(Received for publication, November 7, 1996, and in revised form, February 19, 1997)
From the Max-Delbrück-Centrum für Molekulare Medizin,
Robert-Rössle-Strae 10, D-13125 Berlin, Germany
Cross-linked peptide-oligoribonucleotide complexes derived from distinct regions of the rRNA and individual ribosomal proteins of the 30 S ribosomal subunits from Escherichia coli were isolated and purified. Cross-linking sites at the amino acid and nucleotide level were determined by N-terminal amino acid sequence analysis in combination with matrix-assisted laser desorption/ionization mass spectrometry (MALDI-MS). MALDI-MS analysis performed subsequent to a partial alkaline hydrolysis of cross-linked peptide-oligoribonucleotide complexes allowed for the first time the cross-linked rRNA moiety to be sequenced by this technique. In this manner Lys-44 in S4 was determined to be cross-linked to the oligoribonucleotide at positions 1531-1542 on the 16 S RNA (whereby either U-1541 or A-1542 is the actual cross-link site), Lys-75 in S7 to positions 1374-1379 (C-1378 cross-linked), Met-114 in S7 to 1234-1241 (U-1240 cross-linked), Lys-55 in S8 to 651-654 (U-653 cross-linked), and Lys-29 in S17 to 629-633 (U-632 cross-linked). The novel approach applied here promises to be useful for similar studies on other known protein·RNA complexes.
The quarternary structure of the ribosome from the eubacterium Escherichia coli has been studied intensively for many years, and several individual models concerning either the protein composition and the arrangement of proteins (e.g. Refs. 1-5) or the three-dimensional folding of 16 S RNA including the protein-rRNA interactions (e.g. Refs. 6-9) have been proposed. However, the resolution of these models at the molecular level is still limited, since the contact sites between adjacent proteins and between proteins and rRNA are not precisely known at the amino acid level with only a few exceptions of cross-linked protein pairs (e.g. Refs. 10-12), although a number of protein contact sites have been identified at the nucleotide level (for review, see Ref. 9). More recently, detailed protein-rRNA cross-linking studies at the peptide level (13) in combination with three-dimensional structures of isolated ribosomal proteins (e.g. Refs. 14-17) have provided more insight into the protein-RNA interactions and their functional implications within the prokaryotic ribosome. Additionally, site-directed hydroxyl radical probing of the rRNA neighborhood in reconstituted ribosomal particles has provided information relevant to the three-dimensional orientation of several proteins within the 30 S subunit (18, 19).
Even though discrete peptide regions of individual ribosomal proteins in close contact to the rRNA have been clearly established (13), the analysis of the corresponding sites on the rRNA has still remained a problem. Detailed modeling of ribosomal structures requires precise knowledge of protein-nucleotide and peptide-RNA contact sites concomitantly at the amino acid and nucleotide level. With the help of such information, the three-dimensional structures of ribosomal proteins can be placed into current models of the rRNA, and hence refined models can be adapted to the overall topography of the prokaryotic ribosome as derived from cryo-electron microscopy data (20, 21). For this purpose we have developed a new approach that enables us to determine simultaneously the contact sites at the peptide as well as at the nucleotide level within cross-linked peptide-oligoribonucleotide complexes isolated from ribosomal subunits. The strategy employed is based on N-terminal sequence analysis combined with matrix-assisted laser desorption/ionization mass spectrometry (MALDI-MS).1
MALDI-MS (22) has been successfully applied for e.g. the determination of phosphoproteins, glycoproteins, and oligonucleotides (for review, see Ref. 23-25). Furthermore, applications of MALDI-MS for sequencing of peptides (e.g. Refs. 26-28) and of oligodeoxynucleotides (e.g. Refs. 29 and 30) have been described. Here, we report for the first time the sequencing of oligoribonucleotides by MALDI-MS cross-linked to peptides. The strategy applied here should be a valuable tool for studying protein-RNA interactions, not only in ribosomes but also in other protein·RNA complexes.
Preparation of 30 S ribosomal subunits from E. coli (Eco 30S), chemical cross-linking of ribosomal subunits with 2-iminothiolane, and generation of cross-linked peptide-rRNA heteromers were performed according to Urlaub et al. (13).
