INFORMATION TO USERS
This reproduction was made from a copy of a document sent to us for microfilming. While the most advanced technology has been used to photograph and reproduce this document, the quality of the reproduction is heavily dependent upon the quality of the material submitted.
The following explanation of techniques is provided to help clarify markings or notations which may appear on this reproduction.
1.The sign or “target” for pages apparently lacking from the document photographed is “ Missing Page(s)” . I f it was possible to obtain the missing page(s) or section, they are spliced into the film along with adjacent pages. This may have necessitated cutting through an image and duplicating adjacent pages to assure complete continuity.
2. When an image on the film is obliterated with a round black mark, it is an indication of either blurred copy because of movement during exposure, duplicate copy, or copyrighted materials that should not have been filmed. For blurred pages, a good image of the page can be found in the adjacent frame. I f copyrighted materials were deleted, a target note will appear listing the pages in the adjacent frame.
3. When a map, drawing or chart, etc., is part of the material being photographed, a definite method of “sectioning” the material has been followed. It is customary to begin filming at the upper left hand comer o f a large sheet and to continue from left to right in equal sections with small overlaps. If necessary, sectioning is continued again-beginning below the first row and continuing on until complete.
4. For illustrations that cannot be satisfactorily reproduced by xerographic means, photographic prints can be purchased at additional cost and inserted into your xerographic copy. These prints are available upon request from the Dissertations Customer Services Department.
5. Some pages in any document may have indistinct print. In all cases the best available copy has been filmed.
University Micixxilms International 300 N. Zeeb Road Ann Arbor, Ml 48106 8526208
Li, Shi-Jiang
BIOPHYSICAL INVESTIGATIONS INTO THE SECONDARY STRUCTURE OF TRITICUM AESTIVUM 5S RIBOSOMAL RNA
The Ohio State University Ph.D. 1985
University Microfilms International 300 N. Zeeb Road, Ann Arbor, Ml 48106 PLEASE NOTE:
In all cases this material has been filmed in the best possible way from the available copy. Problems encountered with this document have been identified here with a check mark V .
1. Glossy photographs or pages______
2. Colored illustrations, paper or print ______
3. Photographs with dark background______
4. Illustrations are poor c o p y______
5. Pages with black marks, not original copy_____
6. Print shows through as there is text on both sides of page ______
7. Indistinct, broken or small print on several pages ^
8. Print exceeds margin requirements ______
9. Tightly bound copy with print lost in spine ______
10. Computer printout pages with indistinct print______
11. P ag e(s)______lacking when material received, and not available from school or author.
12. P ag e(s)______seem to be missing in numbering only as text follows.
13. Two pages numbered ______. Text follows.
14. Curling and wrinkled pages ______
15. Other ______
University Microfilms International BIOPHYSICAL INVESTIGATIONS INTO THE SECONDARY STRUCTURE
OF TRITICUM AESTIVUM 5S RIBOSOMAL RNA
DISSERTATION
Presented in Partial Fulfillm ent of the Requirements for
the Degree Doctor of Philosophy in the
Graduate School of the Ohio State University
By
SHI-JIANG LI
*******
The Ohio State University
1985
Reading Committee: Approved By
Professor Alan G. Marshall
Professor Gary E. Means
Professor Robert T. Ross I k A H i s / J . I l l Aavisor Professor Elizabeth L. Gross Department of Biochemistry TO MY FAMILY
1i ACKNOWLEDGMENTS
There are a very large number of people whom I would lik e to
thank. First, I would like to express my appreciation to Professor
Alan G. Marshall for his guidance, inspiration and style throughout
this work.
I would also like to thank C. E. Cottrell for his guidance in NMR
program. Special thanks to T. L. Ricca, and L.-H. Chang for th eir DSC program, J. 0. Alben and K. 0. Burkey for acquiring FT-IR spectra.
All the members of Professor Marshall’ s research group, including T.
C. L. Wang, S.-M. Chen, F. S. Andersen, and the rest of the troupe are acknowledged for their assistance and fruitful discussions.
I would lik e to extend my appreciation to Professor Gary E.
Means, Professor Edward J. Behrman, M.-K. Chang, and J. Houck for their encouragement and help.
A special gratitude is expressed to my family. Without their support this work would not have been possible.
Finally, I wish to acknowledge the U.S.A. Public Health Service
(N .I.H . 1 R01 GM-29274; N .I.H . 1 S10 RR-01458) and the Department of
Biochemistry, The Ohio State University. VITA
October 31, 1946 ...... Born, Wuxi, Jiangsu, China
1970 ...... Graduated from Tsing Hua University, Beijing, China
1970-1978 ...... Assistant Engineer at National 610th Factory, China
1978-1980 ...... M. Sci. Program, The Graduate School, Academia Sinica, Beijing, China
1981-1985 ...... Teaching and Research Assistant, at The Ohio State University
PUBLICATIONS
Zhang, Z.-L. and Li, S.-L. (1982) "Infrared Study on Hydration of Bovine Serum Albumin Membranes" Prog. Biophys. & Biochem. 33-36 (in Chinese).
Li, S.-J., Chang, L.-H., Chen, S.-M., and Marshall, A. G. (1984) "Preparative-Scale Isolation and Purification of Procaryotic and Eukaryotic Ribosomal 5S RNA: Bacillus subtil i s , Neurospora crassa, and Wheat Germ," Anal. Biochem.~T3 8 , 465-471.
Chang, L.-H., Li, S.-J., Ricca, T. L., and Marshall, A. G. (1984) "Theoretical and Experimental On-Line Analysis of Multi-State Melting of Polymers by Differential Scanning Calorimetry," Anal . Chem. 56, 1502-1507.
Li, S.-J., Burkey, K. 0., Luoma, G. A., Alben, J. 0., and Marshall, A.G., (1984) "Base Pairing in Wheat Germ Ribosomal 5S RNA as Measured by Ultraviolet Absorption, Circular Dichroism, and Fourier-Transfonm Infrared Spectrometry," Biochemistry 23, 3652-3658. ~ iv Li, S.-J. and Marshall, A. G. (1985) "Multi-State Unfolding of Wheat Germ Ribosomal 5S RNA Analyzed by D ifferen tial Scanning Calorimetry," Biochemistry 24, in press.
Li, S.-J. and Marshall, A. G. "Identification and Assignment of Base Pairs in the "Tuned Helix" of Intact and RNase T1 Cleavage Fragments of Wheat Germ Ribosomal 5S RNA via 500 Mhz Proton Homonuclear Overhauser Enhancements" Submitted to Biochemistry, July, 1985.
Field of Study: Biophysical Chemistry
v TABLE OF CONTENTS
Page
DEDICATION ...... ii
ACKNOWLEDGMENTS ...... H i
VITA ...... iv
LIST OF TABLES ...... xi
LIST OF FIGURES ...... x ii
BIOPHYSICAL INVESTIGATIONS INTO THE SECONDARY STRUCTURE
OF TRITICUM AESTIVUM 5S RIBOSOMAL RNA
page
CHAPTER I INTRODUCTION ...... 1
A. The general events of protein synthesis in the ribosome ...... 1
B. Ribosomal 5S RNA involvement in protein synthesis ...... 10
1. Possible function of 5S RNA in the ribosomes ...... 10
2. Localization of 5S RNA in ribosomes ...... 15
C. Secondary structural models for 5S RNA and the methods to approach i t ...... 17
1. Comparative sequence analysis ...... 19
2. Chemical modifications ...... 21
vi 3. Oligonucleotide binding ...... 23
4. Ribonuclease digestion ...... 24
5. Optical techniques ...... 25
6. Magnetic resonance methods ...... 27
7. Thermal stability of 5S RNA structure and differential scanning calorimetry ...... 29
8. Other techniques ...... 31
D. Wheat germ cytosol 5S RNA secondary structure 33
CHAPTER I I ISOLATION, PURIFICATION, AND PREPARATION OF NATIVE 5S RNA AND ITS RNASE T1 FRAGMENTS FROM WHEAT GERM 37
A. Introduction ...... 37
B. Isolation and Purification ...... 39
1. Materials and Methods ...... 39
2. Phenol-SDS extraction ...... 39
3. Ion-exchange chromatography ( DE-32) ...... 41
4. G e l-filtra tio n chromatography (G-75) ...... 43
C. Purity of the wheat germ 5S RNA ...... 49
1. Gel electrophoresis ...... 49
2. Gel scanning ...... 54
D. Concentration determination ...... 55
E. 5S RNA fragments by RNase T1 cleavage ...... 58
1. Fragment A (base 1-86, 87 and/or 89) ...... 58
2. Fragment (base 1-25, 57-86, 90-120) and B2 (base 26-56) ...... 62
F. Discussion ...... 66
1. The conditions for enzymatic cleavage of WG 5S RNA by RNase T1 ...... 66
2. The homogeneity of the fragments ...... 67
vi i CHAPTER I I I BASE PAIRING IN WHEAT GERM 5S RNA AS MEASURED BY FOURIER-TRANSFORM INFRARED SPECTROSCOPY ...... 68
A. Introduction ...... 68
B. Theory ...... 71
1. Interferometer ...... 71
2. Determination of number and types of base pairs ...... 76
C. Materials and methods...... 78
1. Preparation of 5S RNA for FT-IR spectroscopy ...... 78
2. FT-IR spectroscopy ...... 80
D. Results and discussion...... 85
1. FT-IR spectra ...... 85
2. Determination of RNA base-pair composition from FT-IR absorbance ...... 85
3. Discussion ...... 94
CHAPTER IV THEORETICAL AND EXPERIMENTAL ON-LINE ANALYSIS OF MULTI-STATE MELTING OF WG 5S RNA BY DIFFERENTIAL SCANNING CALORIMETRY ...... 101
A. Introduction ...... 101
B. Theory ...... 102
1. Temperature-dependence of a two-state equilibrium ...... 102
2. Relation between d9/dT and the DSC experim ent ...... 103
3. simulation of DSC melting curve ...... 105
C. Using model compounds to test the two-state model ...... 107
1. Microcal ori m e te r ...... 107
v iii 2. experimental measurements by DSC technique 110
D. DSC analysis of WG 5S RNA ...... 117
1. Materials and Methods ...... 119
2. Results ...... 122
3. Discussion ...... 131
CHAPTER V IDENTIFICATION AND ASSIGNMENTS OF BASE PAIRS IN 5S RNA SECONDARY STRUCTURE FROM WHEAT GERM BY ^H 500 MHZ FT-NMR ...... 144
A. Introduction ...... 144
B. Water suppression techniques ...... 145
1. Dynamic range problem ...... 146
2. Water suppression techniques ...... 147
C. Nuclear Overhauser Effect (NOE) and its application to the nucleic acid research ...... 151
1. The general discription of NOE ...... 151
2. Application of NOE to the nucleic acid research ...... 157
D. Materials and Methods ...... 158
1. NMR sample preparation ...... 158
2. NMR spectroscopy ...... 159
E. Results and Discussion...... 160
1. Number of base pairs in intact wheat germ 5S RNA ...... 160
2. Resolution of overlapping peaks via change in salt concentration ...... 163
3. Peak id entification (A*U, G*C, G*U) from NOE pattern ...... 165
4. Base-paired segments determined by NOE connectivity ...... 168
ix 5. Temperature-induced shifts and melting as an aid in base pair sequencing ...... 169
6. Use of RNase T1 cleavage fragments for assignment of base paired segments ...... 174
7. Ring current-induced chemical s h ift confirm assignment of Ci8G60-A19u59-c20G58 ...... 177
8. Are the conformations of the fragments the same as in the intact molecule ...... 179
9. 5S RNA secondary structure ...... 183
10. Limitations of the NOE method, and implications for future work ...... 185
CHAPTER VI SUMMARY ...... 188
REFERENCES ...... 190
APPENDIX ...... 203
x LIST OF TABLES
PART I I
page
Table 1. Comparison of WG 5S RNA concentration determination between [P i] method and UV method ...... 57
Table 2. Number of secondary base pairs in WG 5S RNA and yeast tRNAphe ...... 95
Table 3. Melting midpoints (Tm) (°C) of WG 5S RNA ...... 97
Table 4. Melting midpoints (Tm) and molar enthalpy of melting (£H) were obtained from theoretical best fits of Eq. 4.11 to experimental DSC curves for dilute aqueous protein solutions...... 116
Table 5. Thermodynamic parameters for multi-stage melting Of WG 5S RNA...... 128
Table 6. Reproducibility and reversibility of DSC scans for WG 5S RNA in the presence of 0.1M NaCl with no added MgCl2...... 130
Table 7. Effect of DSC scan rate upon Tm for component melting transitions of WG 5S RNA in the presence of 0.1M NaCl with no added MgCl 2...... 133
Table 8. The base-pairing energies ...... 142
Table 9. Total enthalpy and free energy of melting for wheat germ 5S RNA ...... 142
Table 10. Comparison of experimental 1 h NMR chemical shifts with those computed from ring-current for a proposed base pair sequence ...... 178 LIST OF FIGURES
page
Figure 1. Procaryotic ribosomes can be dissociated into about f if ty -f iv e proteins and three RNA molecules ...... 2
Figure 2. Three-dimentional model of the ribosome gives an asymmetric shape to its two subunits ...... 4
Figure 3. In itia tio n stage of protein synthesis ...... 6
Figure 4. Elongation stage of protein synthesis ...... 8
Figure 5. Termination stage which ends the synthesis of a protein and release i t from the ribosome ...... 9
Figure 6. Proposed secondary and te rtia ry structure of Ji. coli 5S RNA and the binding regions with proteins and RNAs. 12
Figure 7. Localization of the 3' endsof 5S and 23S RNA ...... 16
Figure 8. The similarities and differences of most commonly proposed secondary structure of 5S RNA...... 18
Figure 9. The physical approaches to the secondary structure of 5S RNA...... 20
Figure 10. Alignment of the determined plant 5S RNA sequences. 22
Figure 11. Ideal representation of the three processes observable via differential scanning calorimetry ...... 32
Figure 12. Cloverleaf model of wheat germ 5S RNA ...... 34
Figure 13. Fox and Woese model of wheat germ 5S RNA ...... 35
Figure 14. Nishikawa and Takimura model of wheat germ 5S RNA. .. 36
Figure 15. Flow diagram for isolation of crude RNA from dried wheat germ ...... 40
xii Figures (continued)
Figure 16. Flow diagram for purification of crude 5S RNA from phenol-SDS extraction ...... 42
Figure 17. A DE-32 anion exchange elution p rofile for the phenol-SDS extraction of wheat germ ...... 44
Figure 18. Separation of large and small molecules using gel permeation chromatography ...... 46
Figure 19. Elution p rofile from Sephadex G-75 gel permeation Chromatography ...... 48
Figure 20. Compounds used in the synthesis of acrylamide gels. 50
Figure 21. Polyacrylamide slab gel electrophoresis profiles. ... 51
Figure 22. Gel scans of electrophoresis ...... 59
Figure 23. The elution profile from the denaturing column used to separate the RNase T1 fragment A ...... 63
Figure 24. Profiles for the non-denaturing gel filtration column and the non-denaturing electrophoresis ...... 65
Figure 25. Optical system of a typical interferometric spectrometer ...... 72
Figure 26. A typical interferogram from Michel son interferometer 73
Figure 27. Diagram of FT-IR sample preparation ...... 79
Figure 28. The measurement of cell path length is corrected by the fringe method ...... 81
Figure 29. Generation of FT-IR absorbance spectrum for wheat germ 5S RNA...... 83
Figure 30. Determination of infrared extinction coefficient versus frequency for A*U base pairs ...... 86
Figure 31. Determination of infrared extinction coefficient versus frequency for G-C base pairs ...... 87
Figure 32. FT-IR reference spectra of the four ribonucleotide 5 '-monophosphates...... 88
Figure 33. FT-IR spectra of WG 5S RNA in the absence of Mg++. .. 89 Figures (continued)
Figure 34. FT-IR spectra of wheat germ 5S RNA in the presence of Mg++...... 91
Figure 35. Nuber of A*U and G-C base pair of wheat germ 5S RNA as fuction of temperature ...... 96
Figure 36. Block diagram of differential heat capacity calorimeter ...... 108
Figure 37. DSC curves for cytochome c ...... 113
Figure 38. DSC curves for ribonuclease A ...... 115
Figure 39. DSC curves for equimolar mixture of cytochrome c and ribonuclease A ...... 118
Figure 40. Illustration of the DSC systematic response time. ... 121
Figure 41. Differential scanning calorimetry plots for wheat germ 5S RNA...... 123
Figure 42. The data of Figure 41 are re-plotted after 7-point smoothing and linear baseline correction over the transition region...... 125
Figure 43. Reversibility of DSC experiment for wheat germ 5S RNA in the presence of 0.1M NaCl and no added Mg++ ...... 129
Figure 44. DSC power vs. scan temperature, for three scan rate: 96.5 K/hr (top), 67.4 K/hr (middle), and 48.9 K/hr (bottom)...... 132
Figure 45. Free energies of the three proposed secondary structural models of WG 5S RNA...... 143
Figure 46. The representation of the three water suppression techniques: long pulse, Redfield 214 pulse , and 1331 hard pulse.
