosome and the probable func-
tional significance of these interactions in
translation.
However, the picture is notably different when
mitochondria and chloroplasts are included,
because only four of the 11 bridges are conserved
across 3P2O (Table 4). Bridges B2a, B2c, B3 and
B7b are conserved, indicating that these contacts
are probably essential. B2a and B3 involve contacts
near the decoding site at the top of the penultimate
stem in the small subunit. From cryo-EM and crys-
tallography studies, it has been shown that this
part of helix 44 moves during translation,4,8,39
and
these bridges almost certainly play a role in those
motions.
Bridge B2c is also found near the decoding site
and is conserved through all three categories listed
in Table 4. This bridge involves helix 27 in the
small subunit, often referred to as a “conformational
switch” because it has been shown to adopt two
secondary structure conformations. Mutations that
prevent conformational changes in this region
affect fidelity.13,40
This switch lies adjacent to helix
44, which, as pointed out above, changes confor-
mation during the translation cycle. Fidelity may
involve the coordination of the movement of the
helix 27 switch with that of the decoding site in
helix 44. Bridge B2c also includes contacts with
the highly conserved structure, helix 24 in the
small subunit, the tip of which contains the 790
loop, known to be involved in interactions with
the mRNA and tRNA at the ribosomal P-site.41,42
In general, bridges that are positioned further
from the decoding site tend not to be conserved in
the 3P2O model (Figure 6). Bridges B5 and B6
involve contacts between the large subunit and
the penultimate stem of helix 44 in the small sub-
unit, but they are further from the decoding site
than B2a and B3 and are not conserved in the
3P2O model. Another example is bridge B1a,
formed by helix 38 of the large subunit and often
referred to as the “A-site finger”. This structure is
not conserved in mitochondria, which correlates
with the loss of much of the rRNA in the central
protuberance, and with the loss of 5 S rRNA in
mitochondria. Similarly, B7a is not conserved, as
this bridge is positioned away from the decoding
site and closer to the L1 arm of the large subunit,
connecting to the platform of the small subunit.
Bridges B4, B7a and B8 are located further away
from the central region of both subunits, and they
are not conserved.
Modeling a Minimal Ribosome 225
12. Figure 5. Probability of movement. (a) The absence of helix 94 (residues 2630–2643; 2761–2777) from the C. elegans
large ribosomal subunit leaves only eight unaligned nucleotides to span a distance of 43 A˚ if the neighboring helices
are to occupy the same positions as in the H. marismortui crystal structure. The histogram shows the inter-phosphate
distance distribution for all nine-nucleotide fragments in the 70 S crystal structures (1GIX and 1GIY). Since 99.3% of
such fragments are too short to span a 43 A˚ gap, we estimate that there is a 99.3% probability that helices on one or
both sides of the gap must occupy different positions in the C. elegans structure than in the H. marismortui structure.
(b) Absence of helices 81–88 (residues 2259–2421) from the C. elegans large ribosomal subunit leaves a gap of 26 A˚ to
be spanned by 24 unaligned nucleotides. In all, 80.5% of all 25-nucleotide fragments in structures 1GIX and 1GIY
span distances greater than this, so the probability that some part of the C. elegans rRNA must be moved to close the
gap is only 19.5%.
226 Modeling a Minimal Ribosome
13. Table 4. Bridges conserved through evolution
Bridgea
SSU positionsb
LSU positionsb
3
Phylo
3 Phylo þ 2
Org C. elegans Bridge type
B1a S13 (92–94) H38 (886–888) Yes No No RNA/pro-
tein
B2a H44 (1408–1410) H69 (1913–1914, 1918) Yes Yes Yes RNA/RNA
B2b H24 (784–785, 794), H45 (1516–1519) H67 (1836–1837, 1922), H71 (1919–1920, 1932) Yes No No RNA/RNA
B2c H24 (770–771), H27 (900–901) H67 (1832–1833), H67 (1832–1833) Yes Yes Yes RNA/RNA
B3 H44 (1484–1486) H71 (1947–1948, 1960–1961) Yes Yes Yes RNA/RNA
B4 H20 (763–764), S15 (40–44) H34 (717–718), H34 (713, 717) Yes No No RNA/pro-
tein
B5 H44 (1418–1419), H44 (1420–1422), H44 (1474–1476), H44 (1474–
1476)
H64 (1768–1769), L14 (44–49), H62 (1689–1690), H64
(1989)
Yes No No RNA/pro-
tein
B6 H44 (1429–1430, 1474–1476), H44 (1431) H62 (1689–1690, 1702–1705), L19 (Hm24e:R44) Yes No No RNA/pro-
tein
B7a H23 (698, 702) H68 (1848–1849, 1896) Yes No No RNA/RNA
B7b H23 (712–713), H24 (773–776) L2 (162–164, 172–174, 177–178), L2 (177–178, 198–202) Yes Yes ?c
RNA/pro-
tein
B8 H14 (345–347) L14 (116–119) Yes No No RNA/RNA
The positions are based on the crystal structure of the E. coli 70 S ribosome (Yusupov et al.17
, 1GIX and 1GIY).
a
B1b not included because the bridge type is protein/protein.
b
Small subunit (SSU) and large subunit (LSU) positions are listed by helix (H) or protein (S or L) using E. coli numbering.
c
The rRNA for B7b is conserved, but it is not known whether or not protein L2 is present. Therefore, we cannot say whether or not this contact exists in C. elegans.