C18 Reversed Phase (RP) High Performance Liquid Chromatography (HPLC)The rRNA pool derived from 150 A260 cross-linked ribosomal subunits treated
with endoproteinases and RNase T1 (13) was subjected to a
second incubation with endoproteinases Lys-C, Glu-C, or trypsin. Isolation of the cross-linked peptide-oligoribonucleotide complexes was
achieved by RP-HPLC (13). Fractions that showed an absorbance at 220 and 260 nm were collected, and cross-linked peptide-oligoribonucleotide complexes were precipitated with 2 volumes of 98% ethanol in the presence of volume of 3 M sodium acetate, pH
5.5, and 40 µg of glycogen (Boehringer Mannheim) for at least 2 h at
80 °C. After centrifugation at 13,000 rpm for 1 h at
4 °C the pellet was immediately redissolved in 25 mM
Tris-HCl, pH 7.8, 1 mM EDTA, 10% (v/v) acetonitrile, and
0.1% (v/v) Triton® X-100 (hydrogenated, protein grade, Calbiochem).
Precipitated and redissolved fractions were rechromatographed in the
same buffer system as described in Ref. 13 using a gradient as follows: 15-min isocratic elution with 10% buffer B, then 10-45% buffer B for
120 min, and 45-90% buffer B for 10 min. Fractions with an absorbance
at 220 and 260 nm were dried under vacuum and cross-linked peptide·rRNA complexes were redissolved in 50% (v/v) acetonitrile, 0.1% (v/v) trifluoroacetic acid, divided into aliquots for further studies and stored at
80 °C.
For identification of the cross-linked peptide moiety, up to 5 pmol of cross-linked peptide-oligoribonucleotide complex in 50% (v/v) acetonitrile, 0.1% (v/v) trifluoroacetic acid was sequenced in a PROCISETM protein sequencer (Applied Biosystems Inc., Foster City, CA, USA). The SWISS-PROT data bank was used for comparison with the known E. coli protein sequences by the program FASTA (31).
MALDI-MSMass spectra of cross-linked
peptide-oligoribonucleotide complexes were recorded in the linear mode
of a time of flight VG-TofSpec (Fisons, Manchester, United Kingdom)
equipped with a nitrogen laser (337 nm, pulse duration: 4 ns) according
to Thiede et al. (28). The acceleration voltage was 22 kV,
and the spectra were obtained either in the positive or in the negative
mode by summing over 20-50 laser shots. For mass determination of
untreated cross-linked complexes, a 0.8-µl sample in aqueous 50%
(v/v) acetonitrile, 0.1% (v/v) trifluoroacetic acid (containing 1-10
pmol of cross-linked peptide-oligoribonucleotide complex) and 1.2 µl
of matrix solution (a saturated solution of -cyano-4-hydroxycinnamic
acid in water:acetonitrile, 3:2; containing 0.1% (v/v) trifluoroacetic
acid) were mixed on the target and air-dried. Alternatively, the sample
was incubated with ammonium acetate activated ion exchange matrix (AG®
50W-X8 Resin, 100-200 mesh, Bio-Rad) as described by Nordhoff et
al. (32) for 30 min at 37 °C. 5-10 ion exchange beads were
sufficient for incubation. 0.8 µl of the supernatant was mixed with
the matrix as above. Nucleotide compositions of the cross-linked
oligoribonucleotides were determined from the mass spectra of untreated
samples by the program NUCSEQ.2 Partial
alkaline hydrolysis of the cross-linked oligoribonucleotide-peptide complex was carried out by treating 1-10 pmol of vacuum-dried sample
with ammonium hydroxide solution at pH 10.0 for 10 min at 95 °C. The
sample was dried again under vacuum and redissolved in an appropriate
volume of aqueous 50% (v/v) acetonitrile, 0.1% (v/v) trifluoroacetic
acid. Mass spectra were recorded as described above. Alternatively,
vacuum-dried samples were redissolved in 5 µl of water and incubated
with 5
>3
phosphodiesterase (from calf spleen, Boehringer
Mannheim) as described by Kirpekar et al. (29) and by Crain
(33). For mass analysis samples were prepared as above.