Figure 47. Effect of various excitation pulse sequences on NMR spectra of wheat germ 5S RNA ...... 153
Figure 48. Plot of the homonuclear NOE in a two-spin system. . . . 156
Figure 49. 500 MHz low fie ld NMR spectrum of WG 5S RNA ...... 161
Figure 50. Use of salt-induced shifts to resolve overlapping resonances in the downfield *H NMR spectrum of intact wheat germ 5S RNA ...... 164
xi v Figures (continued)
Figure 51. Identification of base pair types from intact wheat germ 5S RNA via proton homonuclear overhauser enhancement difference spectra ...... 166
Figure 52. Ambiguity of NOE connectivity when overlapping resonances occur...... 170
Figure 53. Use of temperature-induced melting and shifts to identify and confirm base pair sequence in intact wheat germ 5S RNA ...... 172
Figure 54. Proposed 500 MHz *H NMR spectra (le ft) and proposed secondary structures (right) of intact wheat germ 5S RNA (top) and three purified fragments produced by RNase T1 cleavage ...... 175
Figure 55. Presaturation of peak C on the RNase Tl-resistant fragment shown in the middle of Figure 54 gave the same NOE pattern (bottom spectrum) as for intact molecule (compare spectrum D in Figure 5 1 ) ...... 180
Figure 56. The same NOE pattern is produced by irrad iation of the A*U resonance at peak A, whether from intact 5S RNA (top), the RNase Tl-resistant fragment (middle), or the common arm fragment (bottom)...... 181
Figure 57. The same NOE pattern is produced by irred iation of the A*U resonance at peak B, whether from intact WG 5S RNA or in the fragment ...... 182
xv CHAPTER I
INTRODUCTION
A. The general events of protein synthesis in the ribosome
The biosynthesis of proteins is a complex process that takes place in the ribosome. Extensive studies during the last three decades have led to a formal description of this process (1 -9 ). A ribosome is a highly specialized and complex structure, approximately
200 A in diameter. The best-characterized ribosomes are those of E. c o li. The E. coli ribosome has a mass of 2,500 kdal and sedimentation coefficient of 70S. It can be dissociated into a large subunit (50S) and a small subunit (30S) (Figure 1) (10). These subunits can be further split into their constituent proteins and RNA's. The 30S subunit contains twenty-one proteins and a 16S RNA molecule, whereas the 50S subunit contains about thirty-four proteins and two RNA molecules, a 23S species, and a 5S species. About two-third of the mass of E. coli ribosome is RNA, whereas the other third is protein.
Eukaryotic ribosomes are appreciably larger than prokaryotic ones, they contain a grater number of proteins, about 80 rather than
55, and they have a extra molecule of RNA (5.8S RNA). An intact eukaryotic ribosome has a sedimentation coefficeint of 80S. Like a 2
< # p i f - 23S RNA 5S RNA
^ l " < » About 34 proteins T 50S subunit
70S- ribosome " \ V ^ ^ ^ 16S RNA
30S subunit About 21 proteins
Figure 1. Prokaryotic ribosomes can be dissociated into about fifty-five proteins and three RNA molecules (10). bacterial ribosome, i t dissociates into a small subunit (40S) containing an 18S RNA instead of the 16S RNA that is in the bacterial ribosome, and a large subunit (60S) containing 28S RNA (or 26S) instead of 23S RNA, as well as a 5S and 5.8S RNA.
Using smal 1-angle x-ray scattering diffraction, and various electron microscopy methods (12, 13), the overall shape and the gross structure of ribosomes and th e ir subunits have been established by many observers. In spite of the differences between the various methods (14) there is a general agreement that the small subunit is a prolate asymmetric particle as seen in Figure 2. It consists of a
"head" comprising about one third, and a "base" comprising the other two thirds of the particle, also it has a cleft between the two parts as well as one platform; the large subunit includes a central protuberance flanked by a ridge on one side and a stalk on the other.
When the large subunit and the small subunit are bound together, the stalk of the large subunit has its base near the constriction on the small subunit, and the head of the small subunit and the central protuberance of the large subunit are approximately aligned. Such a machinery provides functional significance in protein synthesis.
The synthesis of a protein proceeds in three distinct stages.
The first stage is initiation (15), the second is called elongation
(16) and the final stage is called termination (17,18) which is not yet as well understood as the previous two stages above.
Protein synthesis starts with the association of mRNA, a 30S ribosomal subunit, and formylmethionyl-tRNA^ to form a 30S initiation 4
CENTRAL PROTUBERANCE \
PlATTOflH
+
lAPCcsoeuNn MBOSOMC SMALL SUBUNIT CENTRAL 'PROTUBERANCE
MM.LEV STAIN
Figure 2. Three-dimensional model of the ribosome gives an asymmetric shape to its two subunits. The small subunit includes a "head", a " base" and a platform. The large subunit includes a central protuberance, flanked by a ridge on one side and a stalk on the other. Two orientations of the model are shown (14). complex (Figure 3 ). The formation of this complex requires 6TP and three initiation factors IF-1, IF-2 and IF-3. A 50S ribosomal subunit then joins a 30S in itia tio n complex to form a 70S in itia tio n complex; the bound GTP is hydrolyzed in this step. The result of in itia tio n is a complex consisting of a large ribosomal subunit, a small ribosomal subunit, the mRNA and the fMet-tRNA. This complex is now ready for the elongation stage of protein synthesis.
The elongation cycle in protein synthesis consists of three steps: f ir s t , the binding of ami noacyl-tRNA which is called codon recognition; second peptide bond formation; finally, a translocation process.
Each iteration of the cycle requires the participation of molecules called elongation factors (EF's) and also two molecules of
GTP. At the beginning of the iteration a peptidyl-tRNA (a tRNA bearing the nascent chain) is bound to the ribosome at what is called the P s ite . An ami noacyl-tRNA binds to the R s ite , binding in complex with EF-Tu and GTP. This binding is controlled by the matching of the codon on the mRNA to three bases (anticodon) on the tRNA. The second step is now performed and a peptide bond is formed. The tRNA is transferred to the A s ite , where its amino acid accepts the nascent chain catali zed by peptidyl transferase to form a peptide-bond. At this step, an uncharged tRNA occupies the P s ite , whereas a peptidyl-tRNA occupies the A site (the existence of distinct P and A sites was inferred from studies of puromycin, an antibiotic which is an analog of the terminal aminoacyl-adenosine portion of ami noacyl-tRNA. 6
m R NA
SMALL SUBUNIT
IRNA
(MET LARGE SUBUNIT
Figure 3. Initiation stage of protein synthesis: formation of initiation complex consisting of a large ribosomal subunit, a small ribosomal subunit, the mRNA, and the fMet-tRNA (14). 7
The final step of the elongation cycle is translocation. Three
movements occur; f ir s t the uncharged tRNA leaves the P s ite , and the
peptidyl-tRNA moves from the A site to the P site while the mRNA moves
a distance of three nucleotides. Translocation requires a third
elongation factor EF-G (also called translocase), and the GTP bound to
EF-G is hydrolyzed during translocation. After translocation, the R
and A sites are empty, ready to bind an ami noacyl-tRNA to begin
another round of elongation.
Termination of protein synthesis occurs when any of the codons
UAA, UGA, or UAG are reached: these three are called stop signals.
These stop signals are recognized by two release factors protein, RF1
and RF2, instead of ami noacyl-tRNA which does not normally bind to
these stop signals. The binding of a release factor to a termination
codon in the A site activates peptidyl transferase so that it
hydrolyzes the bond between the peptide and the tRNA in the P s ite .
The peptide chain then leaves the ribosome, and the 70S ribosomal unit then dissociates into the 30S and 50S subunits, ready to start the synthesis of another protein molecule (Figure 5).
As far as the synthesis of protein in eukaryotic is concerned, a much more complex mechanism may be involved and fa r less is known about this matter in contrast to the one in prokaryotic ribosome (11).
The details for the differences may be limited to only those which are more closed related to the function and structure of ribosomal 5S RNA, and are discussed la tte r. 8
Figure 4. Elongation stage of protein synthesis: including binding of aminoacyl-tRNA, peptide-bond formation, translocation, and reaction cycle of elongation factor Tu (14). 9
mRNA SMALL SUBUNIT
GTP
TERMINATION TERMINATION FACTORS FACTORS
PROTEIN LARGE SUBUNIT
Figure 5. Termination stage which ends the synthesis of a protein and releases i t from the ribosome (14). 10
B. Ribosomal 5S RNA involvement in protein synthesis.
The principles of protein synthesis, as it occurs in the ribosome are now understood as described above; a su fficiently detailed molecular level process for those numerous components has been investigated. One component, namely ribosomal 5S RNA , has received intensive investigation (19). So far, 238 5S RNAs from prokaryotic, archebacteria, organelles and eukaryotic sources have been sequenced
(20). This small ribonucleic acid first came to attention in 1963
(21), from £ . coli 50S ribosomal subunits, and i t was soon determined that its sequence (22) did not contain modified nucleotides, such as were found in tRNA (23). It is not surprising that 5S RNA plays an important role in the ribosome during protein synthesis.
1. Possible functions of 5S RNA in the ribosome.
Attempts to elucidate the precise function of 5S RNA in the ribosome have led to the following observations.
a. 5S RNA is essential for ribosomal activity. This was demonstrated by total reconstitution experiments using £ . stearothermophilus (24, 25, 26) and IE. coli (27) 50S ribosomal subunits. In the absence of 5S RNA, reconstituted ribosomes lack the a b ility to bind non-enzymatically and enzymatically to ami noacyl-tRNA and thus can not synthesize proteins.
b. The 5S RNA binding proteins were identified from IB. stearothermophilus (28) and £. coli (28, 29) by reconstituting 11
specific 5S RNA-protein complexes from 5S RNA and 50S ribosomal
protein conponents and then verified by partial reconstitution
experiments. The results are that the E^. coli 5S RNA binding proteins
are E-L5, E-L18, and E-L25, the jl. stearothermophil us 5S RNA binding
proteins are B-15 (corresponding to E-L5), and B-L22 (corresponding to
E-L18). The binding regions of these proteins to coli 5S RNA are
shown in Figure 6A (30, 31).
c. The formation of a complex between 5S and 18S RNA has raised
the possibility that their intermolecular interaction could play an
important role in joining the two ribosomal subunits during protein
synthesis (32, 33). The hypothetical interaction of E_. coli 5S RNA with 16S RNA, 23S RNA, and tRNA is shown in Figure 6B (34).
d. Reconstitution experiments have shown that different
prokaryotic 5S RNA's can be incorporated in jl. stearothermophil us 50S
ribosomal subunits to yield biologically active particles, while
eukaryotic 5S RNA cannot (26). Thus, the structure-function relation must be universal, at least for prokaryotes. On the other hand, many mitochondrial ribosomes seem to lack a 5S RNA or its equivalent
(35). This would imply either that the role of 5S RNA is not a
universal one or perhaps that i t is taken over by some other ribosomal
component in these mitochondria.
e. Eukaryotic 5S RNA does not recognize E. coli 5S RNA binding
proteins E-L5, E-L18, and E-25. Instead, eukaryotic 5.8S RNA, free
in solution, is recognized by the binding to E-L18 and E-L25. I t is
therefore suggested that eukaryotic 5.8S RNA, and not eukaryotic 5S
RNA, is similar in function to prokaryotic 5S RNA (36). 12
•»o , Ed enea HE B DE] I BG3 e ra DD EE “OBJ ® e b « ® B E II ? °®H000BBE)B©o8 ® o°B B E 0 EraSE)®® O © ” ® “ o ®o W * BE) ESQ ®SKD» I I I ®DD III mQJO 16 S ED rRNA (RNA
Figure 6. (A), Proposed secondary and tertia ry structure of 5S RNA, and the binding regions of protein L5, L25, and L18 to the molecule. (B), Hypothetical interaction of E.. coli 5S RNA with 16S RNA, 23s RNA, and tRNA (58). 13
f. GTPase and ATPase activities have been found in association with specific 5S RNA-protein complexes isolated from J3. stearothermophilus (37,38), coli (39), and also in rat liver tissue (40). Since the antibiotics that in h ib it the hydrolytic a c tiv itie s of the 5S RNA-protein complexes also in h ib it the ribosome-and elongation factor-dependent GTPase (41), these results suggest the possible involvment of 5S RNA proteins in this function.
On the other hand, ATPases so far detected with eukaryotic protein synthesis are involved in the initiation phase. It is interesting that eukaryotic 5S RNA-protein complexes exhibit ATP hydrolytic activity and that eukaryotic 5S RNA appears to be involved in chain initiation; in particular, eukaryotic 5S RNA plays a role by interacting with initiator tRNA (43,44).
g. A feature about 5S RNA function formerly widely accepted, has recently been challenged by Pace et al. (45, 46); namely, that a conserved 5S RNA segment complementary to tRNA is not required for protein synthesis and translation of natural mRNA. It had been proposed that 5S RNA is directly involved in tRNA binding, namely that the conserved 5S RNA sequence C 43GAAC47 interacts with the conserved tRNA sequence GT CG (loop 4) (47-49). Indeed, the sequence T CG inhibits non-enzymatic and enzymatic binding of ami noacyl tRNA to ribosome (48,50) and i t prevents "magic spot" (pppGpp and ppGpp) formation. However, Pace et a l . (45) demonstrated that the deletion of GAAC from 5S RNA did not affect its a b ility to carry out poly-U-directed synthesis of polyphenylalanine. Furthermore, the deletion of the sequence CCGAA 46, and the deletion of the sequence 14
CCGAACUCAGA52 were tested via in vitro phage MS 2 RNA-directed protein
synthesis, ami noacyl-tRNA binding, 70S in itia tio n complex formation,
and guanosine tetraphosphate synthesis. The unexpected result was
that ribosomes reconstituted from 5S RNA containing the internal
deletions were active in all of the tested functions. It is clear
from these results (46) that CCGAA46 sequence of 5S RNA appears to be
dispensable for accurate translation of natural and synthetic mRNA.
h. The different conformational forms of 5S RNA was found in KB
c e ll, Hela cell and rat liv e r by electrophoretic and chromatographic
methods. Two for KB (51) and Hela cells (52), and three for rat
liver (53) distinct bands representing the native 5S RNA (A form) and
thermodynamically stable denaturated forms (B form) were observed.
The conformational forms can be interconverted by heat or urea
treatment (51-55). NMR analysis (56) has suggested the existence of
an "H" form (fo r high temperature and high Mg++ ion concentration) and
an "L" form (fo r low temperature and absence of Mg++ ions). A
possible functional significance for the conformational conversion is
that a switching mechanism between the forms may provide the necessary movement for translocational events in the ribosome (57).
i. The amount of phylogenetically conserved nucleotides in
single stranded regions is higher than in double stranded regions in
sequence comparative studies; also, the number of adenines in those conserved single-stranded regions is significantly higher than that of
the other nucleotides (58). In addition such structural features as
single- or double-base bulge nucleotides, multiple G'U's, and internal loops also are present in the 5S RNA structure. The possible roles of 15 these irregularities are at present an open question. It is not unreasonable to conclude that these structural features are important for the recognition of, and for interaction with, other ribosomal components or molecules involved in protein biosynthesis (59)
2. Localization of 5S RNA in ribosomes
5S RNA is an integral part of the large subunit of the ribosomes.