14. Although it is not found in the core of the ribo-
some, bridge B7b is conserved at the 3P2O level.
B7b arises from an interaction between protein L2
and helix 23 from the small subunit. L2 is highly
conserved in all organisms and organelles, as is
helix 23. The conservation of B7b is less clear in
the C. elegans mitochondrial ribosome, because L2
has not been found in C. elegans, and half of the
nucleotides involved in helix 24 (residues 712–
713) are lost. If C. elegans contains a bridge corre-
sponding to B7b, the details of the intersubunit
interactions must be different than in other ribo-
somes. We indicate the uncertainty about the exist-
ence of bridge B7b by coloring it yellow in Figure 6.
Conservation of tRNA/rRNA contacts in
the ribosome
Many tRNA interactions with the ribosome are
made and severely broken throughout the trans-
lation cycle as the tRNA moves through the central
cavity, the intersubunit space, between the ribo-
somal subunits. The large subunit interacts with
the acceptor stem and elbow region of the tRNA,
while the small subunit interacts with the anti-
codon stem and loop. Table 5 lists the tRNA inter-
actions with the ribosome and the residue
numbers that are involved in the contacts, based
on the recent T. thermophilus 70 S crystal structure.17
A distance criterion (8.5 to 10 A˚ ) was also used to
determine residues near the different tRNA posi-
tions and mRNA. While this list is more inclusive,
the results are similar to those shown in Table 5
(additional Tables are available at the CRW site†).
On the basis of the crystal structure, the A-site
tRNA has few interactions with the small subunit,
but these are vital to the fidelity of translation.
Residues 530 and 1492–1493 are believed to read
the groove of the mini-helix created by the
codon–anticodon interaction at the A-site,4
and all
three of these residues are universally conserved
in all organisms and organelles. Residues 955 and
1054 are also conserved, but only at the level of
conserved positions in both the 3P and 3P2O
models, indicating that these residues may be
important in positioning the tRNA but are prob-
ably not involved in specific hydrogen bonding.
Several interactions occur between the A-site
tRNA and the large subunit. The greatest degree
of conservation is found in the residues that con-
tact the tRNA acceptor stem. Residues in the loop
between helices 69 and 71 and other nucleotides
throughout domain V position the acceptor stem.
Conserved residues within helix 89 and the loop
of helix 69 interact with the elbow region of the
A-site tRNA in the 3P2O model, but residues
within the A-site finger (helix 38; residues 880–
900) that would normally contact the elbow of the
tRNA are not conserved in some mitochondria.
Similarly, residues in domain I of the large subunit
that contact the elbow of the A-site tRNA in the
3P model are missing from the 3P2O model. In
short, ribosome contacts with the acceptor stem and
anticodon at the A-site are most highly conserved,
while interactions with the elbow region are con-
served in the 3P model, but not in the 3P2O model.
Figure 6. Positioning of bridge contacts onto density
maps of the large subunit (upper) and small subunit
(lower). Red bridges (B2a, B2c and B3) represent posi-
tions that are present in at least 95% of the sequences
through 3P and 3P2O conservation and are also found
in C. elegans. B7b is colored yellow because it is con-
served through 3P and 3P2O, but it cannot be deter-
mined if it is present in C. elegans (it is not known
whether protein L2 is conserved in C. elegans). The other
seven bridges are conserved through 3P, but are not con-
served when organelles are considered, and they are also
absent from C. elegans mitochondrial rRNA (Table 3).
† http://www.rna.icmb.utexas.edu/ANALYSIS/
MINRIB
228 Modeling a Minimal Ribosome
15. The P-site tRNA is at a crucial position in the
ribosome because it carries the nascent peptide
chain. Several tRNA/rRNA contacts are made
with both subunits, and all of these interactions
are conserved through both the 3P and 3P2O
level. Within the small subunit, C1400 and residues
1338–1339 are universally conserved. Universally
conserved residues in the large subunit include
residue 2252 in the P-loop and 2602 in helix 93,
which appear to be crucial for positioning the
acceptor stem in the P-site. Other large subunit
contacts with the acceptor stem and with the
elbow of the tRNA are conserved as positions, but
not as specific residues, again indicating that their
role is probably to help position the tRNAs, for
which a specific hydrogen bonding pattern is
apparently not needed. It is impressive that all of
the contacts between the P-site tRNA and the
rRNA are conserved at some level, reflecting the
importance of the positioning of the tRNA at this
position.
The E-site tRNA also has contacts with both sub-
units and, like the A and P-site tRNAs, all of its
interactions with the small subunit are conserved
at some level. Residue 1382, found in helix 28 of
the small subunit rRNA, is universally conserved.
