This article was downloaded by: [New York University] On: 07 January 2015, At: 21:16 Publisher: Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK
Journal of Biomolecular Structure and Dynamics Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/tbsd20
Analysis of an RNA Pseudoknot Structure by CD Spectroscopy a
Kenneth H. Johnson & Donald M. Gray
a
a
The Program in Molecular and Cell Biology Mail Stop FO 31 , The University of Texas at Dallas , Box 830688, Richardson , TX , 75083-0688 Published online: 21 May 2012.
To cite this article: Kenneth H. Johnson & Donald M. Gray (1992) Analysis of an RNA Pseudoknot Structure by CD Spectroscopy, Journal of Biomolecular Structure and Dynamics, 9:4, 733-745, DOI: 10.1080/07391102.1992.10507952 To link to this article: http://dx.doi.org/10.1080/07391102.1992.10507952
PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is
Downloaded by [New York University] at 21:16 07 January 2015
expressly forbidden. Terms & Conditions of access and use can be found at http://www.tandfonline.com/page/terms-and-conditions
Journal of Biomolecular Structure & Dynamics, ISSN 0739-1102 Volume 9, Issue Number 4 (1992), "'Adenine Press (1992).
Analysis of an RNA Pseudoknot Structure by CD Spectroscopy
Downloaded by [New York University] at 21:16 07 January 2015
Kenneth H. Johnson* and Donald M. Gray The Program in Molecular and Cell Biology Mail Stop FO 31 The University of Texas at Dallas Box 830688 Richardson, TX 75083-0688 Abstract The RNA PK5 (GCGAUUUCUGACCGCUUUUUUGUCAG) forms a pseudoknotted structure at low temperatures and a hairpin containing an A· C opposition at higher temperatures (J. Mol. Bioi. 214,455-470 ( 1990)). CD and absorption spectra ofPK5 were measured at several temperatures. A basis set of spectra were fit to the spectra ofPK5 using a method that can provide estimates of the numbers of A· U, G · C, and G · U base pairs as well as the number of each of 11 nearest-neighbor base pairs in an RNA(Biopolymers 31,373-384 (1991)). The fits were close, indicating that PK5 retained the A conformation in the pseudoknot structure and that the fitting technique is not hindered by pseudoknots or A· C oppositions. The results from the analysis were consistent with the pseudoknotted structure at low temperatures and with the hairpin structure at higher temperatures. We concluded that the method of spectral analysis should be useful for determining the secondary structures of other RNAs containing pseudo knots and A· C oppositions.
Introduction In 1982, Rietveld et al. (I) proposed that the 3' terminal sequence of turnip yellow mosaic virus RNA contained a pseudoknot. Since then, pseudoknots have been found in the RNA components of ribonucleoproteins (2), in ribosomal RNAs (3), and in messenger RNAs (4). A pseudo knot in the regulatory site ofbacteriophage T4 gene 32 messenger RNA has been associated with translational control (5). Pseudoknots can be a major part of the structure of an RNA For instance, a 204 nucleotide stretch of the 3' noncoding region of tobacco mosaic virus RNA contains five pseudoknots (6). To our knowledge, the CD spectrum of an RNA pseudoknot has not been reported. A new spectral analysis method allows an estimation of the number of base pairs and nearest-neighbor base pairs from the CD and absorption spectra of an RNA(7). The method fits a basis library ofCD and absorption spectra to the CD and absorption • To whom correspondence should be addressed. Present address: Baylor College of Medicine, The Center for Biotechnology, 4000 Research Forest Drive, The Woodlands, TX 77381.
733
Johnson and Gray
734
Downloaded by [New York University] at 21:16 07 January 2015
A
B
1fP
uu
.u
c.a 5
u
a.c
u
C·G C•A A•U G-U U U
G U C A25 a 3.
ca
c u
c
G
u
u
u U
11! U21 C·G A·U G-C U·A25 C•G 3.
u tP u
A G
c as·
Figure 1: Models for the secondary structure ofPKS (9. 10): (A) the pseudoknot model, (B) the 5' -hairpin model, and (C) the 3'-hairpin model.