30 S ribosomal subunits from E. coli (Eco 30S) were cross-linked with 2-iminothiolane followed by
a short UV irradiation. Ribosomal proteins cross-linked to 16 S rRNA
and the corresponding peptide-oligoribonucleotide heteromers were then
isolated as described by Urlaub et al. (13). Fig.
1A shows a representative RP-HPLC separation
of cross-linked peptide-oligoribonucleotide heteromers derived from 100 A260 cross-linked Eco 30S after treatment with endoproteinase Lys-C and RNase T1, and Fig. 1B
shows an RP-HPLC separation of the same amount of cross-linked
peptide-oligoribonucleotide heteromers generated after treatment of
cross-linked Eco 30S with endoproteinase Glu-C and RNase
T1. In contrast to our previous studies, the isolation of
cross-linked peptide-oligoribonucleotide via RP-HPLC could be markedly
improved by an additional proteolytic digestion with the same
endoproteinase subsequent to fragmentation of the rRNA with RNase
T1. For further purification of the cross-linked peptide-oligoribonucleotide complexes, those fractions which showed an
absorbance at 220 and 260 nm (the latter corresponding to the cross-linked oligoribonucleotide moiety) were collected, precipitated, redissolved, and rechromatographed as described in the experimental procedures. Examples of rechromatographed peptide-oligoribonucleotide complexes (fractions 1 and 2, respectively) are depicted in the inserts
within the corresponding chromatograms (Fig. 1, A and B). Approximately 80% of the cross-linked complex could be
recovered after precipitation. The cross-linked peptide sequence and
the cross-linked amino acid of the complex was identified by automated N-terminal sequencing (13).
Sequence Analysis of an rRNA-Oligoribonucleotide Cross-linked to Met-114 in S7
Fig. 2A shows the MALDI
mass spectrum of the total peptide-oligoribonucleotide complex from
fraction 1 (see inset of Fig. 1A). The peptide in
the cross-linked complex, a Lys-C fragment, originated from ribosomal
protein S7 (positions 113-129,
Ser-Met-Ala-Leu-Arg-Leu-Ala-Asn-Glu-Leu-Ser-Asp-Ala-Ala-Glu-Asn-Lys, with Met-114 cross-linked to the 16 S rRNA) was determined by N-terminal amino acid sequencing. The mass difference between that of
the total complex (Mr: 4405, see mass peak
k in Fig. 2, A and B) and that of the
cross-linked peptide (with a calculated Mr of
1833 without 2-iminothiolane) resulted in the following possible
compositions of the cross-linked T1 fragment:
A3C4Gp, A3C3U1Gp,
A3C2U2Gp,
A3C1U3Gp, and
A3U4Gp (program NUCSEQ).2 The mass
resolution of the spectrometer employed is 200-300 for peptides in the
linear mode. However, the resolution of the detected masses of the
cross-linked complexes was relatively low compared with those for
peptides (see "Discussion"), and hence unambiguous differentiation
between C (Mr: 323) and U
(Mr: 324) was not possible. Accordingly,
different compositions of this 8-mer oligoribonucleotide were
considered. The E. coli 16 S rRNA gene sequence (34)
contains three T1 fragments that have base compositions
corresponding to those we obtained (positions 559-566, 675-682, and
1234-1241). The covalent attachment of Met-114 in S7 to the
oligoribonucleotide resulted directly from the UV irradiation
subsequent to the cross-linking reaction and thus was independent of
the cross-linking reagent 2-iminothiolane (which reacts exclusively
with the -group of lysine residues (13)). Therefore, the mass of
2-iminothiolane (Mr: 101) was not included in
the calculation of the nucleotide composition in this case.
Fig. 2B shows the MALDI mass spectrum of the cross-linked
complex after partial alkaline hydrolysis. The mass peaks obtained, marked a-k, are analyzed in Table I.