Nucleotides in the 690 loop and 790 loop are uni-
versally conserved throughout 3P2O, but other
nucleotides that contact the E-site anticodon are
only conserved as positions. Ribosomal nucleotides
in the large subunit that interact with the E-site
tRNA are conserved at the 3P level, but only as
positions, indicating a role in positioning the
tRNAs. Interestingly, most of the E-site contacts
within the large subunit are lost when mito-
chondrial sequences are considered, raising the
question of whether or not a true E-site exists in
all mitochondria. Only residues 2235 and 2433–
2434, located in helix 75, are conserved throughout
3P2O, including C. elegans. Residue 199 is found in
3P2O conservation but is missing from C. elegans.
All of these residues contact the acceptor stem of
the E-site tRNA.
Discussion and Conclusions
We have used comparative sequence analysis to
examine conserved features of ribosome structure.
Most of the rRNA and proteins in the 70 S particle
are conserved to some degree across the 3P
domains. The regions that are not conserved at the
3P level tend to be located on the periphery of the
structure. These reflect the fact that Archea,
Bacteria and Eucarya ribosomes vary in size, and
the likelihood that organisms living in widely
different environments have developed different
ribosomal modifications to guarantee translational
accuracy and efficiency. The universally conserved
residues are generally found in areas that are
known to be functionally significant, particularly
near the mRNA binding site on the small subunit,
the peptidyl transferase site on the large
Table 5. Conserved tRNA/rRNA contacts within the
ribosome. (More quantitative details available at the
CRW site (http://www.rna.icmb.utexas.edu/ANAL-
YSIS/MINRIB/).
A-Site
Residue numbers tRNA residues 3P 3P2O C. elegans
Small subunit
5301 34–36 Yes Yes Yes
955 40 Yesa
Yesa
Yes
1054 34 Yes Yesa
Yes
1493 38 Yes Yes Yes
Large subunit
27,30 55,62 Yesb
No No
881–2 17 Yesb
No No
882–3 19 Yesb
No No
898–9 56 No No No
1913–4 25,26 Yes Yesa
Yesc
1914–5 11,12 Yes Yesa
Yesc
1942–3 72–73 Yes Yesa,c
Yes
2452 and 2494 74–76 Yes Yes Yes
2470–2 50–53 Yesb
Yesb
Yes
2482–4 64–65 Yesb
Yesb
Yesc
2553 75 Yes Yes Yes
P-Site
Small subunit
790 38 Yes Yes Yes
966 34 Yesa
Yesa
Yes
1229 28–30 Yesa
Yesa
Yes
1338 41 Yes Yes Yes
1339 40 Yes Yes Yes
1400 34 Yes Yes Yes
Large subunit
1908–9 12, 13 Yesa
Yesa
Yes
1922–3 25, 26 Yes Yesa
Yes
2252 74 Yes Yes Yes
2255–6 3 Yes Yesb
Yes
2585 76 Yes Yesa
Yes
2602 75 Yes Yes Yes
E-Site
Small subunit
693–695 37–9 Yes Yesb
Yes
788–789 37–8 Yesb
Yesa
Yes
937 33 Yesa
Yesa
Yes
1339–40 35–6 Yesb
Yesb
Yes
1382 34 Yes Yes Yes
Large subunit
199 76 Yes Yes No
1850–3 3, 4, 5 Yesb
No No
1852–3 2,71 Yesb
No No
1892 71 Yes No No
2112–3 19 Yesa
No No
2116–7 56 Yesa
No No
2235 73 Yesa
Yesa
Yes
2433–4 76 Yes Yesb
Yes
The residues listed are based on the 70 S crystal structure17
and are listed with E. coli numbering.
a
Residue(s) position(s) conserved in at least 95% of the
sequences, but conservation is not greater than 90% at the
nucleotide level.
b
Residue(s) universally or highly conserved, but one of the
residues is conserved as a position (less than 90% at the nucleo-
tide. level, but position is present in at least 95% of the
sequences).
c
One of the residues in this region is not conserved.
Modeling a Minimal Ribosome 229
16. subunit, and the regions on the large subunit that
interact with elongation factors. A strong corre-
lation also exists between protein conservation
and conservation of regions of the rRNA that inter-
act with those proteins. While this may seem a
natural consequence of the need to preserve
protein-binding sites, the situation becomes quite
different when mitochondria and chloroplasts are
taken into consideration.
As seen in Figure 1(c) and (d), when organelle
sequences are considered, both subunits show
severe truncations. Even though plastid ribosomes
are very eubacterial-like, the minimal models
based on 3P2O conservation are quite simple,
which reflects the diversity of mitochondrial
sequences, and the simplicity of many of them.
The simplification of mitochondrial rRNA
sequences correlates well with the known simpli-
fication of mitochondrial genomes. Mitochondria
import many essential proteins and RNAs.
Although it is not yet fully understood how
evolution has chosen which protein and RNA
genes to move from the mitochondrion to the
nucleus, natural selection must balance efficiency
against the need to regulate levels of gene
expression. The diversity and simplicity of mito-
chondrial genomes must reflect these competing
demands. One interesting facet of this is the
observation that some mitochondrial ribosomal
proteins have C and N-terminal extensions,
apparently to accommodate for the loss of rRNA.34
This explains the increased ratio of protein to
RNA within mitochondrial ribosomes. As more
data on MRPs become available, it will be interest-
ing to understand how proteins replace rRNA
both structurally and functionally.