735
RNA Pseudoknot
Downloaded by [New York University] at 21:16 07 January 2015
spectra of the unknown. The library contains 58 CD and 58 absorption spectra of simple RNAs. The library does not contain the spectra of a pseudo knot or an A· C opposition. Pseudoknots and A · C oppositions could have CD and absorption bands that would not be approximated by the spectra in our library. Bands that are not approximated by the library spectra could hinder our analysis. The method was previously evaluated by analyzing spectra of E. coli 5S RNA (8). The spectra of 5S RNA were closely fit and the analysis produced reasonable numbers of base pairs and nearest-neighbor base pairs. A more complete evaluation of our method is now provided by the analysis of the RNA PK5, which contains either a pseudo knot or an A · C opposition depending on the solution conditions. The detailed structure of the 26 nucleotide synthetic RNA PKS (GCGAUUUCUGACCGCUUUUUUGUCAG) has been examined using NMR spectroscopy (9, 10). PK5 exhibits the pseudo knot conformation in buffers containing Mg ++and at temperatures below 12 oc in buffers without Mg++.The pseudoknot form ofPK5 contains two base paired stems that partially stack to form a continuous A-form helix (Figure lA). On one side ofthe helix, C 12 and C 13 are stacked; however no evidence was found for stacking of G3 on G22. Above 12 oc in buffers not containing Mg ++, PKS has a hairpin loop in the 5' end of the molecule (Figure lB). The 5'-hairpin contains an A· C opposition and a G · U base pair. A possible structure for PKS that was not detected by NMR is the 3'-hairpin (Figure lC).
Materials and Methods PK5 was a gift from Dr. Jacqueline R. Wyatt and Dr. Ignacio Tinoco Jr. (Department of Chemistry, University of California at Berkeley). PKS was dissolved in 50 mM NaCl04 , 8 mM Na +(phosphate), pH 7.0. Before optical measurements were made, the solution containing the RNA was heated to 70 o C. For measurements in the presence ofMg++, concentrated MgS04 was added to give a concentration of 5 mM. The e260 ofPK5 at 20 o C in the presence ofMg ++was determined by hydrolysis (7) to be 8480 L · mol- 1 · cm- 1• CD and absorption spectra at increasing temperatures were taken as before (7). The end effects in both CD and absorption spectra of PKS were corrected as for the oligonucleotides in the reference library (7) using Equation [1]. In this equation, COP n-mer is the corrected optical property (per mol of monomer) of the oligomer, OP n-mer is the optical property of the oligomer, OPRp and OPsp are the optical properties of the terminal mononucleoside monophosphates, and n is the length of the oligomer in nucleotides. COPn-mer
=
{n*OPn-mer -(OPRp)/2 -(OPsp)/2}/(n-l)
[l]
The corrected CD and absorption spectra were fit using the computer program described in (7). Since the sequence ofPKS does not contain the nearest neighbors A-A, G-G, or U-A, the spectra of single-stranded RNAs containing these nearest neighbors were deleted from the basis set before the analysis was begun. The deleted spectra were those of the dinucleotides ApA, GpG, and UpA each at 0 and 40 °C, as well as
736
Johnson and Gray
0
Downloaded by [New York University] at 21:16 07 January 2015
0 ,_.,
0
CIO N ~
8 d
0 CD
N
'-"
..d ~
bO
d
0
~
N
Q) ...... Q)
> «S
a= 0 N N
~
I
RNA Pseudoknot
737
Figure 2: CD spectra ofPK5 atO oC(o), 30 oc (V),and 70 oc (x), and the fits to these spectra(-). Spectra ofPK5 were taken in 50 mM NaCI04 , 8 mM Na+ (phosphate) pH 7, and were corrected for end effects using Equation [I].