Its position in the ribosomes was detected by using the immunoelectron microscopy approach (60, 61). The 3' end of the 5S RNA was localized on the outward surface of the central protuberance of the 50S subunit.
Correspondingly, the 3' end of the 23S RNA was below the stalk on the noninterfacing surface (see Figure 7). It should be mentioned that the mapping of the 3 '-end of the 5S RNA on the 50S subunit simultaneously specifies the localization of its 5 ’-end. Indeed, i t has been proved by direct cross-linking experiments that the complementary terminal sequences 1-10 and 110-119 i n ^ . coli 5S RNA form a double-helical stem (62). Thus, the central protuberance of the 50S subunit is the site of location of the 3 ', and 5'-terminal stem of the 5S RNA. This morphological part of the SOS subunit is the universal and very characteristic feature of both the prokaryotic and eukaryotic ribosomes (63). Figure 7. Localization of the 3' ends of 5S and 23S RNA. (a). Two-dimensional localization of the antibody binding site on crown-shaped ( le f t ) and kidney-shaped (right) profiles. (b). Three-dimensional localization of the 3' ends of 5S and 23S RNA on the 50S model proposed by Tischendorf et a l . (64, 65). 17
C. Secondary structure models for 5S RNA and the methods to approach
i t.
Eighteen years have elapsed since the f ir s t 5S RNA sequence was published ( 22), and since that time the collection is approaching
238 sequences (20). Although a general shape for the secondary
structure is emerging (66-78), there seems to be l i t t l e agreement on the exact base pairing scheme in several regions of the molecule.
Figure 8 shows the volution of various models with their differences and similarities.
In contrast to 5S RNA, the cloverleaf model of tRNA was one of the three possible base pairing schemes proposed in the very first paper reporting the complete tRNA sequence (79). I t was quickly and widely accepted as the correct choice. Nine years la te r, the cloverleaf mode was then confirmed by x-ray crystal lography (80).
The confusion over the correct secondary structure of 5S RNA cannot be explained merely by the fact that 5S RNA is 1.5 times as large as a tRNA. The difficulty in exactly determining its structure may come from two reasons. F irs t, the structural features of 5S RNA are much more complex then those of tRNA: bulges and in te rio r loops are rarely found in tRNA. Second, 5S RNA is part of the ribosome; thus, its structure may reflect, in addition to secondary and tertiary interactions, the necessity of forming quaternary bonds with other ribosomal constituents. Hence the structure of the free molecule may be different from the functional form, this in turn may cause different results under different circumstances. For example, »' J‘ 120
HO Cloverleaf
100 4 0. JO •o' 70
(20 Common 20'
Oslerberg, 120 Nishikowa , , Sjoberg, i^^o 5 j & Takem ura i,_,i 2o Fox a Woese 40 . 30. 20.
Studnicko el. al. 1
Figure 8 . The sim ilarities and differences of most commonly proposed secondary structure of 5S RNA. 19 conformational rearrangement of 5S RNA was observed when i t binds to ribosomal proteins (81-85). In spite of these d iffic u ltie s , a wide variety of experimental approaches has been employed in the study of
5S RNA structure. Until the crystallization and subsequent x-ray diffractio n analysis can be achieved ( 86 ), i t is probably fa ir to say that these experimental methods alone are not sufficient to deduce a correct secondary structure but are extremely important tests of the structural features. Figure 9 shows a breakdown of various studies and where they apply. The information obtained to date is based upon studies that can be grouped into seven aspects which are listed below.
1. Comparative sequence analysis.
The systematic sequence analysis is a very powerful approach to the study of 5S RNA secondary structure. This study begins with the assumption that the molecules under study have essentially the same secondary, structure. The primary structures are then aligned, and there must be a compromise between obtaining a maximum amount of sequence homology and introducing the smallest possible number of insertions and deletions (87, 88 ). I f the base changes between the compared sequences are noted, and compensating base changes maintain the complementarity between two potential pairing regions, then this is taken as evidence for the existence of a true helix at that position. Figure 10 shows the alignment of the determined plant 5S
RNA sequences (89) . A disadvantage of this method is that success in this approach is crucially dependent upon the precision of the 20
(*) Flucrouracil ( F NMR)
u.v. absorption i.r. absorption oligonucleotide bind Raman scattering calorimetry (all detect base- stacked regions)
31P NMR (detects sharp bends)
enzymatic cleavage
’ H NMR (one or two peaks for each base-pair) chemical modification; EPR spin-label —" cross-linkers (detects local motion)
Figure 9. The experimental approaches to the secondary structure of 5S RNA 21 sequence alignment; if there is uncertainty at this stage, any conclusions about secondary structure w il1 be equivocal. Another disadvantage is that the penalty for introducing insertion and deletion has to be chosen a rb itra rily . In other words, an assumption has to be made on the relative importance in evolution of base substitutions and insertions or deletions, while this ratio is unknown. Nevertheless, since the systematic sequence comparison method was f ir s t employed to propose a secondary structure model of 5S
RNA (68 ), many authors have used this approach to further study 5S RNA secondary structure models (73-76). This study has reported two major contributions to the nature of 5S RNA. First, a generalized structure of 5S RNA was proposed and a deeper insight into its secondary structure organization and classification has been gained although some aspects of the specific base positions and base pairing have given controversy. Second, comparative analysis of these 5S RNA primary structures has been used for the construction of phylogenetic trees (90), with the conclusion that the 5S RNA molecule possesses evolutionary significance.
2. Chemical modifications.
Base-specific chemical modification studies have been used in order to discriminate between free nucleotides and those which are involved in secondary or te rtia ry interaction. Depending on the particular groups of the nucleotides involved, every reagent exhibits different stereospecifity. Among the more useful single strand-specific chemical probes are glyoxal and kethoxal 22
C C 1 »
urn *to V * • a u a cnuu I I I LTcartmcw iscunmji II * ft tfllMlMA M*Ufi VI* hMa&US VUMIS V * * via* f*»* U ft ft *J010S*C?U VKlMIlt
B’ E’ D*
l. *!■» C 1 VtCIC • C IV 1 V V I c • * c c c V t. a««.f c( l l t u V 11 IC S * KIM V I S V * C V * 1 s ij* 1 s I v s * CCVCCSIS I * * 1 VtCK V C I V IVVIC-IVVC c i. (scatni* ct t f t c v V 11 S C I* KIM 1 * • V * C lIllU l II SSI* c erne c v s i S I * c ». M U c(«((» V 11 SCI* KIM V * f « * cviisLv l« SIS* c c(cje c v s s 1*1 • MCMKII s«vsc>**ct c m tu ts c IVSCV V1t SCI* KIM 1 III* e v i l t «|v IS IM S * c c v c c v s s IS* s r e x v c s u I V V I C * » C V V 1 v. ru* cd i d VV « c *. V U C U M I ( f 1 * € 1 € * k t » C t-H **S p
Figure 10. Alignment of the determined plant 5S RNA sequences. The number at the top applies to the 6 plant 5S RNA sequences, The sequence of A. yinelandii 5S RNA is for comparison. Boxes labled A-A"1", B- 6 ', etc"., enclose areas involved in base pairing and resulting in helices A, B, etc., in the secondary structure model. Bulges are indicated by nested boxes, non-standard base pairs by. characters within parentheses. Lower-case characters at the 3'-end of some sequences denote residues present in submolar amounts (89). 23
(G-specific) (91-93), diethyl pyrocarbonate (A-and G-specific) (94), bisulfite (C-and U-specific) (95), dimethylsulfate (C-specific as a single strand probe), m-chloroperbenzoate and monoperphthalate (96)
(both A-specific). As far as secondary and tertiary structure is concerned, dimethylsulfate senses the secondary structure (monitors the N-3 of C) (94) and tertiary structure (monitors the N-7 of G)
(94), and diethylpyrocarbonate has been shown to detect stacking of adenines via its N-7 (94). Also, psoralen derivatives ami nomethyl trioxsalen (AMT) and hydroxymethyltrimethylpsoralen (HMT) have been used as double strand-specific probes (97-102). Finally, psoralens intercalate between stacked base pairs in nucleic acids and undergo photochemically induced cycloaddition to pyrimidines on opposite strands (103). Significant information has been obtained to establish the structural features of £ . coli 5S RNA (58) through use of these techniques.
3. Oligonucleotide binding.
The binding of oligonucleotides is a very sensitive method for detection of single-stranded stacked regions in nucleic acids. I t depends upon the formation of base pairs between the oligonucleotide and a complementary sequence in the macromolecules (104, 105).
Particular sequences in the prokaryotic (105,49) and eukaryotic (106)
5S RNA have been found using this method. The results from these studies did not fit with data from other sources very well (other sources such as chemical modification and enzymatic hydrolysis). The possible reason may come from the different accessibilities. 24
Oligonucleotide binding not only requires a single-strandedness, but
also good stacking configuration. Another problem may result from
differen t experimental conditions. A major drawback is from the
observation that oligonucleotide-binding of 6T CG to the conserved 5S
RNA sequence CGAAC (44) does not mean that a tRNA-5S RNA interaction
in this region is required for biosythesis of protein (45, 46).
4. Ribonuclease digestion.
Together with chemical modification and oligonucleotide binding,
ribonuclease digestion provides another technique to investigate 5S
RNA structural features. A number of ribonucleases have been used as structural probes, including RNase T1 (108), RNase T2 (109), SI nuclease (110) (a ll single-strand specific) and cobra venom RNase
(111-116) (double-strand specific). A number of single-stranded regions and double helix regions have been localized, particularly in
£. coli 5S RNA, generating a secondary and tertiary structure.
However, a potential danger in the use of nucleases as structure probes has to be taken into account; namely, th at cleavage of the sugar-phosphate chain at one site may lead to significant unfolding, or other rearrangement of the RNA structure. Thus, i t is important to distinguish between primary and secondary cleavage sites, and to rely only on primary cleavage sites as indicators of structure. Now, however, an elegant method utilizing alternate 5 ‘ and 3' end-labeling has been achieved to overcome such ambiguities described above.
However, absence of enzymatic cleavage of a suspected single-stranded 25 or double-stranded region does not imply that no such structure exists. This may due to the fact that such regions are internally buried within the tertiary structure and thereby inaccessible to the relatively large enzymes, or are involved in some interactions to reduce the reactivity in such cleavages. Emphasis on enzymatic cleavage has increased and has been used to obtain specific 5S RNA fragments, from which much more structural information has emerged from other physical studies (117-119).
5. Optical techniques.
Optical spectroscopy measurements (ultraviolet, Fourier transform infrared and Raman spectroscopy) give the approximate total number of base pairs and the relative and absolute numbers of G*C and A*U pairs. From UV measurements (120, 121), A26Q hyperchromism is used to calculate the upper-limit base pair number in 5S RNA, i f one uses a value of 0.30 for the hyperchromism expected from maximal double-stranded stacking (i.e ., maximum base pairing) and a 10% minimum contribution to hyperchromism from single-stranded stacking
(122). Also, A28 O hyperchromism is informative because hyperchromism at 280 nm is due almost entirely to the unstacking of G*C pairs
(123). Further, since the individual normalized hyperchromism spectra for poly (rG)-poly (rC) and poly (rA)*poly (rU) are markedly different in the 260-290 nM range (124), linear combinations of the two cases may be used to estimate the relative proportions of G*C and A*U pair in 5S RNA (125). For infrared measurement, previous attempts to determine the number of G*C and A*U base pairs in 5S RNA were based on dispersive IR spectroscopy (126). Because the infrared method requires accurate measurement of absorbance (O .D .lO .l) in the presence of strong absorption by the D2O solvent (and ultra-strong absorption by HDO present in varying amounts in d ifferen t samples), single scan dispersive IR spectra are incapable of producing meaningful results.
The replacement of the dispersive IR spectrometer by a Fourier transform instrument and the improved data reduction method have increased the precision of the base pair estimates (127, 128). Much more reasonable base pair numbers were obtained by fittin g the RNA difference spectrum for unpaired minus paired bases with the reference difference spectra between paired and unpaired poly (rA)-poly (rU) and poly (rG)*poly(rC) (127). The FT-IR spectra as a function of temperature can also be used to monitor independently the melting of
G*C and A*U pairs (128). Finally, as far as the Raman technique is concerned, particular peaks such as G-stacking (670 cm-1 ), A-stacking
(725 cm_l ) , U-pairing (1660 and 1688 cm"*), and A-helix content (814 cm"l) are used to estimate the numbers of A*U and G*C pairs by comparing 5S RNA with tRNA, whose base-stacking is known (69-70, 129).
Above a ll, optical measurements provide a unique technique to evaluate the upper limits for possible base-stacking in 5S RNA. A big disadvantage, however, is lack of the specific location of their base pairs. Further, since the measurements require comparison with model compounds, such as synthetic polynucleotides, and a certain flexibility of interpretation, it is not surprising that there are significant variations among the results (120, 126, 127). 27
6 . Magnetic resonance methods. Nuclear magnetic resonance
success in obtaining useful proton NMR spectra of RNAs in 2 HO
solution requires the highest available fie ld (to increase signal to noise ratio and disperse the overlapping peaks) and optimal suppression of the H2O solvent resonance. The simplest water-nulling method in general use for such systems is the "21412"
Redfield sequence (130). This "magic" sequence effectively overcomes the "broad skirt" problem of the soft "long pulse" (131, 132) A variety of water nulling techniques have been developed in recent years (131-136), particularly the 1331 pulse sequence, which have simplified the operating procedure of the Redfield method (130). With these techniques much better spectra were obtained (117-119) compared to those in the earlier periods (137, 138, 139). Fortunately, the secondary and tertiary structure of 5S RNA involves the extremely low fie ld resonance region of proton NMR spectra, which is well separated from the rest of the protons (aromatic and ribose protons) and from the relatively cluttered remainder of the spectrum (140, 141). These low fie ld resonances represent hydrogen bonded imino protons which are from guanosines (Nx) and uridines (N3 ). Since each A-U and 6 *C pair contributes one resonance to this region, and each G*U contributes two, the imino proton spectra can be used to evaluate the numbers of base pairs in the molecule, although the exchange rate with solvent may affect the result, and this method only gives a lower lim it to the total base pair number. A breakthrough technique was the introduction of the Nuclear Overhauser effect (NOE) to RNA research (142). The 28 nuclear Overhauser effect can be viewed as a transfer of saturation from one proton to another proton nearby in space. The magnitude of transfer is inversely proportional to the sixth power of the distanceO between the two protons and becoms too small to detect beyond 4A in most situations (143, 144). When a hydrogen-bonded imino proton of a base pair is presaturated, the NOE's seen are generally limited to the next-nearest neighbor protons. These w ill include the resonances of base pairs immediately above and below the base pair whose proton is presaturated. These inter-base NOEs give the connectivity of a base-pair sequence in the 5S RNA secondary structure (119, 142).
Further, presaturation of an A-U base pair imino proton gives a sharp strong NOE in the aromatic region in addition to the NOEs in the downfield region to its nearest neighbors (145). The proton seen is either a C 2 proton (Watson-Crick) or a Cg proton (reverse Hoogsteen) on the A of the A*U pair. This upfield NOE distinguishes an A*U downfield resonance from a G*C downfield resonance. G*U wobble base resonances can be used as a starting point to elucidate the base sequence connectivity because their strong in tra base NOE can be recognized easily (146-148). The successful application of NOE in tRNA has been applied to 5S RNA research with informative results
(117-119).