Those proteins that maintain conserved inter-
actions with rRNA are generally known to serve
critical roles in translation. Roughly half of the
small subunit proteins (13/24) are conserved
across all three phylogenetic domains and both
organelles (Table 2), indicating their importance in
the structure and function of the small subunit.
For example, proteins S5 and S12 are conserved,
as is helix S27 with which they interact. Mutational
studies have shown the importance of these
proteins in translational fidelity.13
Although all of the bridges are conserved across
the three phylogenetic domains, only B2a, B2c, B3
and B7b are conserved across 3P2O (Table 4),
suggesting that these are ancient and functionally
critical contacts. The functional significance of B2a
and B3 has been suggested by several studies, but,
as far as we are aware, the importance of B2c and
B7b has not been previously suggested. Proteins
may replace RNA in some of the bridges that are
not conserved across 3P2O, but the reduced size
of both subunits in the 3P2O consensus models
indicates that other bridges, particularly those
around the periphery, do not exist in all ribosomes.
The disconnected appearance of the 3P2O
models in Figure 1 arises from the lack of conserva-
tion of critical RNA connections, suggesting that
some functional regions have moved during
evolution. For example, L11 and the rRNA that
binds it are both conserved across 3P2O, while the
adjoining helices (helices 37–42 and helix 45) are
not. Similarly, helix 94, which connects the SRL to
the rest of the ribosomal RNA, is missing in the
3P2O consensus model. The remaining rRNA frag-
ments that must connect the SRL and the L11
region to the rest of the large subunit are too short
to do so without some change in the positions of
the SRL and protein L11 (Table 3). Therefore, the
SRL and L11 must both be in substantially different
positions in mitochondrial ribosomes than in other
ribosomes. Similarly, the 690 region of the small
subunit rRNA is separated from its neighboring
sequence in the 3P2O model, because of the loss
of the S15 protein-binding region. These
observations strongly suggest that, as pieces of
rRNA are added or deleted during evolution, the
relative three-dimensional positions of some highly
conserved RNA domains must change. These
evolutionary shifts may well alter both intra-
subunit and intersubunit interactions. The func-
tional consequences of such changes are unclear.
The polypeptide-conducting tunnel is conserved
in the 3P model, but the truncation of rRNA in the
3P2O model results in a very short tunnel (Figure
2). It is likely that as the need for protein secretion
arose, the protein tunnel evolved, which is why
this region is so highly conserved in the 3P model.
The truncation of the tunnel in many mito-
chondrial sequences is surprising, since many
proteins translated in the mitochondria are
membrane-associated. The strategy for protein
translocation is conceivably very different in
organelles, so that mitochondrial ribosomes have
developed a new mechanism for protein secretion.
The non-homologous proteins that are found in
the mitochondrial ribosome may serve such a role.
The traditional small subunit assembly map,
which was determined for E. coli,43
may not have
universal applicability. S8, one of the central pro-
teins in the assembly path, is conserved across
3P2O (Table 2), but its RNA binding site is not.
Because of the limited knowledge of MRPs, it is
not yet known whether S15 is conserved, although
it is important for the assembly of the 30 S central
domain.44
The RNA binding site for S15 has been
shown to be crucial for recruitment of the protein,45
but this binding site is not conserved at the 3P2O
level, and it is missing from C. elegans mito-
chondria. Perhaps terminal extensions onto these
proteins could fill the gaps left by the loss of
RNA, but it is not yet clear what effects such exten-
sions might have on subunit assembly.
The remainder of this discussion is devoted to
considering the contacts between the ribosome
and the tRNAs. When considering 3P2O conserva-
tion, this is complicated by the fact that many mito-
chondrial tRNAs have truncations in the D and
T-stem/loop regions,46
so their three-dimensional
structures are different from conventional tRNAs.
Both computer modeling47
and transient electric
230 Modeling a Minimal Ribosome
17. birefringence experiments48
have shown that the
altered secondary structure necessitates substantial
conformational changes in the elbow region,
leading to a large increase in the angle between
the two arms of the tRNA.
Examination of contacts between the ribosomal
RNA and tRNA (Table 5) shows that these are
most highly conserved at the P-site, and least con-
served at the E-site. Contacts with the anticodon,
the elbow, and the 30
acceptor terminus of the
P-site tRNA are all conserved across 3P2O, includ-
ing in C. elegans. At the other extreme, conservation
at the E-site is weak enough that it is possible that a
true E-site may not exist in all ribosomes, particu-
larly those from mitochondria, or, if it exists, it
may have a substantially different structural
organization.
The situation is more complicated at the A-site.
Contacts with the anticodon and acceptor terminus
are preserved across 3P2O, but contacts between
the large subunit and the elbow are not. In the
3P2O model, contacts are lost with both the
D-loop (residues 17 and 19) and the T-loop
(residues 55 and 56), and with the T-stem (residue
62). The contact with residue 62 is probably not
critical, since it is not even conserved at the 3P
level. It is more difficult to interpret the loss of the
other four contacts (Table 5); these are conserved
across the three phylogenetic domains, but not
across 3P2O.