Downloaded by [New York University] at 21:16 07 January 2015
the spectra ofpoly[r(A)] at 20 and 80 oc and ofpoly[r(A-U)] at 50 and 70 oc. Thus, the initial basis set contained 48 of the possible 58 spectra. The 48 spectra included spectra of ApU at 0 and 40 oc as well as spectra of double-stranded RNAs containing the A-A, G-G, and U-A nearest neighbors. The fitting procedure was also changed such thatthe weights on the fractions of A-A, G-G, and U-A were increased to 2000 (Table II of Ref. 7). Secondary-structure models were interpreted in terms of the numbers of nearestneighbor base pairs, single-stranded nucleotides, and nucleotides in undefined structures as before (8). Since the spectra of PK5 were corrected for end effects by removing half of the spectra ofthe end nucleotides (Equation [1 )), nucleotides at the ends of the molecule were counted differently from those at the end of 5S RNA in previous work (8). Instead of counting a double-stranded nucleotide at the end of a molecule as part of a nearest-neighbor base pair plus 0.5 single-stranded nucleotide, it was counted only as part of a nearest-neighbor base pair. A single-stranded nucleotide at the end of the molecule was counted as 0.5 of a single-stranded nucleotide instead of a whole single-stranded nucleotide. Therefore, for the 26 nucleotide PK5 the sum of: ( 1) two times the number of nearest-neighbor base pairs, (2) the nucleotides in single strands, and (3) the nucleotides defined as being in undefined structure = 26 - 1 = 25. Interpreting pseudoknot junctions, such as the G3-A4 and U21-G22 junction in Figure 1A., was not specifically dealt with earlier. For the pseudo knot conformation of PK5, Puglisi et al. (9) did not find evidence for the stacking of G3 on G22. Therefore, we treated the junction as the ends of two helices. The magnitude of the normalized deviations (MND) was used to determine how close a model was to our results (8). The MND was calculated as:
[2] where PVi were the values predicted by the models, MVi were our estimates, E. were the expected errors in our estimates, and the sum was over the 15 categories {i}. The 15 categories were: {1-10} the 10 Watson-Crick nearest-neighbor base pairs, {11} (G,U) · (G,U) nearest-neighbor base pairs (nearest-neighbor base pairs containing G · U base pairs, as explained in Ref. 7), and {12-15} each of the fournucleotides in single strands. The expected errors were taken from previous work (7).
Results and Discussion Spectra of PK5
CD and absorption spectra ofPK5 at 0, 30, and 70 oc in the absence ofMg ++,after correction for end effects using Equation [1 ), are shown in Figures 2 and 3. The spectra
738
Johnson and Gray
0 0
Downloaded by [New York University] at 21:16 07 January 2015
I"')
.,.........
0
CQ
N
e ..._., ~
..c::: 0
+J
N
~
co
bD
OJ _... QJ
ttS ....0N >
•
0
N N
g
~~_.--~~_.--~~~--L-~~~ N oN
- -
739
RNA Pseudoknot Figure 3: Absorption spectra of PK5 at 0 spectra (-).
oc (o), 30 oc (V),
and 70
oc
(x), and the fits to these
at 0 oc were similar to the spectra ofPKS in the presence ofMg++between 0 and 40 oc (not shown). The magnitudes of the 210 and 268 nm CD bands at 0 and 30 oc (no Mg++) (Figure 2) were similar to those for mixed sequence double-stranded RNAs (11 ), but were larger than those for mixed sequence double-stranded DNAs (12). This indicated that PKS had predominantly the A conformation at 0 and 30 a c. At 70 °C, the CD bands were dramatically smaller, and the absorption at 260 nm was 12% hyperchromic between 30 and 70 °C, indicating that the secondary structure had melted.