Another technique is electron spin resonance (ESR). ESR spin probes have proved extremely sensitive to structural differences within an RNA molecule at a given temperature, as well as to temperature-induced structural changes (148). ESR spin probes have successfully demonstrated a conformational change in the anticodon 29 following ami noacyl ation of tRNA (149). For 5S RNA research, the specific labeling to the 3 -terminus of the 5S RNA was done, since the
3'-terminal ribose of each RNA is the only sugar with a 2', 3'-diol functionality (the 3'-OH is esterfied to a phosphate in all the other
sugars) (150). Analysis was made separately for the internal motion at the site of attachment of the label and the more interesting gross motion of the RNA backbone by the measurements of the effective rotational correlation times. Interesting structural information might also be obtained from the combination of ESR and NMR techniques.
To date several of these experiments are in progress.
7. Thermal stability of 5S RNA structure and differential
scanning calorimetry.
A simple method for estimating the secondary structure of an RNA molecule was proposed on the basis of the knowledge of its primary sequence (149). The secondary structure which an RNA molecule takes up at equilibrium in any solution is that which minimizes the free energy of the molecule. This problem was f ir s t approached by Fresco et a l. (150). Subsequently, Tinoco et a l . (151-153 ) improved the method for the estimation of secondary structure in ribonucleic acids.
In order to state the topological rules that the secondary structure of RNA obeys, i t is useful to consider the following definitions (20).
F irs t, a double-stranded area is an area consisting of one or more helical segments connected by bulges or, more exceptionally, by small interior loops. Second, a helical segment is a stretch of double 30
helix uninterrupted by bulges or in terio r loops. Third, a standard
base pair is G*C, A*U, or G’U. Finally, a non-standard base pair is
one of the 7 other conceivable base pairs. With these definitions,
the topological rules can be summarized as follows. A hairpin loop
cannot contain less than 3 bases; a helical segment should consist of
a t least two standard base pairs, and any non-standard base pair
occurring in such a segment should be intercalated between two
standard base pairs. I t is extremely rare to find more than one
non-standard base pair in any one double stranded area. When this
approach was applied to tRNA sequences, a wide variety of sequences
could a ll be put into a sim ilar cloverleaf pattern with the anticodon
in the same position, thus demonstrating the interchangeability of
tRNA's in protein biosynthesis in the ribosome (154). However, for
the 5S RNA, the lack of knowledge of its function and the large number
of possibilities for intermolecular pairing led to a number of
proposed models with no obviously correct choice ( 68 , 69, 71, 73, 74,
76, 77). Nevertheless, by applying the rules to 5S RNA via a
computer program, we can compare the results for various proposed models (for details see Chapter IV).
Differential scanning calorimetry (DSC). From its inception in
1887 u n till about 18 years ago, DSC has earned the reputation of being
a difficult, time-consuming technique nonrepeatable. More recently,
the differential scanning calorimeter has been improved dramatically
for easier operation, high sensitivity and good repeatability. In
general, differential scanning calorimetry analysis is based on the
measurement of d ifferen tial power (heat input) necessary to keep a 31 sample and a reference substance isothermal as the temperature is changed (scanned) lin e arly. DSC plots are graphs of the differen tial rate of heating (in cal/sec) versus temperature (see Figure 11). The area under the peak is directly proportional to the heat evolved or absorbed by the reaction, and the height of the curve is directly proportional to the rate of reaction or phase transition.
Many research groups have successfully applied DSC technique to analyze the process of denaturation of biomacromolecules (155-158).
Also, the thermodynamic parameters of structure stabilization of 5S
RNA from E. coli have been intensively investigated (159). A variety of methods to analyze the DSC p ro file were used under different circumstances (157, 158, 160). More recently, an algorithm for the deconvolution of calorimetric curves was proposed, based on fundamental thermodynamic relations and on the sequential determination of the parameters of the elementary transitions without any assumption of the paticular model or mechanism of the reaction
(161-163). In spite of these effo rts, an unequivocal structural interpretation of the calorimetric data is impossible without the acquisition of additional information.
8 . Other techniques.
In addition to those techniques mentioned above for RNA structural studies, there are other useful methods available to approach this project. The more frequently used methods include
CD: (Circular Dichroism) (125,164), ORD (Optical Rotatory Dispersion) iue 1 Iel ersnain f h tre rcse osral via observable processes three the of representation Ideal 11. Figure Differential Power, mcal/sec 0 nohri Reaction Endothermic ifrnta sann clrmty (233). calorimetry scanning tial differen xtemc Reaction Exothermic Baseline e perature Tem etCpct hne( P) C Change Capacity (A Heat 32
33
(165, 166), small-angle x-ray scattering (167), electron microscopy
(12, 13), and low-angle neutron scattering (168).
D. Wheat germ cytosol 5S RNA secondary structure.
Wheat germ, (species Triticum Aestivum), a flowering plant, has been obtained extensive attention since the plant its e lf is easily available and is an ideal system for the study of protein synthesis.
I t is therefore logical that the detection of the secondary structure of wheat genm 5S RNA has become a popular project to reveal the universal structure of 5S RNA from eukaryotic organisms. Moreover, since plants and fungi are approximately equally divergent, and considerably more divergent than insects, from the line ascending to vertebrates( 169, 170), the elucidation of wheat germ 5S RNA structure is phylogenical ly significant. Wheat germ 5S RNA was sequenced indirectly quite early (171-173) and was later revised by a technique which combines rapid gel sequencing methods and classical sequence analysis of oligonucleotides generated by ribonuclease T1 (174). From this sequence (174), the three most cited secondary structure models can be adapted to this sequence as shown in Figure 12, 13, 14. In this thesis, a variety of biophysical methods is used to investigate its secondary structure. These techniques include Fourier Transform
Infrared Spectroscopy (FT-IR), Differential Scanning Calorimetry
(DSC), and Fourier Transform Nuclear Magnetic Resonance Spectroscopy
(FT-NMR). 34
5* 3 ' C-OH 120 C PG »C G -U a • u , U • A I G • C OG G* U iA* U •U • G—HO 10-^ c u / au a g r in ^ If U VA C \GC C A # u a ’ III 20—C-G A _ III AAaU#AA—100 ■ • » /a C C c N G IV a /U UAGGC — CA GGUCCU-CCA (G [C •••••• ^ AG A ACUCCGAAGU% CUAGGAyGGg G
5 0 _ _ > G * U r C • G gU*A l 60—GG »U»u A C *G U* A U*G—70 G A G G GC
Figure 12. Cloverleaf model of wheat germ 5S RNA (187). 35
s' y C-OH 120 c PG .C G »U A »U I r . t C »G
-iccv-,1'? i c\ r *• uo.;° iv c;a / m U a GGCC CACGACC UGCUCC GGGUCCU0 f* I u •••• ••••• ••••• •••••• u . A c UCCGa rGUGCUiiGG GCGAG* aC U A G G A ^ g G* •^ agaa ti / aGua< ^ I ^ \ 50 \ 60 60 N
Figure 13. Fox and Woese model of wheat germ 5S RNA (72). 36
S’ 3' C-OH 120 PG .£ G »U A *U U • A G • C C • G n V j o o /.C C c 's so » g Aaa 2 0 Ay U „ u G A a ; U” u•••• a g g c — ••ca C U •••*••*( W a- c{ g •••• c u c c g g •••••••• g u c c u - c c ^ y. ' \c rOCCGAAGUu c GUGCUuGG G— CGAG. ACUAGG A ^ G c " * Figure 14. Nishikawa and Takimura model of wheat germ 5S RNA (71). CHAPTER I I ISOLATION, PURIFICATION, AND PREPARATION OF NATIVE 5S RNA AND ITS RNASE T1 FRAGMENTS FROM WHEAT GERM. A. Intoduction. Ribosomal 5S RNA has been sequenced from 238 biologically distinct species (20). Much of the published work on 5S RNA has involved "chemical" manipulations (e.g., primary nucleotides sequencing, binding to ribosomal proteins, nuclease digestions, ologonucleotide binding) which require only microgram of the molecule. Therefore, most published precedures for isolation and purification of 5S RNA have not been optimized for high yield , and involve procedures not well suited to preparative scale-up (e.g., cutting up an electrophoresis gel to isolate the band of interest (180)). With a few exceptions, most of these experiments have been designed to discover unpaired bases in the polynucleotide sequence. More recently, a variety of “physical" techniques designed to characterize base-paired and base-stacked segments have been applied to 5S RNA. Each of these techniques requires large amounts of sample, so that a better isolation was required: for example, FT-IR (121, 128) and Raman (69, 70) 2-5 mg/sample, l \\ FT-NMR 37 38 12 mg/sample, 19F FT-NMR 10 mg/ml, 31P NMR 50 mg/sample,(117-119,125, 139, 181-183), differential scanning calorimetry (DSC) (159) 20 mg/sample, and crystallizatio n (86) for future X-ray diffractio n analysis (> 100 mg). In lig h t of these facts, a specific procedure for isolating large quantities of electrophoreticaly pure 5S RNA from wheat germ became necessary. This procedure is presented below. Two types of isolation strategies have previously been employed. In the first (181), ribosomes are isolated via ultracentrifugation, with subsequent extraction of ribosomal RNAs, and fin a lly gel-permeation chromatography on Sephadex G-100. In the second approach (180), RNAs are extracted from cells directly with phenol and then resolved by gel electrophoresis; the band containing 5S RNA is then cut for further sequencing. In the f ir s t isolations, 5S RNA from wheat germ was obtained in small quantity from ribosomes (184, 185). Later, phenol extraction followed by electrophoresis gel sectioning was introduced (172) based on the e a rlie r procedure (186). Finally, preparative electrophoresis was replaced by ion-exchange chromatography followed by gel-permeation chromatography (187). The present procedure employes a phenol-SDS (Sodium Dodecyl Sulfate) extraction of RNA from large quantities of wheat germ followed by deproteinzation of the crude extract using DE-32 ion-exchange chromatography and gel permeation chromatography by using Sephadex G-75. Different combinations of methods have been found optimal for increasing release of the 5S RNA from the c e ll, as well as suppressing ribonuclease activity, removing high-molecular weight contaminants, and suppressing 5.8S RNA release. 39 B. Isolation and Purification 1. Materials and methods. All chemicals were reagent grade and used without further purification. Wheat germ was a gift from International Multifoods, Columbus, Ohio. A flow diagram for isolation of mixed RNAs from wheat germ is shown in Figure 15, whereas a flow diagram for purification of crude 5S RNA is shown in Figure 16. 2. Phenol-SDS extraction. The details of the phenol-SDS extraction procedures are described step by step in Figure 15. Several important items are crucial at this stage. F irs t, the dry wheat germ has to be swollen for four hours in sodium acetate buffer in the presence of bentonite. Almost no 5S RNA could be recovered if this step was omitted. Second, the concentration of SDS is chosen to optimize the maximum release of 5S RNA from cell membranes while not the DNA from nuclear membrane. Too much SDS in the solution would also release some unidentified macromolecules, which would cause separation difficulties later in the precedure. In this experiment, the percentage of SDS in the buffer solution is around 0.15% to 0.2%. Third, the activity of ribonuclease from wheat germ itself and other sources such as glassware, dirty aparatus, and the hands of researchers, is inhibited by the a combination of the three things. First, the bentonite mentioned above is an inhibitor for ribonuclease, and the swollen cells are in a buffer containing 6 mg/ml of bentonite. 40 600g dried wheat germ (ca. 20% H2O) Swell for 5 hr in 1800 ml of buffer containing 0.01 M NaOAc, pH 5.0, 0.1 M NaCl, 10 g Bentonite Centrifuge at 2,500 x g foe 4 min Precipitate Supernatant (discard) Add 2000 ml phenol ( 88 %) Add 2000 ml 0.01M NaOAc, pH 5.0, 0.1 M NaCl, 0.15% SDS Stir for 40 min Centrifuge at 10,000 x g for 15 min Aqueous layer Phenol layer and debris (discard) Add 1/2 volume of phenol ( 88 %) Stir for 10 min Centrifuge at 10,000 x g for 15 min Aqueous layer Phenol layer and debris (discard) Add 2.5 volumes of cold (-20°c) 95% ethanol Let stand overnight Centrifuge at 10,000 x g for 15 min Precipitate Supernatant (discard) Dissolve in 200 ml 0.01 M tris-base, pH 7.5, 0.3 M NaCl Centrifuge at 10,000 x g for 15 min Precipitate (discard) Supernatant (crude extract) Proteins and mixed RNA's ( 18S RNA, 26S RNA, 5S RNA, tRNA’ s) Figure 15. Flow diagram for isolation of crude RNA from dried wheat germ. 41 Second, phenol is a powerful solvent to separate the proteins including ribonuclease from aqueous phase (RNAs stay in the aqueous phase). Third, it is well known that the hydrophobic long tail of SDS denatures the ribonuclease. In summary, phenol separation, SDS denaturation, and bentonite inactivation eliminate the ribonuclease activity. Finally, low temperature extraction (4°C) is performed, in order to suppress the release of 5.8S RNA, which w ill increase the ease of purification in the next step. 3. Ion-exchange chromatography. Ion exchange may be defined as the reversible exchange of ions in solution with ions electrostatically bound to some sort of insoluble support medium. In the present experiment, an anion exchanger (DE-32) with the ionizable group, diethyl ami noethyl (-OCHzCHzNf 0285 )2) is used. Because successful perfomance with the ion exchange technique is largely based on proper preparation of the exchanger, the DE-32 powder is treated in an orderly procedure. In the case of DE-32, treatment with hydrochloric acid (0.5N) results in the conversion of a ll the diethyl ami noethyl groups to their charged form (C2H4N+H(02^ ) 2 ) • Placing positive charges on the functional groups results in mutual repulsion and hence a maximum amount of swelling (exposure of the charge groups). Once the matrix is fu lly swollen the acid is washed away until the pH is 4.0, then the resin is treated with sodium hydroxide (0.5N). This converts the functional groups of DE-32 to their free base form, then the base is washed away until the 42 Crude mixed RNA’s and proteins Apply to Watman DE-32 column (5 x6 cm) pre-equilibrated with more then 100 volume of 0.01 M Tris-base, pH 7.5, 0.3 M NaCl Wash with 10 volumes of same buffer u n til! A260 < 0 .3 Elute with O.ol M tris-base, pH 7.5, 1.0 M NaCl Collect fraction with A26O > 2 .0 Precipitate with 2.5 volume cold (-20°c) 95% ethanol Dissolve in 0.01 M tris-base, pH 7.5, 2.0 M NaCl Centrifuge at 10,000 x g for 10 min and discard debris Dilute the solution with water to 1.0 M NaCl Apply to Sephadex G-75 column (5 x 150 cm ) Elute with same buffer 35-45 mg of electrophoretically homogeneous 5S RNA from 600 g dry wheat germ can be obtained (see Figure 19) Figure 16. Flow diagram for purification of crude 5S RNA from phenol-SDS extraction. 43 pH is 8.0; this is the form that can be most easily equilibrated with the desired counterions. About 20g of DE-32 is treated as described above and packed onto a 5 x 50 cm column. Fifteen times the DE-32 volume of the equilibrium buffer is passed through the column (10 mM tris-base, 0.3 M NaCl, pH 7.5, 2 l i t e r ) . The precipitate in ethanol is then concentrated by centrifugation at 0°C, 10,000g for 10 min, and dissolved in 200 ml of the sample buffer. The supernatant is then loaded onto the column subsequently, and 1.5 lite rs of the same buffer are used to wash the column until the absorbance at 260 nm is below 0.3 O.D. At this stage, the neutral compounds, positively charged compounds, and the majority of proteins (at pH 7.5, 0.3 M NaCl, few proteins can bind to the anion exchanger) as well as weakly negatively charged compounds are eluted from the column. Only the nucleotides such as 26S, 18S, 5.8S, 5S, and tRNAs remain bound to the column. In this way, the nucleotides can be separated from other biomolecules. Finally, an elution buffer (10 mM tris-base, 1.0 M NaCl, pH 7.5) is added to the column resulting in strong ionic competition for the nucleotides binding sites. The mixed nucleotides are eluted out, then collected and precipitated with precooled ethanol (-20°C). The elution profile of the ion-exchange chromatography is shown in Figure 17. 