Mass peaks a-d correspond to oligoribonucleotides
containing no cross-linked peptide. These were cleaved off during the
hydrolysis of the total cross-linked complex. The higher mass
peaks e-k are fragments of the partially hydrolyzed cross-linked
complex to which the peptide is still covalently attached. Note that
the non-cross-linked cleavage products of the oligoribonucleotide
(mass peaks a-d), as well as the fragment lacking the Gp at
the 3 end (mass peak j) have 5
-hydroxyl groups and cyclic
2
,3
-phosphate termini (marked by
H20 in Table I) due
to the partial alkaline hydrolysis (35). The given nucleotide sequence
from the 5
to 3
terminus (see Table I) was determined from the mass
differences of peaks e-k and the fact that
T1-oligoribonucleotides contain a guanosine
3
-monophosphate at their 3
termini. Comparison of the mass
peaks e and j leads to the identification of a C or U
residue as the cross-linked nucleoside 5
to Gp (bold letters, see
Table I). Despite this C/U ambiguity, the sequence of the cross-linked
T1-oligoribonucleotide matches only one of the three possible 8-mer T1-oligoribonucleotides (see above), namely
positions 1234-1241 on the 16 S RNA. Thus, U-1240 could be identified
as the site which cross-linked to Met-114 in S7. Fig. 2C
shows the sequence of the T1-oligoribonucleotide
cross-linked to Met-114 in S7 as determined by the mass fragments after
partial alkaline hydrolysis.
|
The MALDI-MS analysis of the
peptide-oligoribonucleotide complex derived from ribosomal protein S4
cross-linked to 16 S rRNA (positions 35-56,
Gln-Ala-Pro-Gly-Gln-His-Gly-Ala-Arg-Lys-Pro-Arg-Leu-Ser-Asp-Tyr-Gly-Val-Gln-Leu-Arg-Glu, with Lys-44 cross-linked, fraction 2 in Fig. 1B) is
illustrated in Fig. 3. Fig. 3A shows the mass
spectrum of the untreated cross-linked complex, and Fig. 3B
shows the spectrum of the cross-linked complex after partial alkaline
hydrolysis. Interestingly, the mass of the total complex (mass
peak k, Mr: 6242 and 6243, respectively) as well as
the mass of the cross-linked peptide derivatized with the
2-iminothiolane spacer (mass peak a, Mr: 2567 and 2568, respectively) occur in both spectra. Complexes in which the
peptide is cross-linked to the oligoribonucleotide via 2-iminothiolane
seem to be less stable under MALDI conditions than complexes derived
after direct photoinduction (e.g. the S7 Met-114/U-1240
cross-link described above). After partial alkaline hydrolysis the mass
spectrum revealed an almost perfect sequence ladder of the cross-linked
T1-oligoribonucleotide (mass peaks b-j, Fig.
3B). The mass peaks a-k are listed and
interpreted in Table II. The sequence of the
cross-linked T1-oligoribonucleotide obtained correlates
exclusively with the T1 fragment of the 16 S rRNA directly
at its 3 terminus. Mass peak b corresponds to a fragment of
the hydrolyzed complex consisting of the derivatized peptide
cross-linked to UAOH or CAOH and thus
corresponds to the extreme 3
end of the rRNA. Mass peaks
c-j differ by the masses of the nucleotide 3
-monophosphates
shown, whereas mass peaks k and j differ by a
dinucleotide of either AC or AU at the 5
end of the 12-mer
oligoribonucleotide. Fig. 3C shows the sequence of the
T1-oligoribonucleotide from positions 1531-1542, with
either U-1541 or A-1542 directly cross-linked to Lys-44 within the
Glu-C fragment of S4.
|
A second peptide-oligoribonucleotide complex derived from ribosomal protein S7 cross-linked to the 16 S RNA could be isolated after treatment of cross-linked ribosomes with Glu-C and RNase T1 (fraction 3 in Fig. 1B, chromatogram after purification not shown) and analyzed as described above. The peptide (positions 74-89, Val-Lys-Ser-Arg-Arg-Val-Gly-Gly-Ser-Thr-Tyr-Gln-Val-Pro-Val-Glu, with Lys-75 cross-linked) was found to be cross-linked to a 6-mer T1 fragment of the composition A3C2Gp, A3C1U1Gp, or A3U2Gp (mass spectrum not shown). Four T1 fragments in the 16 S RNA have these base compositions (positions 129-135, 823-828, 934-939, and 1374-1379). The results of the mass spectrum obtained subsequent to a partial alkaline hydrolysis of the cross-linked complex (mass spectrum not shown) resulted in the oligoribonucleotide sequence of AACACGp, AACAUGp, AAUACGp, or AAUAUGp, which correlates only with the T1 fragment from positions 1374-1379 in the 16 S RNA. The site of cross-linking of the oligoribonucleotide to Lys-75 in S7 was identified as the nucleotide C-1378.