Downloaded by [New York University] at 21:16 07 January 2015
Fits to the Spectra of PK5
CD and absorption spectra ofPKS a tO, 5, 10, 12, 15, 20, 25, 30, 35, 40, 50, 60, and 70 oc (no Mg ++)were fit using our method of analysis (7). The fits were all to within 2.5 times the error of the measurement. The fits were not as close as those to the spectra of E. coli 5S RNA(8). However, they were as close as the fits to the spectra of various polymers whose nearest-neighbor base pair contents were well predicted by our method (7). The closeness of the fits showed that none of the conformations displayed by PK5 had large CD or absorption bands that were not common to the simple RNAs in the basis set. IfPK5 was not predominantly in the A conformation, the spectra would be expected to be poorly fit by the spectra of the basis set of RNAs. Thus, the constraints of the pseudo knot or 5'-hairpin conformations do not alter the RNA greatly from the A conformation. Results from the Fits
Figure 4 shows the numbers of A· U, G · U, and G · C base pairs derived from the fits to the spectra ofPK5 at different temperatures. There were two transitions, one from 0 to 25 °C, and one from 35 to 70 a c. A G · U base pair was formed during the first transition and melted in the second transition. A· U and G · C base pairs were lost during both transitions. Wyatt et al. (10) found a transition centered at 53 oc and a broad transition with low hyperchromicity centered at 12 oc for PK5 in 50 mM NaC104, 0.5 mM EDTA.10 mM sodium phosphate, pH 6.4. In our buffer, both transitions were very broad. Table I contains the numbers of base pairs, nearest-neighbor base pairs, and singlestranded nucleotides given by our analysis at 0, 30, and 70 °C, and by three models for PK5. Figure 5 shows the MNDs (Equation 2) for each model at each temperature. Models with low MNDs fit our data better than models with high MNDs. Structure at 0°C
From 0 to 12 a c. the pseudoknot had a lower MND than did the other models. At oac (Table I) our results gave one more A· U base pair than was predicted by the pseudoknot model. The extra A · U base pair was represented partly as 0.5 of an (AU) · (A-U) nearest-neighbor base pair, conceivably because of the stacking of A4 and US, which has been shown by NMR spectroscopy (9).
740
Johnson and Gray
Downloaded by [New York University] at 21:16 07 January 2015
--------~----------~------i
0 CD ~
t..)
..._,
0
Q)
l-4
0
•
= «
Journal of Biomolecular Structure and Dynamics Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/tbsd20
Analysis of an RNA Pseudoknot Structure by CD Spectroscopy a
Kenneth H. Johnson & Donald M. Gray
a
a
The Program in Molecular and Cell Biology Mail Stop FO 31 , The University of Texas at Dallas , Box 830688, Richardson , TX , 75083-0688 Published online: 21 May 2012.
To cite this article: Kenneth H. Johnson & Donald M. Gray (1992) Analysis of an RNA Pseudoknot Structure by CD Spectroscopy, Journal of Biomolecular Structure and Dynamics, 9:4, 733-745, DOI: 10.1080/07391102.1992.10507952 To link to this article: http://dx.doi.org/10.1080/07391102.1992.10507952
PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is
Downloaded by [New York University] at 21:16 07 January 2015
expressly forbidden. Terms & Conditions of access and use can be found at http://www.tandfonline.com/page/terms-and-conditions
Journal of Biomolecular Structure & Dynamics, ISSN 0739-1102 Volume 9, Issue Number 4 (1992), "'Adenine Press (1992).
Analysis of an RNA Pseudoknot Structure by CD Spectroscopy
Downloaded by [New York University] at 21:16 07 January 2015
Kenneth H. Johnson* and Donald M. Gray The Program in Molecular and Cell Biology Mail Stop FO 31 The University of Texas at Dallas Box 830688 Richardson, TX 75083-0688 Abstract The RNA PK5 (GCGAUUUCUGACCGCUUUUUUGUCAG) forms a pseudoknotted structure at low temperatures and a hairpin containing an A· C opposition at higher temperatures (J. Mol. Bioi. 214,455-470 ( 1990)). CD and absorption spectra ofPK5 were measured at several temperatures. A basis set of spectra were fit to the spectra ofPK5 using a method that can provide estimates of the numbers of A· U, G · C, and G · U base pairs as well as the number of each of 11 nearest-neighbor base pairs in an RNA(Biopolymers 31,373-384 (1991)). The fits were close, indicating that PK5 retained the A conformation in the pseudoknot structure and that the fitting technique is not hindered by pseudoknots or A· C oppositions. The results from the analysis were consistent with the pseudoknotted structure at low temperatures and with the hairpin structure at higher temperatures. We concluded that the method of spectral analysis should be useful for determining the secondary structures of other RNAs containing pseudo knots and A· C oppositions.