4. Gel filtr a tio n chromatography (G-75) Comparison with the ion-exchange method, in which separation is based on charge, separation via the gel permeation chromatography 44 1.0 N 0.3 It NaCl NaCl o CO CM < F R ACTION Figure 17. A DE-32 anion exchange elution p rofile for the phenol-SDS extract of wheat germ. The earlier eluting peak contains impurities while the peak eluted by 1.0 M NaCl contains a purified mix of RNAs. 45 technique is based on molecular size. The various media which are used for this purpose are spherical beads composed of a spongelike matrices which containing pores of relatively restricted size distributions. When a mixture of d ifferen t sized molecules are loaded on the top of a column containing these beads, the large molecules cannot easily diffuse into the pores and are eluted from the column with little or no resistance (Figure 18), whereas the small molecules diffuse into the pores of the gel beads; therefore, th eir movement is slowed while they diffuse through the pores in the beads of the gel matrix. As a result, the bigger the molecules are, the fast the molecules elute o ff the column. In this way, the mixture of the RNA nucleotides can be separated re latively easily (see Figure 18). For this purpose, Sephadex G-75, a gel filtr a tio n medium, is used. The number 75 in the G-75 refers to the water regaining value of the dry gel multiplied by a factor of 10. The gel is produced by permitting a microorganism, Leuconostoc mesenteroides, to ferment sucrose into large polymers of glucose. The polyglucose units are purified and cross-linked by treatment with epichlorohydrin into a particular class of gel beads, (Epichlorohydrin introduces glyceryl groups which cross-link the polyglucose units). About 210g Sephadex G-75 are treated following the specification from the manufacturer. About four liters of buffer (1.0 M NaCl, 10 mM tris-base, pH 7.5) was prepared for the swelling the 210g G-75 gel. During the gel preparation, the swollen gel has to be cooked in this boiling buffer for 3 hours. In this experiment, a column which is 150 cm long and 5 cm in diameter is employed . Intensive care has to be taken so that 46 Migrates Rapidly f f Eluant Flow Migrates Slowly Figure 18 . Separation of large and small molecules using gel permeation chromatography. Molecules are represented schematically as large and small filled circles. 47 the operating pressure does not exceed 75 cm of water in order to avoid crushing the gel and maintaining correct flow rate la te r on. After packing, the column must be equilibrated for a 36 hour period. The mixture of 26S, 18S, 5.8S, 5S and tRNAs is collected by centrifuging at 10,000g at 0°C fo r 10 min. The precipitate is dissolved into 15 ml buffer (1.0 M NaCl, 10 mM tris-base, pH 7.5). About 250 mg of mixed RNAs are loaded onto the 5 x 150 cm, 6-75 column. The flow rate is adjusted in such a way that the separation not only has better resolution, but also saves time. The flow rate is set at 4.5 sec/drop; every fraction tube contains about 16 ml of volume (350 drops). Each time, about 40 mg purified 5S RNA can be obtained. The elution profile is shown in Figure 19. In Figure 19A, although the three peaks are well resolved, the background absorption is too high to be acceptable. The cause of this phenomenon is not because of bad column resolution instead, i t is caused by some kind of contaminants which are not yet identified. There is an easy way to get rid of such contaminants. Before loading the column, the precipitate is dissolved in 2.0 M NaCl, 10 mM tris-base, pH 7.5, and the undissolved contaminants removed by centrifugation. Then the solution is diluted with water to 1.0 M NaCl, and loaded onto a Sephadex column. The improvement by this technique may be seen in Figure 19B: i t is clear that the peaks are well resolved while the higher baseline background has been removed. iue 9 () Euin rfl fo Spae G7 gl permeation gel Sephadex G-75 from profile Elution (A). 19. Figure ABSORBANCE rRNA hoaorpy f ie RA, x 5 c clm, 5 ml/h, 35 cm column, 150 x RNAs, 5 mixed chromatography of e rd f h cnaiet; e tx fr details. for text see contaminents; the of rid get B. h sm a aoe xet . M al xrcin to extraction M NaCl 2.0 except above same The as (B). denote arrows The ml/tube. 16 h asrac o unknown contaminants. of absorbance the FRACTION S tRNA 5S 48 49 C. Purity of WG 5S RNA. 1. Gel electrophoresis. Electrophoresis techniques are often used as a principal tool for characterizing macromolecules and for assaying th eir p u rities. This method is based on the fact that molecules such as RNA and proteins possess a charge and therefore are able to move when placed in an electric fie ld . Since each molecule is expected to possess a unique charge and size, it will migrate to a unique position within the electric field in a given length of time. Therefore, if a mixture of RNAs is subjected to a charged fie ld , each of the RNAs would be expected to concentrate into a tight band. The electrophoresis gel consists of several chemical compounds, such as acrylamide, N,N'-methylene-bis(acrylamide), tetramethylenediamine (TEMED) (see Figure 20) and ammonium persulfate. The ammonium persulfate produces free radicals to activate acrylamide molecules to produce a long polymer chain. The N,N'-methylenebis- ( acrylamide) then hooks the long polymer chains together by cross-linking them to one another to form a net of acrylamide chains. TEMED serves as a catalyst for the gel formation because of its a b ility to exist in a free radical form. In the present experiment, non-denatured and denatured (see section E.2 .b.) slab gel electrophoresis techniques are used. In the slab gel technique, acrylamide is polymerized into a thin rectangular slab between two glass plates. Sample wells are made at one end of 50 A o ii c h 2= c h - c - n h 2 B HoCs H3 p o o ^ II II CH2=CH-C-NH-CH2-NH-C-CH = CH2 D N ?H c h 2= c h - c h 2- n h - c - ch I CHo=CH-CH2-NH-C-CH II I O OH Figure 20. Compounds used in the synthesis of acrylamide gels. (A), Acrylamide. (B), Tetramethylenediamine (TEMED). (C), N.N'-Methylene-bis (acrylamide). (D), N,N’-dial lyltartardiamide. 51 5.8S o m w < DISTANCE (cm) — Figure 21. Polyacrylamide slab gel electrophoresis profiles for RNA extracts, (a ), Wheat germ mixed RNAs. (b ), Wheat germ 5S RNA afte r G-75. The purity is more than 98%. 52 the gel by placing a comb-shaped jig into the reaction mixture before it polymerizes. After polymerization the jig is removed leaving sample wells molded into the acrylamide. The advantage of this technique is the ease with which a number of experimental samples can be compared. Since a ll of the samples are present in the same gel the electrophoresis conditios are quite constant from sample to sample. In this way, large rRNA, 5.8S RNA, 5S RNA, and tRNAs bands are easily distinguished (see Figure 21). For quick assay, particular gel compositions are chosen so that the gel electrophoresis gives not only better resolution, but also short running time. This is achieved by using the following recipes. a. Stock solution 1 ). TBE 10X (1.0 M tris-B orate, pH 8.3, 20 mM EDTA). 2). Monomer solution (Acrylamide/bisacrylamide is 38%/2%). b. Gel composition (20 ml gel solution to make one piece of slab gel). 1). TBE 10X solution 2 ml 2 ). Monomer solution 4 ml 3 ). Urea 10 9 4 ). Ammonium persulfate 7 mg 5). Water 6.7 ml 6 ). TEMED 30 ul c. Sample Buffer 1). 80% (v/v) formamide 2 ). 10 mM NaOH 3 ). 1 mM EDTA 53 4 ). 0.1% Brominphenol blue. d. Sample preparation for gel electrophoresis About one O.D. (A 260) sample is taken to be precipitated with cold ethanol in a small v ia l. Then, a fte r centrifugation, the precipitant is placed in a vacum desiccator for 10 min. Then 15 ul sample of buffer is added into the vial which contains the dried sample. The vial is vortexed and then heated in 65°c water bath for 5 min to denature the samples. At this point, the sample is ready to be loaded onto the wells in the gel. e. Making the slab gel. A few commemts are in order. F irs t, since resolvation of urea in water is an endothermic reaction, one must warm i t up in order to dissolve the urea. However, the warmed gel solution will polymerize too fast to handle before one can make the slab gel. To avoid such a problem, one f ir s t cools down the solution in ice bath. Second, the catalyst TEMED is added only at the final step ju s t before the gel is put into the slab gel aparatus. About 30 min is needed for the gel to polymerize. Then wells must be cleaned to remove the heavier density solution in order that the sample may be loaded properly. Running the electrophoresis, the electric power supply should be set for the constant voltage mode. 240 V are applied to the slab gel and sample. About 4 watts are required under these conditions. Passing cooling water though the gel aparatus is unnecessary. The constant power mode is not suggested for the novice, who might leave the equipment unwatched, because leakage in the upper tank w ill burn out the power supply. The total running time for complete electrophoresis is two and half hours; and the gel 54 is then fixed in 1.0 M acetate for 15 min. Subsequently the gel is transferred to the methylene blue stain solution for another 15 min period. The destain procedure is done by changing d is tille d water several times until clear bands of RNAs can be seen. 2. Gel scanning Gel scanning is performed by using the DU-8 gel scanning system on a DU-8 UV-Visible spectrophotometer. The system consists of a UV-Visible computing Spectrophotometer, Digital plotter, sample transport mechanism, tube gel holder and gel scan Compuset". The destained slab gel is cut into pieces suitable for the gel tube. To obtain a good quality picture of gel scan, i t is necessary to f i l l up the gel tube with destilled water (make sure no a ir bubble is in i t ) . The typical printer-operator dialogue is as following: A -l, T-2, PROG-3? 3 73 GEL SLIT? .10 PROG #? 00 (.00-.99) 1 A -l, T-2? 1 3 A 550.0 74 GEL START? 10.0 (1-210) 8 READ AVG? 3 75 GEL END? 100.0 (10.1-210) 11 SLIT WIDTH? 05 76 AREA CALC? YES .1,.2,.5,1,2,5NM (YES/NO) 45 OFFSET -.3A? YES 1-AUTO, 2-SELECT? 2 (YES/NO) FACTOR? (Y/N) NO 46 SUPPRESS? 80 MIN PEAK? 0.000 (0-4.000) (-0.300 TO 4) 0.000 81 SENSITIVITY? 6 47 SPAN (.001-4)? (1-9) 2.000 85 SAVE PROG? YES 48 CHT SPEED? 1 (YES/NO) (.1,.2,.4,1,2, PROG #? 06 5,10,30 CM/MIN) PUSH GO TO START 72 GEL SPEED? 1 (.1,.2,.4,1,2, 5,10,30 CM/MIN 55 The results of gel scans are shown in Figure 21. The purity of the 5S RNA was more than 98%. D. Concentration determination. An accurate concentration determination of 5S RNA is very difficult. Usually, a value of 22.4 O.D. (Aggg) is used for 1.0 mg of 5S RNA. However, since these experiments are done under d ifferen t ionic conditions, the absorption will be different even though they contain the same amount of sample. This fact is due to the hyperchromism of the molecule (see Chapter I I I ) . In these experiments, a ll concentration measurements are made using the following procedure unless otherwise specified. The procedure for determing the concentration of wheat germ 5S RNA is based on the determination of mononucleotide 5' phosphate concentration. Aliquots of original stock solutions of ribonucleic acids were hydrolyzed in 0.3 M NaOH at 37°c for 18 hr. The molar concentration of nucleotide in each digest was calculated by using published extinction coefficients (18) at A 26O the 5'-phosphates in alkaline solution and expressed as moles of phosphate. For 5'AMP, 5'UMP, 5'6MP, and 5'CMP, the extinction coefficients are 15.3, 7.3, 11.5, and 7.4 per mM per cm, respectively. The following formula is used for calculating the average extinction coefficient of 5S RNA (in millimoles of phosphate residues after alkaline hydrolysis); f-j: fraction of particular base in 5S RNA, c-j: extinction c o iffic ie n t of particular mononucleotide 5' phosphate. 56 According to primary sequence of wheat germ 5S RNA, f^ is 29/120, fu is 26/120, fG is 34/120, and fc is 31/120. E26o of 5S RNA in phosphate is then f A x CA + f U x CU + f G * CG + fC x cC = 29/120 x 15.3 + 26/120 x 7.3 + 34/120 x 11.5 + 31/120 x + 31/120 x 7.4 = 10.45 (mM-1 cm-*). I f one measures the absorption of the hydrolyzed 5S RNA in an alkaline solution at A260» the concentration of WG 5S RNA expressed in mM is shown by a260 [5S RNA],^ = ------/ 120 10.45 Such calculation for the determination of WG 5S RNA concentrations is rather complex. Therefore, in order to make these measurements simpler and quicker, a comparative measurement is used based on the determination of method shown above. I f we assume that in the absence of Na+ and Mg++, both methods give the same concentration; then by using the data as reference from phosphate determination method, then a value of 23.7 O.D. (A2 5q) corresponds to 1 mg of WG 5S RNA. In the following four different conditions, the [P i] determination method and UV measurement method are compared (see Table 1). From Table 1, one can find that the values from the UV method are always below the corresponding values from [P i] determination. This phenomenon is due to the different molecular folding and different hyperchromism. The bigger the factor is , the more the molecules fold. This is consistent with UV melting measurements (187). 57 Table 1. Comparison of WG 5S RNA concentration determination between [P i] method and UV method. Salt Condition [Pi] {MW of WG UV (23.7 O.D. Factor between Na+ , Mg++ 5S is 38,700) as 1 mg) the two methods No, No 6.147 6.147 1 100 mM, No 6.649 6.147 1.082 100 mM, 10 mM 6.473 5.760 1.124 No, 10 mM 4.354 3.778 1.153 Note: Using this Table and UV measurement together with the corresponding factors to make quick measurement of concentration of WG 5S RNA, the following formula applies: A26O x Dilution Factor [mg/ml] = x Factor from Table 1. 5S RNA 58 E. WG 5S RNA fragments by RNase T1 cleavage. 1. Fragment A (bases 1-86, 87 and/or 89). a. RNase T1 digestion. 40 mg of wheat germ 5S RNA are dissolved in 40 ml of buffer which contains 0.3 M Na+, 10 mM tris-base, at pH 7.5. The solution is incubated in the ice bath. Using this concentration of wheat germ 5S RNA (1 mg/ml), 3.5 ul of RNase T1 ( 350 units from Boehringer Mannheim) per mg of 5S RNA were required for digestion. Every 10 min, the solution was vortexed to make the reaction completely homogeneous. After 137 min, the intact molecules are completely cleaved at the f ir s t order cleavage sites. To stop the reaction, 40 ml of a 5% SOS solution is added together with 40 ml of phenol. The mixture of these solutions is then stirred for 2 min and then centrifuged. The aqueous phase was then removed. The phenol extraction procedure is repeated twice in order to remove the RNase T1 completely. The aqueous phase solution is then precipitated with cold ethanol. It is also may be interesting and useful to mention how this condition was found. The amount of 1 mg in 1 ml of WG 5S RNA is prepared under the condition described above. The enzymatic reation continues for 47 min, then every 10 min a small amount of sample (100 ul) was taken out and the reaction is stopped immediately. These samples are then subjected to a denaturing electrophoresis procedure . In this way, the best condition required for complete cleavage of intact molecule was found. Figure 22 shows the time-dependence of the cleavage reaction. 