While the non-standard conformation of mito-
chondrial tRNAs might be invoked to explain loss
of contacts with the elbow at the A-site, how are
such contacts then preserved at the P-site? One
possibility is that the regions of the rRNA respon-
sible for these contacts are located in different
positions in mitochondrial ribosomes than in other
ribosomes. Another puzzle is that, since precise
tRNA positioning is presumably critical to trans-
lational fidelity and efficiency, how can mito-
chondrial ribosomes have lost contact with the
elbow of the A-site tRNA? Either A-site positioning
is not as critical as it is at the P-site, or contact with
the tRNA elbow at the A-site is provided by some
other region of the rRNA or by a ribosomal
protein.
Materials and Methods
Comparative analysis
The sequence alignments used for this analysis are
maintained at the CRW site24
†. rRNA sequences were
manually aligned to maximize sequence and structural
identity using the alignment editor AE2 (T. Macke,
Scripps Clinic, San Diego, CA). The rRNA alignments
are sorted by phylogeny and cell location. The sequences
used in this analysis are at least 90% complete and
publicly available. Specific data for each sequence are
accessible from the CRW site’s relational database
management system (RDBMS).
The numbering systems from the E. coli 16 S and 23 S
rRNA sequences (GenBank Accession no. J01695) are
used as the references for position numbers for both
16 S and 23 S rRNAs.
Secondary structure and conservation diagrams were
developed entirely or in part with the interactive
graphics program XRNA (Weiser & Noller, University
of California, Santa Cruz). The PostScript files output by
XRNA were converted into PDF using ghostscript
(version 7.00‡).
Computer modeling and graphics
The models are based on the crystal structures 1FJG
for the 30 S subunit of T. thermophilus,15
1FFK for the
50 S subunit of H. marismortui,14
and 1GIX/1GIY for the
70 S ribosome from T. thermophilus.17
They were
generated and evaluated using Insight II and QUANTA
(Molecular Simulations, Inc.). Areas that are unresolved
in the large subunit crystal structure were modeled
using various techniques. These regions include helix
38, helix 69, and the L1 and L11 binding domains of the
23 S rRNA. The L11–RNA crystal structure49
was super-
imposed onto the 50 S structure (1FFK) using tools in
the Insight II program. The L1 binding domain of the
rRNA was built using the phosphate positions of the
70 S structure as a guide. Helices 38 and 69 were built
manually using Insight. All of these modeled regions
were compared with density maps from cryo-EM studies
of intact 70 S particles.19
Helices 38 and 69 of the large
subunit and the L11-binding domain fit the density
rather well while the L1 domain has to be moved slightly
from the body of the structure to fit the density.
RIBBONS50
was used to create the images in Figure 2.
It was also used to create image files that were processed
by POV-Ray§ to generate Figures 1, 4 and 6.
Protein conservation
Protein conservation was determined from the website
maintained by Dr Nikos C. Kyrpides{ at the University
of Illinois. Information on MRPs is limited to a few
organisms, especially bovine,51
yeast52
and rat systems.53
Therefore, the list of conserved proteins in Table 2 is
incomplete and can be expected to grow as more data
become available.
For this work, we define protein conservation by the
presence of the protein in the ribosome of the species
being studied. The degree of amino acid or domain
conservation within that protein is not considered. It is
conserved if homologous proteins are found in the
genomes of divergent species including in the dataset
being considered for the study (3P or 3P2O).
Probability of movement
In evaluating the model of the C. elegans mitochondrial
ribosome, we have estimated the probability, P(move),
that a deletion has forced movement of some region of
the rRNA to close the resulting gap. This probability is
† http://www.rna.icmb.utexas.edu
‡ http://www.cs.wisc.edu/~ghost/index.htm
§ http://www.povray.org
{http://geta.life.uiuc.edu/~nikos/Ribosome/
rproteins.html
Modeling a Minimal Ribosome 231
18. based on distributions of inter-phosphate distances
measured in the two crystal structures of the 70 S particle
(1GIX and 1GIY). For a deletion where unaligned
sequence is available to span the gap, consecutive
phosphate distances are measured (p þ n, where n is the
sequential phosphate number determined by the length
of the unaligned sequence) generating a distribution of
distances. These can be compared with distances of con-
secutive phosphate groups created by this deletion, and
a probability of movement is then calculated.
Two examples are shown in Figure 5. The first shows
the distribution of distances found in the crystal struc-
ture for phosphate groups that are separated by eight
nucleotides, corresponding to a connection between
phosphate i and phosphate i þ 9. Within the secondary
structure of C. elegans the absence of helix 94 of the large
subunit leaves eight nucleotides of unaligned sequence
to span a gap of 43 A˚ . The distribution of distances
found in the crystal structure shows that for eight
nucleotides, 99.3% of the population has a distance less
than 43 A˚ . In this case, then, P(move) would be 0.993.
The second example in Figure 5 shows the distribution
of distances for phosphate groups that are spaced by 24
nucleotides, connecting phosphate i with phosphate
i þ 25. In C. elegans, the deletion of helices 81–88 (resi-
dues 2259–2421) results in a gap of 26 A˚ to be spanned
by 24 nucleotides of unaligned sequence. Only 19.5% of
the distances in the population are smaller than this, so
P(move) would only be 0.195. In this case the nucleotides
at the ends of the gap could be connected without diffi-
culty, and the affected domains could easily have the
same relative positions in C. elegans as in H. marismortui.