Sequence Analysis of rRNA-Oligoribonucleotides Cross-linked to Lys-29 in S17 and to Lys-55 in S8Peptide-oligoribonucleotide
complexes derived from the 16 S rRNA cross-linked to either the
ribosomal protein S17 (fraction 4 in Fig. 1A, positions
19-35,
Ser-Ile-Val-Val-Ala-Ile-Glu-Arg-Phe-Val-Lys-His-Pro-Ile-Tyr-Gly-Lys, with Lys-29 cross-linked) or the ribosomal protein S8
(chromatogram not shown, positions 33-63,
Val-Ala-Ile-Ala-Asn-Val-Leu-Lys-Glu-Glu-Gly-Phe-Ile-Glu-Asp-Phe-Lys-Val-Glu-Gly-Asp-Thr-Lys-Pro-Glu-Leu-Glu-Leu-Thr-Leu-Lys, with Lys-55 cross-linked) were isolated after digestion with RNase T1 and Lys-C (for the 16 S-S17 complex), or trypsin (for
the 16 S-S8 complex). Sequence analysis of the cross-linked peptide
moieties by N-terminal amino acid sequencing and of the
T1-oligoribonucleotides by MALDI-MS were carried out as
described above. For S17, a 5-mer T1-oligoribonucleotide
was found to be cross-linked to Lys-29 with the base sequence of
AACCGp, AACUGp, AAUCGp, or AAUUGp, and the cross-linked nucleotide was
identified as either C or U directly 5 to the 3
Gp (mass spectra not
shown). Four T1 fragments of the 16 S rRNA coincide with
the calculated sequences (positions 320-324, positions 629-633,
positions 918-922, positions 1339-1343).
The corresponding analysis of the S8 complex revealed a 4-mer
T1-oligoribonucleotide cross-linked to Lys-55 in S8, with
either C or U directly 5 to the 3
Gp as the actual cross-linking site (mass spectra not shown). Because of its base composition
(C3Gp, C2U1Gp,
C1U2Gp, or U3Gp), no sequence could
be clearly assigned. Comparison of these calculated compositions with
the 16 S RNA sequence (34) shows sequence similarity with 13 possible
T1 fragments.
To determine and sequence the oligoribonucleotides cross-linked to peptides in 30 S ribosomal subunits, we have successfully applied MALDI-MS. To date, similar studies using MALDI-MS have only been carried out on cross-linked protein-DNA complexes (36), but without oligonucleotide sequence determination.
A novel strategy for isolation and purification of peptide-oligoribonucleotide complexes derived from cross-linked ribosomal subunits was recently established that led to the identification of peptides from 13 different ribosomal proteins cross-linked either to the 16 S or the 23 S RNA (13). However, in this previous study, only the cross-linking sites within the peptides, but not within the corresponding cross-linked oligoribonucleotides could be identified. A second proteolytic cleavage subsequent to the rRNA fragmentation enabled us to obtain cross-linked complexes in sufficient amounts to be rechromatographed to purity and analyzed simultaneously by N-terminal amino acid sequencing and MALDI-MS. By combining these techniques, we were able to identify the cross-linked oligoribonucleotide moiety that contacts distinct regions of the ribosomal proteins S4, S7, S8, and S17 from E. coli. Moreover, the cross-linking sites could be analyzed at the nucleotide/amino acid level.