59 Figure 22. Gel scans of the electrophoresis. (A), the intact WG 5S RNA was digested by RNase T1 at 0°c for 47 min. The cleavage mixtures of the molecule were applied to denaturing gel electrophoresis. Peak a and peak b are complementary fragments of the intact molecule. The uncleaved 5S RNA peak was approximately 30% afte r 47 min reaction. (B), 107 min digestion. Only 5% of the intact 5S RNAis remaining. (C), 137 min digestion. The intact 5S RNA molecules are all cleaved. This result makes the following separation much easier. (D), The purified RNase T1 fragment of WG 5S RNA from the denatured Sephadex G-75 gel filtr a tio n chromatography. The Y-axes are the gel absorbances at 550 nm. 60 a 5S J L i J L i 1 I 1 1 1 1 j DISTANCE (cm) Figure 22 61 b. Oligonucleotide separation: denaturing column. Since the cleavage described above produces partial digestion, the 5S RNA molecule is s t ill held together by its secondary structure binding. This has been proved though use of a non-denaturing electrophoresis procedure which shows the same gel pattern as for the intact 5S molecule (The non-denaturing gel technique is shown below). Sequencing papers (171-172) for this molecule have shown the same results: RNase T1 under these conditions preferentially cleaves the G residues at positions 86, 87, and/or 89 as shown in Figure 13). It is reasonable then to conclude that the two pieces of the 5S molecule are hydrogen-bonded and folded together. Obviously, the separation of the two pieces of the molecule can be achieved only by a denaturing method. For this purpose, 4 M urea buffer solution was chosen for a Sephadex G-75 column at room temperature. This condition is easily checked out by UV melting measurement, which shows that under these conditions there is no hyperchromism present throughout a temperature range from 25°C to 80°C. One then analyzes to see that these digested fragments are completely separated as single strands without any base stacking or base pairing at room temperature. RNase T1 cleaved RNA fragment separation from the digest precipitate was achieved by Sephadex G-7 5 gel fitra tio n chromatography (150cm x 2.5cm) in 4M urea, 50 mM NaCl, ImM EDTA, 10 mM tris-base, pH 7.0 at room temperature. The eluted fractions of the sample cannot be precipitated by ethanol because of the presence of 4 M urea. Therefore the urea is first removed by dialysis against 10 mM NaCl, 62 tris-base, 100 mM NaCl, pH 7.0. At this point, the urea-free sample can be was recovered by ethanol precipitation once again. The elution profile from the denaturing column is shown in Figure 23. 2. Fragments B1 (base 1-25, 57-85, 90-120) and B2 (base 26-56). a. RNase T1 digestion. In contrast to the fragment A digestion, a different digestion condition is chosen in order to obtain a completely different fragment. According to the sequence papers (171, 172), the RNase T1 cleavage sites favor at f ir s t the bases at the 86 , 87 and/or 89 positions. The second order cleavage sites are at positions 25 and 56. In the proposed secondary structure (see Figures 12, 13, and 14), such cleavage sites would s p lit the molecule as two complementary parts. These two pieces of the molecule can be separated by using non-denaturi ng column. This experiment was performed as follows: 1 mg of 5S RNA is dissolved in 1 ml of buffer prepared in 0.3 M NaCl, 10 mM tris-base pH 7.5 buffer solution at room temparature. 10 ul of RNase T1 (from the same company as before) was added to the solution to start the cleavage reaction. Each 10 min, 200 ul of solution were removed to which 5% SDS and phenol were added to stop the reaction and remove the RNase T l. Such samples were then subjected to electrophoresis in a non-denaturing gel. The gel consists of 10% acrylamide, 0.5% bis(acrylam ide), 0.08% TEMED, and 0.5% ammonium persulfate. The buffer is 0.1 M KC1 , 5 mM MgCl 2 , and 50 mM 63 o co CM < oLU z < c a o c O F R ACTION Figure 23. The elution profile from the denaturing column (Sephadex G-75, 2.5 x 150 cm) used to separate the RNase T1 fragment A. The f ir s t peak was collected and treated as described in the text. 64 tris-b o rate, pH 7.8. Gel were run at 3V/cm for 16 hours at room temperature, and RNA was visualized by staining with methylene blue (as described in section C .I.) . To achieve a successful pattern on the non-denaturing electrophoresis as compared to denaturing electrophoresis, the running power is set to 24 volts instead of 240 volts, and the sample concentration should be as high as possible; also, the amount of sample put in the wells of the slab gel should be as small as possible, thereby increasing the resolution of non-denaturing electrophoresis bands. Figure 24 shows a gel scan of the non-denaturing electrophoresis. The f ir s t and the second peaks are complementary fragments. The third peak represents small pieces of cleaved ribonucleotides, and has no secondary structure according to J-H NMR experiments (see Chapter V). b. Oligonucleotide separation (non-denaturing column). D ifferent methods must be employed from that for fragment A in order to obtain a purified sample of the fragments Bj and B 2 . A non-denaturi ng column instead of the denaturing column was employed. The column condition is identical with the conditions described in section B.4, except that the column diameter is 2.5 cm instead of 5 cm. The elution profile is shown in Figure 24A. The fir s t peak (called the core of the molecule) does not contain just one polynucleotide chain, but rather three oligonucleotide chains which are held together by the secondary structure. The elution profile of 65 FRACTION DISTANCE DISTANCE (cm) profile from G-75 gel permeation chromatography of the sample as above. Both methods give the results, same cm column, cm 14 m l/hr, 7ml/tube. confirming (B) that The gel the scan was sample of separated the and purified. mixed 5S RNA mixed fragments 5S RNA from T1 RNase digestion, 2.5 x 150 Xhe Xhe the non-denaturing electrophoresis. (A) Elution B 09Zy OSS Figure 24. Profiles for the non-denaturing gel filtration column and < 66 the column is almost identical with non-denaturing electrophoresis result (see Figure 24B), confirming that the sample is completely separated and purified, because the electrophoresis technique is very sensitive to low levels of RNA fragments. F. Discussion. 1. The conditions for enzymatic cleavage of WG 5S RNA by RNase Tl. In order to obtain high yield recovery and uniform cleavage, the amount of the enzyme, the concentration of the WG 5S RNA, the temperature and the time of the reaction, and also the buffer conditions have to be considered. For mild cleavage conditions (0°c, 350 units of RNase T l), the major cleavage sites are the first-o rd er positions; for strong cleavage conditions (room temperature, 1,000 units of RNase T l), major cuts occur at both f ir s t and second order positions. Using different isolation strategies, fragments A, Bi, and B2 are obtained. For the purpose of la te r purification, the intact molecules are designed to be cut completely at their particular positions. Although this may cause a decrease in the yield , better purity of these fragments can be obtained. This is a kind of compromise. The reason is that both columns cannot separate the intact 5S RNA molecule from its biggest fragments. 67 2. The homogeneity of the fragments. Since the f ir s t order cleavage sites are located at the primary sequence positions 86 , 87, and/or 89 (G residues), the fragments may d iffe r between themselves by one or two bases. Hence, the fragments are not homogeneous. However, i f these positions of 86 , 87, and/or 89 are located in a single-stranded region (this may also be proved by this experiment), they may not contribute to any resonances at the low field region of ^H NMR spectrum, which detects only the base-paired protons (see Chapter V). A unique position located at G41 was not cut even with very strong doses of RNase Tl (1000 units at 1 mg/ml of 5S RNA). This result was not sim ilar with enzymatic studies done on _T. u t ilis and yeast 5S RNA. In particular, the G 41AAC sequence is conserved in 5S RNA's. The inaccessibility of enzyme to this position may imply that the G41 may be involved in the tertiary interactions of the molecule. CHAPTER I I I BASE PAIRING IN WHEAT GERM 5S RNA AS MEASURED BY FOURIER-TRANSFORM INFRARED SPECTROSCOPY A. Introduction 5S RNA appears to be present and essential for protein synthesis in virtually every studied prokaryotic or eukaryotic ribosome (19). It is therefore widely speculated that a universial secondary structure exists, and several models have been proposed (69-78). These models necessarily d iffe r in both the total number and the relative number of the three main base-pair types (A*U, G*C, and G*U). Choosing between the existing models or devising a new model therefore requires techniques that can discriminate between the four bases with respect to base stacking and/or base pairing. Ultraviolet (UV) absorption (187), optical rotatory dispersion (ORD) (165, 166), and circular dichroism (125, 164) have been extensively used in studies of 5S RNA secondary structure. Although these studies have contributed to our understanding of the gross features of 5S RNA secondary sturcture, more detailed knowledge of the extent and kind of base pairing is required. Although UV spectra of polynucleotides are sensitive to base stacking, 68 69 analysis for the proportion of individual base-pair types can be obscured by the high degree of overlap between spectra for each of the various bases or base pairs. In this situation, infrared vibrational spectroscopy appears to be one method well suited to this type of study (191). For the analysis of the base-pairing content of RNAs, the 1750-1450 cm"! region of the IR spectrum is of particular interest. The IR absorption bands in this region originate from the in-plane vibrations of the C=0, C=C, and C=N groups of the heterocyclic bases. The frequencies and intensities of the bands are sensitive to interactions such as double-stranded complex formation of complementary polynucleotides. These effects allow IR spectroscopy to be used for the characterization of 5S RNA secondary structure. The first serious attempts to determine RNA base pairing from IR hyperchromism were based on comparing a single IR spectrum of native JL* c°1i 5S RNA with the best f i t linear combination of reference spectra of poly(rA)•poly(U) and poly(rG)*poly( rC) for base-paired segments and 5 '-3 ' APA, 5 '-3 ' UPU, 5 '-3 ' CPC, and 5'-GMP for unpaired segments (Appel 1979, Stults 1981). In both cases, the deduced total base-pair number turned out to be too large by at least 50% when compared to other available measurements (69, 70, 125, 139, 190). A more recent and more accurate infrared determination gave 33 base pairs for coli 5S RNA (127); the improvement was ascribed to more accurate base-line compensation for the solvent (D 2O and HDO). I t is certain that strong absorption by the solvent leads to noisy and imprecise background ("base line") in the IR difference spectra 70 from which the weak RNA absorption bands are measured. The more important improvement of the measurement is fittin g the RNA difference spectrum for unpaired minus paired bases (90°C minus 20°C) with the reference difference spectra between paired and unpaired poly ( rA)*poly(rU), and poly(rG)*poly(rC), rather than simulating an RNA extinction spectrum at a single temperature. Since the IR absorption of RNA, even at 90°c, is smaller at a ll wavelengths (1450-1750 cm_l) than the summed spectra of its component mononucleotides, any attempt to simulate the RNA IR spectrum at just one temperature w ill overestimate the apparent number of paired bases because the reduced intensity in RNA compared to mononucleotides w ill be interpreted as additional base pairs. However, a difference measurement, (90°C)- rn/\ (30°C), should remove that intensity discrepancy, leaving an RNA difference spectrum that more accurately reflects base pairing. More recently, in a study of yeast 5S RNA, the precision of the base-pair estimates was further improved. The high dynamic range, signal-averaging capability, and facility for digitized data manipulation (filtering, base-line correction, difference spectra) of Fourier Transform Infrared (FT-IR) spectroscopy provides major advantages over the dispersive IR instruments from which all published IR spectra of 5S RNA had been obtained. The results can then be analyzed to yield reliab le estimates for the numbers of G*C and A*U base pairs in the 5S RNA. In this chapter, the FT-IR technique is used to discover the numbers and types of base pairs in eukaryotic wheat germ 5S RNA. 71 B. Theory 1. Interferometer. Figure 25a shows the optical system of a typical interferometric spectrometer. A polychromatic beam from source S is chopped, and after traversing any sample present, passes through a circular aperture and is collimated. The parallel beam is directed at an angle of 45° onto a beamsplitter, T. Part of the radiation is reflected and the remainder transmitted: the reflected part fa lls at normal incidence on a plane mirror, which is capable of movement along a lin e perpendicular to its surface, and is returned to be partially transmitted by T and fin a lly to form, with the help of a condensing system, a reduced image of the entrance aperture on detector D. That part of the original parallel beam transmitted by the beamsplitter falls at normal incidence on a fixed plane mirror, is returned to T and after reflection is also condensed onto the detector. Since the two parts of the original beam follow different paths before falling on the detector, as shown in Figure 25b, there is a relative phase delay determined by the path difference xP^y-xP 2y within the interferometer. As the path difference (x) is steadily increased from zero by movement of the plane mirror, detector output fa lls from a maximum (in plane) to a minimum and therefore passes through a succession of maxima and minima with a tendency for the fluctuations to diminish. The resulting curve is known as an interferogram (Figure 26a) 72 a n u b Figure 25. (a) Optical system of a typical interferometric spectrometer; (b) The original beam passes through d ifferent paths, xPjy and xP 2y (231). 73 a Distance b Figure 26. (a ), a typical interferogram from Michelson interferometer, (b), Resultant electric vector R for two equal rays with phase difference (231). 74 To understand how the spectrum is obtained from the interferogram, it is helpful to proceed in stages. First of all, if the source is assumed to emit a single frequency V]_, the two part of the beam on recombination w ill go alternately in and out of phase as the path difference x is steadily increased. The resultant at any instant of two rays with unit electric vectors and differing in phase by a= 2iLvix is obtained by vector addition (Figure 26b). The electric vector of the resultant, R, is 2cosa/2, and since the power P associated with a ray is proportional to the square of the amplitude of the electric vector, we have 4cos 2a/2 cx l+cosa=l+cos2iCvix. For any component v, the power of the detector is: P( x) = B(v) cos2i[vx dv (3.1) Assuming that the constant part of power which is invariant with x is zero, for all components, the detector power is represented by the equation: P(x) v)* cos2*tvx dv. This is the interferogram equation which depends on the spectrum according to a cosine Fourier integral. This relation may be written as: B ( v) = 4 / F(x) cos21j>x dx, (3.2) J 0 showing how the spectrum may be calculated by the operation of the Fourier transformation on the measured interferogram F(x). This relation is the basis of the Fourier transform infrared spectroscopy. Infrared interferometers are to ta lly dependent upon a computer for data collection, Fourier transformation to yield a single beam 75 spectrum, and subsequent computation which results in an absorbance spectrum that has been corrected for instrumental contribution, solvent, and vibrational absorptions of the sample which might mask that of the desired group (see Figure 29). The interferometer has a number of advantages over the conventional grating instruments (192). These include the Jacquinot and F ellg ett advantages. For the Jacquinot advantage, the transmission measurement comes from the larger circular entrance aperture, and therefore the greater optical energy throughout of the interferometer. The resolution of a grating instrument is proportional to the width of its entrance s lit . A narrow s lit is necessary for high resolution, but this restricts the optical through out. In contrast, the interferometer allows a large circular aperture without serious limitation on the resolution, which is proportional to the distance of moving mirror travel. For the Fellgett advantage (193), if the system is detector noise limited, then the signal increases in proportion to the time of observation n, but the noise, being random, only increases in proportion to n1/ 2 , so that signal/noise is proportional to n*/2. when there are m spectral elements (m is equal to the spectrum width divided by the width of a spectral line) contributing energy to the signal and they are observed sequentially, as in most dispersive spectrometers, then each element is observed for a time n/m and the S/N is proportional to (n/m)l/2. There is then a net gain of ml / 2 in S/N for the system in which all elements are observed simultaneously as opposed to observing them sequentially, and this is the F ellg ett advantage. I t should be noted 76 that this becomes very important in high resolution research { V < 2 cm-1). A further advantage of interferometers for biological studies is that the sample is placed afte r the interferometer, and thus receives only a fraction of the infrared radiation from the source. I t is thus much easier to maintain a constant sample temperature during data collection than with a dispersive instrument in which the sample is placed before the monochrometer and receives the fu ll intensity of the focused infrared beam. 2. Determination of number and types of base pairs A relation has been developed (191, 194) between the mid-infrared absorbance of the bases of RNA and its base-pair composition. The method is based on the increase in IR absorption observed on unstacking of the bases (as by increase in temperature) and an assumed correspondence between base stacking and base pairing. The method begins by two assumptions. F irs t, the four ribonucleotides only have stacked (s) or unstacked (u) states. Second, a ll stacked bases are paired and that all base pairs are either A*U or G*C (e.g., no G*U pairs). L e tffv )^ and £(v)j>u be the respective molar extinction coefficients at v of base i in the stacked and unstacked states. £ (v )a ,u, £(v)(jfU, e (v )G,u, and £(v)c , u can be determined experimentally from spectra of the mononucleotide 5 '-phosphates. £ (v )a , s and £ (v )y ,s cannot be measured separately, but their sum can be determined from a spectrum of poly(rA)*poly(rU), from which 77 poly ( rA)"poly(rU), from which g ( v )a .jj = ( 1/ 2) L e(v)A>s + e (v)ll»s ^ (3.3) Sim ilarly, from a spectrum of poly (rG)-poly (rC ), e (v )G*U = d /2 ) L c (v)G, s + e ( v )q , s ] (3.4) Finally, i f all base pairs are "melted" ( i . e . , unstacked) at 90UC and a ll base pairs are present at 30”C, and no G-U's are present, then e ( v) - e( v) RNA,90°C RNA,30UC e(v) + e (v) A,u U,u - e(v)A-u J x fA>u 2 e(v) + g( v ) G ,u C,u + L e ( v)G*C J x f G*C 2 (3.5) Let e(v) ++ e(v) e(v) A,UA,,U U,U - g ( v )a .