Supplemental information
Figures 1 and 4, showing phylogenetic conservation
and the alignment of C. elegans mitochondrial rRNA,
can be viewed from several perspectives in movies at
the Harvey laboratory website (http://www.uracil.cmc.
uab.edu/Publications). Secondary structures and
additional information about the comparative analyses
can be found at the CRW site maintained by the Gutell
laboratory (http://www.rna.icmb.utexas.edu/ANAL-
YSIS/MINRIB/). Conservation of tRNA binding sites
were studied using a distance criterion (8.5 and 10.0 A˚ )
to determine the conservation of residues near the
mRNA and tRNAs in the crystal structures (1GIX and
1GIY). The findings are comparable to the results pre-
sented in Table 5.
Acknowledgments
We acknowledge Jung C. Lee for work with tRNA con-
tacts and Dr Joachim Frank for helpful discussions. This
work was funded by grants from the National Institute
of Health to S.C.H. (GM-53827), R.R.G. (GM 48207), and
R.K.A. (GM 61576).
References
1. Guerrier-Takada, C., Gardiner, K., Marsh, T., Pace, N.
& Altman, S. (1983). The RNA moiety of ribonuclease
P is the catalytic subunit of the enzyme. Cell, 35,
849–857.
2. Kruger, K., Grabowski, P. J., Zaug, A. J., Sands, J.,
Gottschling, D. E. & Cech, T. R. (1982). Self-splicing
RNA: autoexcision and autocyclization of the riboso-
mal RNA intervening sequence of Tetrahymena. Cell,
31, 147–157.
3. Nissen, P., Hansen, J., Ban, N., Moore, P. B. & Steitz,
T. A. (2000). The structural basis of ribosome activity
in peptide bond synthesis. Science, 289, 920–930.
4. Carter, A. P., Clemons, W. M., Brodersen, D. E., Mor-
gan-Warren, R. J., Wimberly, B. T. & Ramakrishnan,
V. (2000). Functional insights from the structure of
the 30S ribosomal subunit and its interactions with
antibiotics. Nature, 407, 340–348.
5. Fourmy, D., Recht, M. I., Blanchard, S. C. & Puglisi, J.
D. (1996). Structure of the A site of Escherichia coli 16S
ribosomal RNA complexed with an aminoglycoside
antibiotic. Science, 274, 1367–1371.
6. VanLoock, M. S., Easterwood, T. R. & Harvey, S. C.
(1999). Major groove binding of the tRNA/mRNA
complex to the 16 S ribosomal RNA decoding site.
J. Mol. Biol. 285, 2069–2078.
7. Agrawal, R. K., Heagle, A. B., Penczek, P., Grassucci,
R. A. & Frank, J. (1999). EF-G-dependent GTP
hydrolysis induces translocation accompanied by
large conformational changes in the 70S ribosome.
Nature Struct. Biol. 6, 643–647.
8. Frank, J. & Agrawal, R. K. (2000). A ratchet-like inter-
subunit reorganization of the ribosome during trans-
location. Nature, 406, 318–322.
9. Agrawal, R. K., Penczek, P., Grassucci, R. A. & Frank,
J. (1998). Visualization of elongation factor G on the
Escherichia coli 70S ribosome: the mechanism of
translocation. Proc. Natl Acad. Sci. USA, 95,
6134–6138.
10. Agrawal, R. K., Spahn, C. M., Penczek, P., Grassucci,
R. A., Nierhaus, K. H. & Frank, J. (2000). Visualiza-
tion of tRNA movements on the Escherichia coli 70S
ribosome during the elongation cycle. J. Cell Biol.
150, 447–460.
11. Stark, H., Rodnina, M. V., Rinke-Appel, J.,
Brimacombe, R., Wintermeyer, W. & van Heel, M.
(1997). Visualization of elongation factor Tu on the
Escherichia coli ribosome. Nature, 389, 403–406.
12. Ogle, J. M., Brodersen, D. E., Clemons, W. M., Jr,
Tarry, M. J., Carter, A. P. & Ramakrishnan, V. (2001).
Recognition of cognate transfer RNA by the 30S ribo-
somal subunit. Science, 292, 897–902.
13. Lodmell, J. S. & Dahlberg, A. E. (1997). A confor-
mational switch in Escherichia coli 16S ribosomal
RNA during decoding of messenger RNA. Science,
277, 1262–1267.
14. Ban, N., Nissen, P., Hansen, J., Moore, P. B. & Steitz,
T. A. (2000). The complete atomic structure of the
large ribosomal subunit at 2.4 A˚ resolution. Science,
289, 905–920.
15. Wimberly, B. T., Brodersen, D. E., Clemons, W. M., Jr,
Morgan-Warren, R. J., Carter, A. P., Vonrhein, C. et al.
(2000). Structure of the 30S ribosomal subunit.
Nature, 407, 327–339.