The covalent linkage between the peptide and the oligoribonucleotide
generated by the cross-linking reaction led to a series of complexes
which resembled derivatized peptides. Thus, their masses could be
measured by MALDI-MS under standard conditions for peptides by using
-cyano-4-hydroxycinnamic acid as a matrix (28). MALDI-MS has the
advantage of being much less sensitive to salt contaminations than the
ESI-MS technique (28, 29, 37) and generally generates data of mixtures
that are easy to interpret. Accordingly, the cross-linked complexes
could be analyzed by MALDI-MS directly after partial hydrolysis with
ammonium hydroxide solution or treatment with phosphodiesterase,
avoiding further purification steps. However, MALDI mass spectra of
oligoribonucleotide-containing samples are often poorly resolved and
show heterogeneity when compared with peptide or protein mass spectra.
In particular, a variety of metal counterions that interact with the
phosphate backbone of the oligoribonucleotides caused broad or multiple mass peaks of the total complex with increasing numbers of nucleotides. This is a well known fact from MALDI-MS analysis of
oligodeoxynucleotides (32). In the case of the complexes derived from
ribosomal proteins S7 (Met-114 cross-linked to U-1240; Lys-75
cross-linked to C-1378) and S4 (Lys-44 cross-linked to U-1541 or
A-1542), the exact mass of each total cross-linked complex could be
determined only after performing ion exchange according to Nordhoff
et al. (32). Increased mass accuracy of the total complexes
was also observed when the samples were incubated directly with
saturated ammonium sulfate solution (37). The nucleotide compositions
of all the cross-linked complexes were calculated from the mass
differences between the sequenced peptides plus their iminothiolane
spacer (except in the case of the S7 fragment cross-linked via Met-114
to the 16 S RNA; see "Results") and the measured mass of the total
complexes. Different compositions with respect to the nucleotides C and
U were considered, due to the low mass resolution of the total complex (see above). Additionally, the samples could be analyzed only in the
linear mode of the mass spectrometer, which has a decreased resolution
in comparison with the reflectron mode. Consequently, the calculated
composition of cross-linked oligoribonucleotide moiety was not
sufficient for an unambiguous identification of the corresponding
positions in the 16 S RNA.
Despite the decreased mass accuracy, a partial alkaline hydrolysis of the total complexes followed by MALDI-MS analysis enabled us to obtain adequate sequence information of the cross-linked oligoribonucleotide for a localization of the corresponding positions in the 16 S RNA. By applying this sequencing strategy we identified Met-114 in S7 cross-linked to U-1240 on the 16 S RNA, thereby confirming previous results by Möller et al. (38) who determined Met-114 to be cross-linked to the oligoribonucleotide stretch 1238-1243 after direct UV irradiation. Furthermore, our data provide direct evidence for an additional interaction site between S7 and the 16 S RNA. We were able to isolate a second cross-linked peptide-oligoribonucleotide complex in which Lys-75 in S7 was cross-linked to U-1378 in the 16 S RNA. Nucleotides 1377-1378 have also been cross-linked with 2-iminothiolane in earlier studies to the intact S7 protein (39), but without determination of the cross-linked peptide. The two sites in S7 (Lys-75 and Met-114) which were cross-linked to nucleotides U-1378 and U-1240, respectively, are consistent with the data of Dragon et al. (40), who generated mutations in G-U base pairs in helices 41 and 43 (using here and in the following the helix nomenclature of Ref. 9). The mutation sites are close to the cross-linked oligoribonucleotide regions (positions 1234-1241 and 1374-1379, respectively) and alter the binding of S7 to the rRNA.
In S4 two sites which cross-linked to the 16 S RNA have now been
established, namely the peptide stretch 77-90 with Lys-82 cross-linked
(13) and 35-56 with Lys-44 cross-linked (this work). MALDI-MS analysis
of the latter after partial alkaline hydrolysis of the cross-linked
peptide-oligoribonucleotide complex revealed an almost perfect sequence
ladder of the cross-linked oligoribonucleotide moiety. The
cross-linking site to Lys-44 in S4 was found to be the extreme 3
terminus of the 16 S RNA. Cross-linking studies with a different
reagent (41) as well as studies with base probes (7) and free or
directed hydroxyl radical probing (18, 42) have demonstrated an
interaction of S4 with parts of helices 16, 17, 18, and 23a of the 16 S
RNA. Nevertheless, the specificity of 2-iminothiolane as a
cross-linking reagent has been demonstrated by the isolation of
homologous peptide stretches derived from ribosomal proteins of related
organisms cross-linked to the 16 S RNA (13). Therefore, our data show
that the 3
terminus of the 16 S RNA, which is known to be very
flexible, is also able to contact the ribosomal protein S4.