Introduction In 1982, Rietveld et al. (I) proposed that the 3' terminal sequence of turnip yellow mosaic virus RNA contained a pseudoknot. Since then, pseudoknots have been found in the RNA components of ribonucleoproteins (2), in ribosomal RNAs (3), and in messenger RNAs (4). A pseudo knot in the regulatory site ofbacteriophage T4 gene 32 messenger RNA has been associated with translational control (5). Pseudoknots can be a major part of the structure of an RNA For instance, a 204 nucleotide stretch of the 3' noncoding region of tobacco mosaic virus RNA contains five pseudoknots (6). To our knowledge, the CD spectrum of an RNA pseudoknot has not been reported. A new spectral analysis method allows an estimation of the number of base pairs and nearest-neighbor base pairs from the CD and absorption spectra of an RNA(7). The method fits a basis library ofCD and absorption spectra to the CD and absorption • To whom correspondence should be addressed. Present address: Baylor College of Medicine, The Center for Biotechnology, 4000 Research Forest Drive, The Woodlands, TX 77381.
733
Johnson and Gray
734
Downloaded by [New York University] at 21:16 07 January 2015
A
B
1fP
uu
.u
c.a 5
u
a.c
u
C·G C•A A•U G-U U U
G U C A25 a 3.
ca
c u
c
G
u
u
u U
11! U21 C·G A·U G-C U·A25 C•G 3.
u tP u
A G
c as·
Figure 1: Models for the secondary structure ofPKS (9. 10): (A) the pseudoknot model, (B) the 5' -hairpin model, and (C) the 3'-hairpin model.
735
RNA Pseudoknot
Downloaded by [New York University] at 21:16 07 January 2015
spectra of the unknown. The library contains 58 CD and 58 absorption spectra of simple RNAs. The library does not contain the spectra of a pseudo knot or an A· C opposition. Pseudoknots and A · C oppositions could have CD and absorption bands that would not be approximated by the spectra in our library. Bands that are not approximated by the library spectra could hinder our analysis. The method was previously evaluated by analyzing spectra of E. coli 5S RNA (8). The spectra of 5S RNA were closely fit and the analysis produced reasonable numbers of base pairs and nearest-neighbor base pairs. A more complete evaluation of our method is now provided by the analysis of the RNA PK5, which contains either a pseudo knot or an A · C opposition depending on the solution conditions. The detailed structure of the 26 nucleotide synthetic RNA PKS (GCGAUUUCUGACCGCUUUUUUGUCAG) has been examined using NMR spectroscopy (9, 10). PK5 exhibits the pseudo knot conformation in buffers containing Mg ++and at temperatures below 12 oc in buffers without Mg++.The pseudoknot form ofPK5 contains two base paired stems that partially stack to form a continuous A-form helix (Figure lA). On one side ofthe helix, C 12 and C 13 are stacked; however no evidence was found for stacking of G3 on G22. Above 12 oc in buffers not containing Mg ++, PKS has a hairpin loop in the 5' end of the molecule (Figure lB). The 5'-hairpin contains an A· C opposition and a G · U base pair. A possible structure for PKS that was not detected by NMR is the 3'-hairpin (Figure lC).