16. Schluenzen, F., Tocilj, A., Zarivach, R., Harms, J.,
Gluehmann, M., Janell, D. et al. (2000). Structure of
functionally activated small ribosomal subunit at
3.3 A˚ resolution. Cell, 102, 615–623.
17. Yusupov, M. M., Yusupova, G. Z., Baucom, A.,
Lieberman, K., Earnest, T. N., Cate, J. H. & Noller,
H. F. (2001). Crystal structure of the ribosome at
5.5 A˚ resolution. Science, 292, 883–896.
232 Modeling a Minimal Ribosome
19. 18. Agrawal, R. K. & Frank, J. (1999). Structural studies
of the translational apparatus. Curr. Opin. Struct.
Biol. 9, 215–221.
19. Gabashvili, I. S., Agrawal, R. K., Spahn, C. M.,
Grassucci, R. A., Svergun, D. I., Frank, J. & Penczek,
P. (2000). Solution structure of the E. coli 70S ribo-
some at 11.5 A˚ resolution. Cell, 100, 537–549.
20. Ramakrishnan, V. & Moore, P. B. (2001). Atomic
structures at last: the ribosome in 2000. Curr. Opin.
Struct. Biol. 11, 144–154.
21. Gutell, R. R., Larsen, N. & Woese, C. R. (1994).
Lessons from an evolving rRNA: 16S and 23S rRNA
structures from a comparative perspective. Microbiol.
Rev. 58, 10–26.
22. Gutell, R. R., Weiser, B., Woese, C. R. & Noller, H. F.
(1985). Comparative anatomy of 16S-like ribosomal
RNA. Prog. Nucl. Acid. Res. Mol. Biol. 32, 155–216.
23. Gutell, R. R. (1992). Evolutionary characteristics of
16S and 23S rRNA structures. In Proceedings of the
Conference on the Origin and Evolution of Prokaryotic
and Eukaryotic Cells. The Origin and Evolution of the
Cell (Hartman, H. & Matsuno, K., eds), World
Scientific, Singapore.
24. Cannone, J. J., Subramanian, S., Schnare, M. N.,
Collett, J. R., D’Souza, L. M., Du, Y. et al. (2002). The
Comparative RNA Web (CRW) site: an online data-
base of comparative sequence and structure infor-
mation for ribosomal, intron, and other RNAs. BMC
Bioinformatics, 3, 2.
25. Frank, J., Zhu, J., Penczek, P., Li, Y., Srivastava, S.,
Verschoor, A. et al. (1995). A model of protein syn-
thesis based on cryo-electron microscopy of the
E. coli ribosome. Nature, 376, 441–444.
26. Malhotra, A., Penczek, P., Agrawal, R. K., Gabashvili,
I. S., Grassucci, R. A., Junemann, R. et al. (1998).
Escherichia coli 70 S ribosome at 15 A˚ resolution by
cryo-electron microscopy: localization of fMet-
tRNAfMet and fitting of L1 protein. J. Mol. Biol. 280,
103–116.
27. Wuyts, J., Van de Peer, Y. & De Wachter, R. (2001).
Distribution of substitution rates and location of
insertion sites in the tertiary structure of ribosomal
RNA. Nucl. Acids Res. 29, 5017–5028.
28. Yamaguchi, K. & Subramanian, A. R. (2000). The
plastid ribosomal proteins. Identification of all the
proteins in the 50 S subunit of an organelle ribosome
(chloroplast). J. Biol. Chem. 275, 28466–28482.
29. Yamaguchi, K., von Knoblauch, K. & Subramanian,
A. R. (2000). The plastid ribosomal proteins. Identifi-
cation of all the proteins in the 30S subunit of an
organelle ribosome (chloroplast). J. Biol. Chem. 275,
28455–28465.
30. Agrawal, R. K., Linde, J., Sengupta, J., Nierhaus, K.
H. & Frank, J. (2001). Localization of L11 protein on
the ribosome and elucidation of its involvement in
EF-G-dependent translocation. J. Mol. Biol. 311,
777–787.
31. Koc, E. C., Burkhart, W., Blackburn, K., Koc, H.,
Moseley, A. & Spremulli, L. L. (2001). Identification
of four proteins from the small subunit of the mam-
malian mitochondrial ribosome using a proteomics
approach. Protein Sci. 10, 471–481.
32. Koc, E. C., Burkhart, W., Blackburn, K., Moyer, M. B.,
Schlatzer, D. M., Moseley, A. & Spremulli, L. L.
(2001). The large subunit of the mammalian mito-
chondrial ribosome. Analysis of the complement of
ribosomal proteins present. J. Biol. Chem. 276,
43958–43969.
33. Suzuki, T., Terasaki, M., Takemoto-Hori, C., Hanada,
T., Ueda, T., Wada, A. & Watanabe, K. (2001).
Structural compensation for the deficit of rRNA
with proteins in the mammalian mitochondrial ribo-
some. J. Biol. Chem. 276, 21724–21736.
34. Suzuki, T., Terasaki, M., Takemoto-Hori, C., Hanada,
T., Ueda, T., Wada, A. & Watanabe, K. (2001). Struc-
tural compensation for the deficit of rRNA with pro-
teins in the mammalian mitochondrial ribosome.