MALDI-MS analysis of the S17 complex showed that four T1
fragments of the 16 S RNA have to be considered to be putative
cross-linked oligoribonucleotides to Lys-29 in S17, namely positions
320-324, 629-633, 918-922, or 1339-1343. The site of cross-linking
was identified to be either C or U directly 5 to the 3
Gp of the 5-mer T1 fragment. By comparison with previous
cross-linking studies on the 16 S RNA, we conclude that U-632 of the 16 S RNA T1-oligoribonucleotide 629-633 is cross-linked to
Lys-29 in S17, since exactly the same region was also identified to be
cross-linked with 2-iminothiolane to the intact S17 protein (43). Other
cross-linking studies with bis-(2-chloroethyl)methylamine (41) and
chemical footprinting data (7) point to helix 11 of the 16 S RNA as a
second interaction site to S17. In the three-dimensional structure of
S17 from Bacillus stearothermophilus, two loop regions (loop
1 and 3, see Fig. 4A) were proposed to
interact with the the 16 S RNA (17). Our results indicate that loop 1 in S17, which harbors the cross-linked lysine-residue (Lys-29 in
E. coli; Lys-31 in B. stearothermophilus, to
compare, see Ref. 12), interacts with the bulged region of helix 21 (positions 629-633) on the 16 S RNA (Fig. 4A). Thus, our
data provide further evidence for a close proximity of helixes 11 and
21 in the 16 S RNA (9).
For the S8 complex, the sequence of the cross-linked
oligoribonucleotide moiety could only be determined to be
C3Gp, C2U1Gp, C1U2Gp, or U3Gp with either C or U
directly 5 to the 3
Gp as the actual cross-linked nucleotide. By
comparison of our data with those derived from chemical footprinting
(7, 41), cross-linked studies at the rRNA level (38), and binding
studies with S8 (44, 45), we conclude that either positions 589-592 or
651-654 are the cross-linked oligoribonucleotides in the 16 S RNA to
Lys-55 in S8. Both regions are located within or close to helix 21, and the latter possibility is in agreement with one T1 fragment
found to be cross-linked to the intact S8 protein (38). Therefore, our
data strongly suggest that U-653 is the site that is cross-linked to
Lys-55 in S8. Very recently, the three-dimensional structure of S8 from
B. stearothermophilus has been published (16). The cross-linked amino acid (whereby Lys-55 in E. coli
corresponds to Gln-56 in B. stearothermophilus) is
positioned within a loop structure of the protein (Fig. 4B),
which has been suggested to be part of a more extended RNA binding site
in S8 (17). As with the S17 complex discussed above, the corresponding
cross-linked nucleotide in the S8 complex (U-653) and its adjacent area
belong to a non-base-paired region of the 16 S RNA (Fig.
4B). Obviously, loop structures within ribosomal proteins
together with looped or bulged structures within the rRNA are crucial
for protein-RNA interactions in ribosomes (13, 47) similar to other
protein-RNA complexes (e.g. Ref. 46). Additionally, it may
be that only loop regions have sufficient flexibility to allow the
bridging reaction by the cross-linking reagent to occur. The
interaction of discrete loop regions in both proteins S8 and S17 with
helix 21 is consistent with neutron scattering experiments which reveal a very close proximity of these proteins in the 30 S subunit (3). The
data we present here should be very useful in combination with the
three-dimensional structures of ribosomal proteins S8 and S17 and the
current models of the 16 S RNA (9) and the 30 S ribosomal subunit (20,
21) for modeling a highly resolved S8·S17·16 S RNA complex.
We are grateful to Dr. Steven W. White for
providing the S8 pdb-file. We also thank Helga Neubauer and Monika
Owald for their valuable technical assistance.