Materials and Methods PK5 was a gift from Dr. Jacqueline R. Wyatt and Dr. Ignacio Tinoco Jr. (Department of Chemistry, University of California at Berkeley). PKS was dissolved in 50 mM NaCl04 , 8 mM Na +(phosphate), pH 7.0. Before optical measurements were made, the solution containing the RNA was heated to 70 o C. For measurements in the presence ofMg++, concentrated MgS04 was added to give a concentration of 5 mM. The e260 ofPK5 at 20 o C in the presence ofMg ++was determined by hydrolysis (7) to be 8480 L · mol- 1 · cm- 1• CD and absorption spectra at increasing temperatures were taken as before (7). The end effects in both CD and absorption spectra of PKS were corrected as for the oligonucleotides in the reference library (7) using Equation [1]. In this equation, COP n-mer is the corrected optical property (per mol of monomer) of the oligomer, OP n-mer is the optical property of the oligomer, OPRp and OPsp are the optical properties of the terminal mononucleoside monophosphates, and n is the length of the oligomer in nucleotides. COPn-mer
=
{n*OPn-mer -(OPRp)/2 -(OPsp)/2}/(n-l)
[l]
The corrected CD and absorption spectra were fit using the computer program described in (7). Since the sequence ofPKS does not contain the nearest neighbors A-A, G-G, or U-A, the spectra of single-stranded RNAs containing these nearest neighbors were deleted from the basis set before the analysis was begun. The deleted spectra were those of the dinucleotides ApA, GpG, and UpA each at 0 and 40 °C, as well as
736
Johnson and Gray
0
Downloaded by [New York University] at 21:16 07 January 2015
0 ,_.,
0
CIO N ~
8 d
0 CD
N
'-"
..d ~
bO
d
0
~
N
Q) ...... Q)
> «S
a= 0 N N
~
I
RNA Pseudoknot
737
Figure 2: CD spectra ofPK5 atO oC(o), 30 oc (V),and 70 oc (x), and the fits to these spectra(-). Spectra ofPK5 were taken in 50 mM NaCI04 , 8 mM Na+ (phosphate) pH 7, and were corrected for end effects using Equation [I].
Downloaded by [New York University] at 21:16 07 January 2015
the spectra ofpoly[r(A)] at 20 and 80 oc and ofpoly[r(A-U)] at 50 and 70 oc. Thus, the initial basis set contained 48 of the possible 58 spectra. The 48 spectra included spectra of ApU at 0 and 40 oc as well as spectra of double-stranded RNAs containing the A-A, G-G, and U-A nearest neighbors. The fitting procedure was also changed such thatthe weights on the fractions of A-A, G-G, and U-A were increased to 2000 (Table II of Ref. 7). Secondary-structure models were interpreted in terms of the numbers of nearestneighbor base pairs, single-stranded nucleotides, and nucleotides in undefined structures as before (8). Since the spectra of PK5 were corrected for end effects by removing half of the spectra ofthe end nucleotides (Equation [1 )), nucleotides at the ends of the molecule were counted differently from those at the end of 5S RNA in previous work (8). Instead of counting a double-stranded nucleotide at the end of a molecule as part of a nearest-neighbor base pair plus 0.5 single-stranded nucleotide, it was counted only as part of a nearest-neighbor base pair. A single-stranded nucleotide at the end of the molecule was counted as 0.5 of a single-stranded nucleotide instead of a whole single-stranded nucleotide. Therefore, for the 26 nucleotide PK5 the sum of: ( 1) two times the number of nearest-neighbor base pairs, (2) the nucleotides in single strands, and (3) the nucleotides defined as being in undefined structure = 26 - 1 = 25. Interpreting pseudoknot junctions, such as the G3-A4 and U21-G22 junction in Figure 1A., was not specifically dealt with earlier. For the pseudo knot conformation of PK5, Puglisi et al. (9) did not find evidence for the stacking of G3 on G22. Therefore, we treated the junction as the ends of two helices. The magnitude of the normalized deviations (MND) was used to determine how close a model was to our results (8). The MND was calculated as:
[2] where PVi were the values predicted by the models, MVi were our estimates, E. were the expected errors in our estimates, and the sum was over the 15 categories {i}. The 15 categories were: {1-10} the 10 Watson-Crick nearest-neighbor base pairs, {11} (G,U) · (G,U) nearest-neighbor base pairs (nearest-neighbor base pairs containing G · U base pairs, as explained in Ref. 7), and {12-15} each of the fournucleotides in single strands. The expected errors were taken from previous work (7).