[j (3 .6 ) 2 and g( v) + g( v ) G,u C,u b = ------g( v )q .g (3.7) 78 Rearranging equation 3.5, we obtain € (tJ) - £ ( i l ) RNA, 90 °C RNA,30°C b = fA*U + ----- fG-C 2a a » where a is defined as the normalized (per mole of phosphate) absorbance difference between stacked and unstacked states for an equimolar mixture of A and U (Eq. 3.6) and b is defined as the normalized absorbance difference between stacked and unstacked G and C (Eq. 3 .7 ). Equation 3.8 is a linear equation of the form y = b + mx from which the fraction of paired A*U's in the RNA is equal to b ( i . e . , the y intercept) and the fraction of paired G-C's is equal to m ( i . e . , the slope). The number of paired A*U (or G*C) is obtained by multiplying fA.u (or fg-c) by tbe total number of bases in the RNA (e.g., 120 for wheat germ 5S RNA). C. Materials and Methods 1. Preparation of 5S RNA for FT-IR spectrometry Wheat germ 5S RNA was isolated and purified as described in Chapter II. The preparation of 5S RNA for FT-IR measurements is outlined in Figure 27. Basically, wheat germ 5S RNA (ca. 10 mg) in the absence of MgCl 2 was prepared by dissolving the RNA in lOmM sodium cacodylate (pH = 7.0), 10 mM ethylenediamine tetraacetate acid (EDTA), and 100 mM NaCl and then dialyzed against the same buffer as above in order to remove Mg++. The solution was further dialyzed 79 Wheat Germ 5S RNA With MgCl2 Dissolve the 5S RNA in Dissolve the 5S RNA in 10 mM cacodyl ate 10 mM cacodylate 10 mM EDTA 100 mM NaCl, 10 mM MgCl 100 mM NaCl, pH=7.0 pH=7.0 Dialyze the 5S RNA solution Dialyze the 5S RNA solution against the buffer same as above against the same buffer as 4°C, 500 Volume, 4h above, 4°C, 1000 Volume, 24h Dialyze the 5S RNA solution against 10 mM cacodylate 100 mM NaCL, pH=7.0 4°C, 1000 Volume, 24h Precipitate with 95% Precipitate with 95% cold ethanol cold ethanol Lyophilize twice from Lyophilize twice from 99.8% D20 99.8% D20 Dissolve in Dissolve in 10 nM cacodylate 10 mM cacodylate pH meter reading 7.0 pH meter reading 7.0 100 mM NaCl, 99.8% D20 100 mM Nacl, 10 mM MgCl2 99.8% D20 \ Ready for FT-IR measurement Ready for FT-IR measurement Figure 27. Diagram of FT-IR sample preparation. 80 against 1000 volume of lOmM sodium cacodylate, lOOmM NaCl, pH = 7.0 for 24h at 4°C. After dialysis, the 5S RNA was precipitated with 95% ethanol and lyophilized twice from a 99.8% deuterium oxide solution. For FT-IR measurements, the 5S RNA was dissolved in lOmM sodium cacodylate, pH meter reading 7.0, and lOOmM NaCl, in 99.8% D 2O. Wheat germ 5S RNA in the presence of MgCl 2 was prepared by using the same procedure except that the f ir s t step (in the presence of EDTA) was omitted and lOmM MgCl 2 was included in all buffers. The concentration determination procedures were followed according to Chapter I I . The final wheat germ 5S RNA concentration in the absence of Mg++ is 35.7 mg/ml, whereas i t is 25.5 mg/ml in the presence of Mg++. 2. Fourier transform infrared spectroscopy a. Path length measurement. For quantitative work it is essential to obtain accurately the optical path length of each cell so that comparison can be made between substances not examined in identical c ells. The path length of the empty cell can be measured very accurately by use of the interference fringes (channel spectra) caused by reflection of a portion of the lig h t from the internal surfaces of the cell windows (Figure 28). Reflected and unreflected light are combined and the phase difference between the two results in a secondary centerburst in the interferogram, which in turn results in a sine wave in the transmittance spectrum. The fringe method is carried out using the following equation: iue 8 Te esrmn o cl pt lnt i cretd y the by corrected is length path cell measurement The of 28. Figure o o 81 5n d(mm) ------t (3.9) (vi - Vg) where and V 2 are the starting and finishing wave number values, n is the number of fringe maxima between v^ and V£ ( vj_, v 2 in wave number). In Figure 28, there are 5 maxima between 2280 cm'l and 1720 cm“l. A cell thickness of 0.0446 mm is then calculated. b. Measurement of IR spectra. The infrared absorption spectra of wheat germ 5S RNA were measured at 1 cm-* resolution in cells with CaFg windows and 0.044 mm path length. WG 5S RNA spectra at the elevated temperature were obtained by wrapping the stainless steel cell holder with coils of nichrome wire connected to a dc power supply. Asbestos tape was used to insulate the nichrome wire. Sample temperature was determined by using a calibrated iron-constantan thermocouple attached directly to the infrared c e ll. Spectra in the 1000-4000 cm~l region were detected with a Digilab FTS-14D interferometer fitted with a standard triglycine sulfate pyroelectric detector. For each spectrum, 256 digitized i nterferograms were accumulated and then Fourier transformed to give a single-beam spectrum. A single-beam reference spectrum was acquired from 10 mM sodium cacodylate (pH meter reading 7 .0 ), and 100 mM NaCl, in 99.8% D 2O. Each absorbance spectrum was computed from the logarithm of the ratio of wheat germ 5S RNA to buffer at the same temperature, as illustrated in Figure 29. Absorbance spectra were corrected for small differences in water vapor content (due to slight 83 Figure 29. Generation of FT-IR absorbance spectrum for wheat germ 5S RNA (C) from single-beam transmittance spectra of wheat germ 5S RNA solution (A) and buffer (B). Each absorbance data point, A(v), in (C) is obtained from A(V) = lo g io d (V)RNA " T( v^buffer^ in which T(v)rm and T(v)bUffer are the transmittances from (A) and (B). The small spikes visible in (A) and (B) arise from water vapor due to incomplete purging of the spectrometer. Finally, different HDO concentrations in sample and buffer (compare 1457 cm-1 region in (A) and (B}) would produce an interfering absorbance band in (C). That interference has been removed by absorbance subtraction, using an HDO absorbance spectrum obtained from transmittance spectra of 99.8% and 97.3% D 2O solutions at that temperature. ABSORBANCE INTENSITY 40 60 20 60 1 1400 FREQUENCY FREQUENCY iue 29. Figure c 1)(cm 84 85 differences in purging the sample compartment) and for small differences in HDO content in the sample and reference buffers, as described in the legend for Figure 29. D. Results and Discussion 1. FT-IR spectra. It is necessary and interesting to show the three reference spectra. Figure 30 shows the infrared spectra for unpaired and paired base A and U. The difference spectrum (A )-(B ), is taken to represent the infrared extinction due to A*U base pairs. Figure 31 shows similar spectra for G*C pairs. Finally, Figure 32 shows FT-IR extinction coefficient spectra of the four ribonucleotide 5 '-monophosphates. The reason for including these reference spectra is to point out that only 5'-CMP exhibits significant absorption at 1505 cm“l . FT-IR spectra for wheat genm 5S RNA in the presence and absence of Mg2+ are presented in Figure 33 and Fig. 34, respectively. FT-IR spectra of yeast tRNA closely resemble those previously reported from dispersive IR (194) and are not shown. 2. Determination of RNA base-pair composition from FT-IR absorbance The theoretical basis for calculation of RNA base-pair compostion has been given in section B.2. However, i t is necessary to sum up those assumptions which are made during the derivation. The assumptions are as follows: 86 Abs= 0 .0 7 5 . 5'a MP ♦ 5'UMP 0.5 E Abs = 0 .0 7 o 0 .5 poly-rA : poly- rU ro O x 00 0 .7 5 Difference (A-B) 0.25 -0 .2 5 1400 1500 1600 1700 1800 FREQUENCY (cm"') Figure 30. Determination of infrared extinction coefficient versus frequency for A*U base pairs. (A), Spectrum consisting of 1/2 the sum of the spectra of 5 ‘-AMP and 5'-UMP. (B), Spectrum for double-stranded poly(rA)*poly(rU). (C), Difference spectrum, (A )-(B ), representing the infrared spectrum of an A*U base pair. All extinction coefficients are normalized per mole of phosphate. 87 Abs = 0.065 5'GMP ♦ 5'CMP 0 .5 E Abs * 0 .0 8 o ~ 0.5- poly rGipoly rC ro O >< CO 0 .5 Difference (A-B) -0 .5 1400 1500 1600 1700 1800 FREQUENCY (cm-') Figure 31. Determination of infrared extinction coefficient versus frequency for G-c base pairs. (A), Spectrum consisting of 1/2 the sum of the spectra of 5'-GMP and 5 '-CMP. (B), Spectrum for double-stranded polyfrG)*poly(rC). (C), Difference spectrum, (A)-(B), representing the infrared spectrum of a G*C base pair. All extinction coefficients are normalized per mole of phosphate. 88 Abs=O.I3 . 5' GMP Abs = O .I3 5' CMP E o Abs = 0.15 ro i 5 'AMP O 1.0 Abs= 0.15 5' UMP 1 4 0 0 1500 1600 1700 1800 FREQUENCY (cm-') Figure 32. FT-IR reference spectra of the four ribonucleotide 5 1-monophosphates. 89 Figure 33. FT-IR spectra of wheat germ 5S RNA in the absence of Mg++. (A) FT-IR spectrum of wheat germ 5S RNA at 90°C. (B) FT-IR spectrum of wheat germ 5S RNA at 30°C. (C) Difference spectrum, or base-pair spectrum, (A) - (B), representing the infrared intensity due to base pairing of the RNA bases. (D) Simulation of (C), computed from the reference A*U and G*C base-pair spectra using base pair numbers listed in Table 2 (14 A*U and 17 G-C pairs). (E) Difference spectrum, (D) - (C), representing the difference between experimental and simulated base-pair content. -0.125 ^ X 10 3(M“'c m “ 0. 5 .2 -0 0. 5 .2 -0 0.125 0.25 0.25 0.25 0.75 0.25 0.75 0 0 4 1 Wheat germ 5S RNA + Mg + RNA 5S germ Wheat 1500 FREQUENCY FREQUENCY iue 33. Figure 601700 1600 C "1) (CM 0 0 8 1 90 91 Wheat germ 5S RNA 0 .7 5 0.25 0.75 o 0.25 rO 0.5 0.25 0.25 0.125 - 0.I25L 1400 1500 1600 1700 1800 FREQUENCY (CM-1) Figure 34. FT-IR spectra of wheat germ 5S RNA in the presence of Mg++. (A), (B), (C), and (E) as in Figure 33. (D) Simulation of (C), computed from the reference AU and GC base-pair spectra using base pair numbers listed in Table 2 (14 A‘ U and 18 G*C pairs). 92 a. All RNA bases are either paired or unpaired. b. There are only two possible types of base pairs, A'U and G*C (e.g., no G-U pairs). c. Base stacking in single-stranded RNA segments are ignored. d. At 30“C, a ll base pairs in the native wheat germ 5S RNA are present. e. At 90”C, a ll bases in wheat germ 5S RNA are completely unstacked and unpaired. f. Any te rtia ry base pairs behave as Watson-Crick pairs. Among these assumptions, at least item b may introduce a large error. In the case of tRNA, since there are only very few G*U base pairs present in the molecule, the error can be truly ignored. However, in the case of 5S RNA, most proposed models include five to seven G*U pairs. This may result in a 20% to 30% errors in the case of the number of G*C pairs. I t is obvious that the formula which was derived from section B.2 has to be modified. The equations 3.13 and 3.14 are no longer valid in the presence of G-U base pairs. With this consideration, the equation can be rederived and represented: e(v )RNA»90uC “ £(v )RNA,30“C 2 e ( v )a . u + c ( v ) u , u = L - ...... ------v)A.yJ X fA.y 2 e ( v )q , u + e(v)c»U + L — ...... - e(v)G.cJ X fG.c 2 e (v )g ,u + e( v )u , u + [ ------r ( v )q .yJ x f q. jj (3.24) 93 The le f t hand side of Eq. 3.24 is directly measurable. I f the various bracketed expressions on the right were available from suitable reference compounds, then accurate measurement of the extinction coefficients at three IR wavelength would suffice to determine the desired fractions of each base pair present. Although self-consistent values for the extinction coefficients associated with paired and unpaired A*U and G*C can be assigned from reference spectra of the two corresponding poly( rA)-polyl rll) and poly( rG) *poly( rC) duplexes and the four ribonucleotide monophosphates: 5'-AMP, 5'-UMP, 5‘-GMP, and 5'-CMP, there is as yet no reliable estimate for the IR hyperchromism of a G*U base pair because poly(rG) does not pair up with poly(rll) when the two homopolymers are mixed. There are two possible approaches to the G*U problem. The simplest is to assume that the e(v) at every wavelength changes by the same amount for a G*U pair as for a G*C pair. In that case the second and third terms in brackets on the right of Eq. 3.24 are equal, leading to: £ ( V)RNA»90UC - e( v)RNA,30oC 2 e(v)A,u + e(v)u,u = [ - e (v )A.(jJ 2 e (v)G’U + e tv )C»U + (_ ------_ e ( v) q.cJ x ( fq• q + fQ-U) (3.25) 2 This calculation w ill lead to a much higher content of G-C in the molecule than is actually the case. 94 A better approach is suggested by close inspection of the spectra of the four ribonucleotide monophosphates in figure 32. Since only 5'-CMP exhibits significant IR absorbance at 1505 cm-*, 5(1505) should be affected by G-C stacking but not by A*U or G*U stacking. Thus, the fraction of G*C pairs can be deduced from the RNA hyperchromism at 1505 cnr* alone. The hyperchromism at 1620 cm-! is due to mainly to A*U stacking with some G*C stacking contribution. Since fQ.Q is now known, as is the G*C contribution to hyperchromism at 1620 cm-* (middle term in brackets on the right of Eq. 3.24), f/\-u can be computed from the 1620 cm“l hyperchromism. F in ally, a ll three base-pair types (A»U, G*C, and G*U) are expected to contribute to the hyperchromism at 1575 cm“l . Since fg.Q and f/\.u are now known, fg.y can be obtained from (fg-c + ^G’ U^ computed from hyperchromism at 1575 cm“l with Eq. 3.25, under the assumption that£(v)g.Q =£(v )g. u at 1575 cm"l only. The base pair numbers resulting from these calculations are listed in Table 2 along with values determined by other techniques, and the melting midpoints (Tm) deduced from th e ir temperature dependence (Figure 35) are listed in Table 3. 3. Discussion. Figure 33 and 34 illu s tra te the good agreement between the experimental RNA hyperchromism and that simulated from linear combination of reference spectra in the proportion listed in Table 2. The r e lia b ility of the data reduction is further evidenced by the close agreement between the relative and absolute numbers of A*U and 95 Table 2. Number of secondary base pairs in wheat germ 5S RNA and yeast tRNAP^e. Precision of experimental values is 10%. Species Source A-U G-C G-U Total Yeast X-ray and *H NMR 8 12 1 21 tRNAPhe FT-IR 7.4 11 — 18.4 UV Hyperchromism 15 21 (G-C + G-U) 36 t -Mg++ 14 17 5 36 FT-IR Wheat +Mg++ 14 18 3 35 Cl overleaf Model 12 17 6 33 Revised Fox/Woese Model 9 17 5 29 Studnicka et a l ./ Nishikawa & Takemura Model 10 20 4 34 t Data from (190) iue 5 Nme o AU o ad * (• ) ae ar o wet germ wheat of pairs base ) • ( G*C and (o) A*U Number of 35. Figure Base Pairs 20 20 30 - 0 1 IQ- - - S N a fnto o tmeaue (o) i the in (top), temperature: of function RNA5S as rsne f g+ (otm, n h asne f Mg++. of absence the in (bottom), Mg++; of presence 0 0 0 0 0 0 0 90 80 70 60 50 40 30 20 ~Cr eprtr (°C) Temperature ha Germ RNA 5S Wheat Wheat Germ RNA 5S O MgCl2 NO 10 MgCl2 mM 96 97 Table 3: Melting midpoints (Tm)(°C) of wheat germ 5S RNA. Technique Type of base pair -Mg++ +Mg++ UV hyperchromism A-U, G-C, G-U 54 69 t Circular Dichroism A-U, G-C, G-U 48 63 t FT-IR A-U 54 58 G-C 56 76 t Data from (190). 