Systematic analysis of protein components of the
large ribosomal subunit from mammalian mito-
chondria. J. Biol. Chem. 276, 21724–21736.
35. Okimoto, R., MacFarlane, J. L., Clary, D. O. &
Wolstenholme, D. R. (1992). The mitochondrial gen-
omes of two nematodes, Caenorhabditis elegans and
Ascaris suum. Genetics, 130, 471–498.
36. Morosyuk, S. V., Cunningham, P. R. & SantaLucia, J.
(2001). Structure and function of the conserved 690
hairpin in Escherichia coli 16 S ribosomal RNA. II.
J. Mol. Biol. 307, 197–211.
37. Morosyuk, S. V., SantaLucia, J. & Cunningham, P. R.
(2001). Structure and function of the conserved 690
hairpin in Escherichia coli 16 S ribosomal RNA. III.
J. Mol. Biol. 307, 213–228.
38. Gabashvili, I. S., Agrawal, R. K., Grassucci, R.,
Squires, C. L., Dahlberg, A. E. & Frank, J. (1999).
Major rearrangements in the 70S ribosomal 3D
structure caused by a conformational switch in 16S
ribosomal RNA. EMBO J. 18, 6501–6507.
39. VanLoock, M. S., Agrawal, R. K., Gabashvili, I. S., Qi,
L., Frank, J. & Harvey, S. C. (2000). Movement of the
decoding region of the 16 S ribosomal RNA accom-
panies tRNA translocation. J. Mol. Biol. 304, 507–515.
40. Velichutina, I. V., Dresios, J., Hong, J. Y., Li, C.,
Mankin, A., Synetos, D. & Liebman, S. W. (2000).
Mutations in helix 27 of the yeast Saccharomyces
cerevisiae 18S rRNA affect the function of the decod-
ing center of the ribosome. RNA, 6, 1174–1184.
41. Lee, K., Varma, S., SantaLucia, J., Jr & Cunningham,
P. R. (1997). In vivo determination of RNA struc-
ture–function relationships: analysis of the 790 loop
in ribosomal RNA. J. Mol. Biol. 269, 732–743.
42. Yusupova, G. Z., Yusupov, M. M., Cate, J. H. &
Noller, H. F. (2001). The path of messenger RNA
through the ribosome. Cell, 106, 233–241.
43. Held, W. A., Ballou, B., Mizushima, S. & Nomura, M.
(1974). Assembly mapping of 30S ribosomal proteins
from Escherichia coli. Further studies. J. Biol. Chem.
249, 3103–3111.
44. Agalarov, S. C., Sridhar Prasad, G., Funke, P. M.,
Stout, C. D. & Williamson, J. R. (2000). Structure of
the S15,S6,S18–rRNA complex: assembly of the 30S
ribosome central domain. Science, 288, 107–113.
45. Serganov, A., Benard, L., Portier, C., Ennifar, E.,
Garber, M., Ehresmann, B. & Ehresmann, C. (2001).
Role of conserved nucleotides in building the 16 S
rRNA binding site for ribosomal protein S15. J. Mol.
Biol. 305, 785–803.
46. Wolstenholme, D. R., MacFarlane, J. L., Okimoto, R.,
Clary, D. O. & Wahleithner, J. A. (1987). Bizarre
tRNAs inferred from DNA sequences of mito-
chondrial genomes of nematode worms. Proc. Natl
Acad. Sci. USA, 84, 1324–1328.
47. Steinberg, S. & Cedergren, R. (1994). Structural com-
pensation in atypical mitochondrial tRNAs. Nature
Struct. Biol. 1, 507–510.
48. Frazer-Abel, A. A. & Hagerman, P. J. (1999). Determi-
nation of the angle between the acceptor and
Modeling a Minimal Ribosome 233
20. anticodon stems of a truncated mitochondrial tRNA.
J. Mol. Biol. 285, 581–593.
49. Wimberly, B. T., Guymon, R., McCutcheon, J. P.,
White, S. W. & Ramakrishnan, V. (1999). A detailed
view of a ribosomal active site: the structure of the
L11–RNA complex. Cell, 97, 491–502.
50. Carson, M. (1997). Ribbons. Methods Enzymol. 277,
493–505.
51. Pietromonaco, S. F., Denslow, N. D. & O’Brien, T. W.
(1991). Proteins of mammalian mitochondrial ribo-
somes. Biochimie, 73, 827–835.
52. Graack, H. R. & Wittmann-Liebold, B. (1998).
Mitochondrial ribosomal proteins (MRPs) of yeast.
Biochem. J. 329, 433–448.
53. Goldschmidt-Reisin, S., Kitakawa, M., Herfurth, E.,
Wittmann-Liebold, B., Grohmann, L. & Graack, H.
R. (1998). Mammalian mitochondrial ribosomal
proteins. N-terminal amino acid sequencing, charac-
terization, and identification of corresponding gene
sequences. J. Biol. Chem. 273, 34828–34836.
Edited by D. E. Draper
(Received 29 April 2002; received in revised form 4 June 2002; accepted 5 June 2002)
234 Modeling a Minimal Ribosome