Results and Discussion Spectra of PK5
CD and absorption spectra ofPK5 at 0, 30, and 70 oc in the absence ofMg ++,after correction for end effects using Equation [1 ), are shown in Figures 2 and 3. The spectra
738
Johnson and Gray
0 0
Downloaded by [New York University] at 21:16 07 January 2015
I"')
.,.........
0
CQ
N
e ..._., ~
..c::: 0
+J
N
~
co
bD
OJ _... QJ
ttS ....0N >
•
0
N N
g
~~_.--~~_.--~~~--L-~~~ N oN
- -
739
RNA Pseudoknot Figure 3: Absorption spectra of PK5 at 0 spectra (-).
oc (o), 30 oc (V),
and 70
oc
(x), and the fits to these
at 0 oc were similar to the spectra ofPKS in the presence ofMg++between 0 and 40 oc (not shown). The magnitudes of the 210 and 268 nm CD bands at 0 and 30 oc (no Mg++) (Figure 2) were similar to those for mixed sequence double-stranded RNAs (11 ), but were larger than those for mixed sequence double-stranded DNAs (12). This indicated that PKS had predominantly the A conformation at 0 and 30 a c. At 70 °C, the CD bands were dramatically smaller, and the absorption at 260 nm was 12% hyperchromic between 30 and 70 °C, indicating that the secondary structure had melted.
Downloaded by [New York University] at 21:16 07 January 2015
Fits to the Spectra of PK5
CD and absorption spectra ofPKS a tO, 5, 10, 12, 15, 20, 25, 30, 35, 40, 50, 60, and 70 oc (no Mg ++)were fit using our method of analysis (7). The fits were all to within 2.5 times the error of the measurement. The fits were not as close as those to the spectra of E. coli 5S RNA(8). However, they were as close as the fits to the spectra of various polymers whose nearest-neighbor base pair contents were well predicted by our method (7). The closeness of the fits showed that none of the conformations displayed by PK5 had large CD or absorption bands that were not common to the simple RNAs in the basis set. IfPK5 was not predominantly in the A conformation, the spectra would be expected to be poorly fit by the spectra of the basis set of RNAs. Thus, the constraints of the pseudo knot or 5'-hairpin conformations do not alter the RNA greatly from the A conformation. Results from the Fits
Figure 4 shows the numbers of A· U, G · U, and G · C base pairs derived from the fits to the spectra ofPK5 at different temperatures. There were two transitions, one from 0 to 25 °C, and one from 35 to 70 a c. A G · U base pair was formed during the first transition and melted in the second transition. A· U and G · C base pairs were lost during both transitions. Wyatt et al. (10) found a transition centered at 53 oc and a broad transition with low hyperchromicity centered at 12 oc for PK5 in 50 mM NaC104, 0.5 mM EDTA.10 mM sodium phosphate, pH 6.4. In our buffer, both transitions were very broad. Table I contains the numbers of base pairs, nearest-neighbor base pairs, and singlestranded nucleotides given by our analysis at 0, 30, and 70 °C, and by three models for PK5. Figure 5 shows the MNDs (Equation 2) for each model at each temperature. Models with low MNDs fit our data better than models with high MNDs. Structure at 0°C
From 0 to 12 a c. the pseudoknot had a lower MND than did the other models. At oac (Table I) our results gave one more A· U base pair than was predicted by the pseudoknot model. The extra A · U base pair was represented partly as 0.5 of an (AU) · (A-U) nearest-neighbor base pair, conceivably because of the stacking of A4 and US, which has been shown by NMR spectroscopy (9).
740
Johnson and Gray
Downloaded by [New York University] at 21:16 07 January 2015
--------~----------~------i
0 CD ~
t..)
..._,
0
Q)
l-4
0
•
= «