98 phe G-C pairs of tRNA determined by FT-IR and by high fie ld proton Phe FT-NMR (Table 2). In fact, tRNA is a severe test case, because of the large number of non-Watson-Crick base pairs and derivatived bases: there are 9 normal G*C and 7 normal A*U pairs, 3 G*C base pairs with methylated G or C, and 1 A*^ base pair. Further evidence for self-consistency is the close agreement between the base-pair ratio: G*C/(G*C + A*U) = 0.58 (UV absorbance) vs. 0.56 (FT-IR) for wheat germ 5S RNA in the presence of Mg++. Because FT-IR can distinguish between A-U and G*C base pairs, i t is possible to monitor independently the melting of these two types of pairs (Figure 35). It is interesting to find that the stability of A*U pairs increases only slightly on addition of Mg++ (Tm increases from 54 to 58°C), whereas the G*C s ta b ility increases by much more (Tm increases from 56 to 67°C). The remaining question, however, with those assumptions described in section D .2., is how to ju s tify the accuracy of the measured number of A*U, G*C, and G*U base pairs. Although the spectra of appropriately prepared 1:1 mixtures of poly(rA)*poly(rU) (Figure 30) and poly(rG)*poly(rC) (Figure 31) probably give accurate description of A*U pairs and G*C pairs, it is not so easy to provide accurate models for unpaired A, C, G, and U bases. For example, the heat-denatured homopolymers are not suitable, poly(rG) exhibits some residual base stacking even at 90°C, as evidenced by its UV absorption vs. temperature p ro file . Therefore the mononucleotide 5 '-phosphates were chosen as reference spectra to represent unpaired nucleotide bases in RNA. 99 Since mononucleotide 5 '-phosphates reference spectra do not re fle c t neighbor effect and dependence of IR hyperchromism on sequence, (e .g ., the homogeneity of the poly(rA)*poly(rU) and poly ( rG)*poly(rC) reference polymers is totally different from the heterogeneity of the 5S RNA sequence arrangement), the theoretical treatment may not f i t the real experimental situation and w ill introduce error (+10$) (195). As far as the secondary structure models of WG 5S RNA are concerned, FT-IR technique provides a powerful method to estimate the variety of the proposed models. Figures 12, 13, and 14 show three popular secondary base-pairing schemes for eukaryotic 5S RNA, each adapted to the WG 5S RNA sequence. The base-pair content for each model is listed in Table 2, for comparison to experimental UV, CD, and FT-IR results. The experimental results give a base-pair total slightly higher than predicted by the models but with similar distribution among A*U and G*C pairs. Since the models do not include tertiary pairs, one would expect to find more pairs experimentally than in the models. The secondary base-pairing model that most closely matches the experimental base-pair totals and base-pair ratios is the cl overleaf (69,70), although experimental impression does not rule out the modified Fox-Woese model (72). Because all three optical techniques (UV, CD, IR) re fle c t base phe stacking and/or helicity and because the x-ray structure of tRNA proves that single-stranded bases can be stacked without being paired, the experimental base-pair estimates in Table 2 should be taken as upper lim its. Proton FT-NMR and proton intra-molecular NOE experiments at 500 MHz should establish lower lim its, since base-pair hydrogens that exchange rapidly with HgO may be broadened beyond detection (see chapter V). Combination of the optical, DSC, and *H NMR results should go far toward establishing the correct secondary structure for ribosomal 5S RNA in solution. CHAPTER IV THEORETICAL AND EXPERIMENTAL ON-LINE ANALYSIS OF MULTI-STATE MELTING OF WG 5S RNA BY DIFFERENTIAL SCANNING CALORIMETRY A. Introduction. An interesting property of biological macromolecules is their ability to undergo structural changes with temperature. Lipids in aqueous environments undergo gel-liquid-crystal transitions, proteins undergo unfolding or denaturation transitions, and base-paired nucleic acids unwind. These transitions have proved to be highly cooperative in nature and hence are sharply influenced by solvent. Calorimetry suggests its e lf as an ideal technique for measurement of thermodynamic parameters associated with such phenomona. In nucleic acid research, especially for 5S RNA as in this case, adiabatic scanning microcalorimetry is a useful method to get data for the enthalpy change,AHeXp, which is required to melt the compact native conformation of the WG 5S RNA to a random co il. Also, DSC can measure the molecular free energy which is a major thermodynamic parameter of the s ta b ility of the molecule. By these measurements, a comparison can be made between the experimental data and theoretical prediction for different melting enthalpies (AH^h) and stabilities according to Tinoco rules which are from study of a number of 101 102 oligonucleotides (151, 154). Based on these rules, the models which have the largest differences between the experimental and theoretical values^Ht h and^G-t^ may not present a favorable structure for the 5S RNA molecule. On the other hand, DSC measurement can also be used to check the theoretical values of enthalpy and free energy of base pairing in nucleic acids. This chapter then w ill focus on the theoretical melting curve derivation, model compounds testing, and their application to the structural analysis of wheat germ 5S RNA. B. Theory 1. Temperature-dependence of a two-state equilibrium. An assumption is made that each component melting process can be treated as a two-state equilibrium, in which A and B are the unmelted and melted forms, respectively. Let A = B and Keq = [B]/[A] (4.1) where Keq = exp ( -AH/RT) exp(As/R) (4.2) in which Keq is the equilibrium constent, AH and AS are the molar enthalpy and entropy of melting, R is the gas constant, and T is the absolute temperature. The fraction melted, 0, can then be defined as 9 = [B]/( [A] + [B]) (4.3) It is readily shown that 0 = Keq/(1 + Keq) (4.4a) 103 so that 1 - 0 = 1/(1 + Keq) (4.4b) 0 Then from Eq. 4.4b y ield Keq = (4.5) 1-0 According to Van't Hoff equation it can be shown that 3 In Kea ah q - (4.6a) 3 T RT2 Substituting Keq by eq 4 .5 , we have 0 3 In 1-9 Ah (4.6b) 3T RT2 Taking the partial derivative of Eq. 4.6b i t can be w ritten as: 9 0 AH 0 (1 - 0) (4.6c) 3 T RT2 Combining Eqs, 4.2, 4.4a, 4.4b, 4.6c, and assumingah and as are independent of T, then the temperature dependence of 0 is given by 1 1 exp [ ( AH/R) ( ------—)] d0 AH T Tm (4.7) dT RT2 1 1 (( 1 + exp[(- Ah/R)(------) ] ) 2 T Tm where Tm is the melting midpoint: when T = Tm, then 0=1/2, Keq=l, and AS= AH/Tm. 2. The relationship between d0/dT and the DSC experiment. The heat capacity at a constant pressure, Cp, is given by the equation dH-j dqD( i ) dqp Cp =1 ni Cp( i ) = I n-j— = I ------= — (4.8) i i dT i dT dT 104 in which H-j is the molar enthalpy, n-j the number of moles, and Cp(i) the molar heat capacity for the ith component, and dqp(i) is the heat required to raise the temperature of the ith component by dT degrees at constant pressure. If the molar enthalpy of melting, 4H, is independent of temperature, then for two-component system of A and B, dT d0 = n [(1 -0 ) Cp(A) + OCp(B)] + AH ( — )] [4.9] in which n = nA + ng, n0 = ng as the number of moles of A that has melted to become B, n(1-9) = nA which is the number of moles of unmelted A remaining, also Cp(A) and Cp(B) are the molar heat capacities of components A and B. Heat capacity of the aqueous buffer solution is compensated by the reference solution in this DSC experiment, and may be omitted. In DSC experiments, the power required to keep the sample temperature the same as the temperature of the reference buffer solution is recorded as the temperature of the system increases linearly with time. The output of the instrument is thus dqp/dt versus t . However, since dT/dt is assumed to be constant, dqp/d t is readily converted to dqp/dT such that: dqp dqp/dt [4.10] dT dT/dt The Cp term in Eq. 4.9 simply determines the baseline of a plot of dqp/dT vs. T, and w ill be omitted for the time being. Eqs. 4.7 and 4.9 then may be combined to give a new equation: 105 exp[(-AH/R)( - - - )] dqD n(dH)2 T Tm — = ------L4.11J dT RT2 1 1 2 1 + expL( -^H/R)( - - - )J T Tm The area under a plot of dqp/dT vs.. T is readily derived from Eq. 4.11. So that: r T 2 Area = J dqp - n Hmelting [4.12] T1 in which T^ << Tm and T2 >> Tm to ensure that all enthalpy contributions are recorded within the observed temperature range. 3. Simulation of DSC melting curves The right-hand side of Eq 4.11 is used to f i t experimental DSC plots of dqp/dT V£. T, in order to obtain initial estimates of Tm andzlH for such f it s . I t then becomes useful to derive theoretical expressions for the height and width of the DSC peak for atwo-state melting transition. The maximum DSC peak height is found by setting the first derivative of Eq 4.11 to zero, d dcln d n WH)2 — (—£) = - L ( ------— ) 0(1 - 0)J =0 14.13] dT dT dT rj2 Solving for 0, 1 RT r 0 = ------[4.14] 2 dH The exact temperature at the DSC peak height maximum can be obtained by substituting Eq. 4.14 into Eq. 4.6 and solving for T. However, since biological polymers typically melt in a range from 290K to 350K, with enthalpy values varying from 10 to 150 Kcal per 106 component, a typical value for (RT/a H) is (2 cal M"1 K“^)(320 K) /(40,000 cal M"l) which equals 0.016. Thus, to a good approximation (namely, 3.2$ for 8 and 0.38 K for Tm in this example), values for 0 and T are given as: 1 0 = at (dqp/dT)niax L4.15J 2 at (dqp/dT)max L4.16J j ^ P j n (a h )2 and L4.17J dT max 4RTm2 The same low level of error as in Eqs. 4.15-4.17 is obtained when the full width at half maximum height of such a plot is given by T = 3.52 L4.18] For example, for a true Tm = 320K and ah = 40 kcal nr*, Eq. 4.18 gives AT = 17.9 K, compared to the true value for at = 17.8 K from numerical iteration of Eq. 4.11. Similarly, the true maximum of dqp/dT vs. T does not occur directly at Tm but rather at Tm-0.38. The principal parameters to be extracted from DSC analysis are Tm (half-melted temperature) and a h (enthalpy of melting) for each component DSC peak. The Tm and a h can be obtained by fittin g a theoretical two-state melting DSC curve to Eq. 4.11. To a good approximation, Tm is given by the temperature at the observed maximum of a plot of dqp/dT vs. T. However, there'are at least four ways to estimateAHs including (a) curve fittin g of the entire DSC curve to Eq. 4.11; (b) integration of the full DSC curve; (c) combination of observed DSC peak hight and center using Eq. 4.17 (155); and 107 (d) combination of the observed DSC peak width and center using Eq. 18 (corrected for the approximation in an earlier work (198). Close agreement between (b) and any of the other estimates constitutes strong evidence for a single two-state phase transition. C. Using model compounds to test the two-state model and computer software. 1. Microcalorimeter. There are a number of d ifferen t types of calorimeters that be generally classified as differential scanning calorimeters. Several commercial instruments of moderate sensitivity have been available; such as the popular Perkin-Elmer DSC-2. These instruments are excellent for many studies involving neat samples or samples at very high concentration since they can resolve changes in heat capacity of the order of 1 part in 100. Because of their small sample size (ca. 40 ul) and the lack of s tric t adiabatici ty, however, the sensitivity is not high enough for many types of biological studies (e .g ., nucleic acid research). In more sophisticated differential heat capacity instruments, thermally induced changes in heat capacity as small as 1 part in 10,000 can be detected above the noise level. With such high sen sitivity, the transitions of most biological samples can be studied in the concentration range of 0.01 to 1.0% (0.5% for 5S RNA). A block diagram of such a calorimeter is shown in Figure 36. A matched pair of cells (Hastelloy C-276) of 1 ml capacity are each 108 Cell Power Supply u Dill power T rim Pols control! o S 1 * X Y Recorder AdioOotic Shield -vw w Temp O Shield Prooe Power Supply -C3- Au» Heoter Sensor /W W v H — 'W 'A A B oth I (,H«oJer_ Figure 36. Block diagram of differential heat capacity calorimeter 1230). 109 fitted with a main and an auxiliary heater. Power input to the two cells is matched by ajusting a pair of trimming potentiometers in parallel to the main heater. The thermopiles detect any off-balance signal between sample and reference. The signal from the thermopiles is amplified and drives a feedback network in order to keep both cells a t the same temperature (_+ 100 microdegrees of n u ll). The current through the auxiliary heater is also fed to a squaring computer during the scan. Since the squared current is proportional to the power dissipated in the lagging c e ll, i t can then be related by an appropriate calibration constant to the difference in heat capacities between the cells. From this a continuous recording of the differential heat capacity with temperature can be obtained. The highly sensitive microcalorimeter used in these experiments was produced by the Micro-Cal company. The absolute heat capacity can not be measured accurately, since the process of emptying, cleaning and refilling the cells introduces errors in heat capacity that are considerably larger than the noise level of the instrument its e lf, also these problems can cause base-line nonreproducibility. However, the absolute capacities of nucleic acids are not considered here, only the thermal induced structural changes are of interest. Hence, from now on no emphasise on the absolute value of the Y axis in the DSC plots w ill be made. 110 2. Experimental measurements using the DSC technique. a. Sample preparations. Cytochrome c and ribonuclease A were chosen as model compounds (from Sigma and Boehringer Mannheim). These two proteins were chosen because f ir s t they have previously been shown to melt reversibly, with large AHme] t j ng values, in what appears to be a single stage (158), and also because they are not expected to interact when mixed together at these concentrations. DSC samples were prepared according to a previously described method (158). Protein concentrations were determined by UV/vis absorbance on a Beckman DU-8 spectrophotometer by using the known extinction coefficients for the two proteins. The cytochrome c sample was 0.64% w/v, and the ribonuclease A sample was 0.88% w/v. A third sample containing both proteins was 0.30 umol ml-1 in each. Protein aggregation was minimized by adjusting pH (MI412 microcombination pH probe) to 4.5 at 24°C in 0.04 M glycine buffer. b. DSC data collection. DSC measurements were conducted according to the manufacturer's recommended procedures (Microcal MC-1 differen tial scanning microcalorimeter, equipped with a Welch Duo-Seal Model 1400 vacuum pump, and liquid cells of 0.7 ml effective volume). Each protein sample was referenced against the same volume of buffer. The reported scan rates were measured d irectly, and were generally 20% larger than the nominal scan rate reading from the MC-1 control panel, depending on the "baseline shift" shifting. Once an MC-1 scan begins, all Ill remaining operations are controlled from an Apple I I Plus microcomputer (64K RAM, with an Apple Super Serial card). Although i t is not essential, the use of a Saturn Systems Inc. Accelerator II card speeds computations by a factor of three. The Apple waits until a previously specified lower temperature (15°C) has been reached, and then begins acquiring data. At specified temperature intervals, the thermocouple (x) and differential heating output (y) voltages from the MC-1 are conditioned to 0-2 volt, digitized by an 11-bit analog-to-digital converter, and sent to the RS-232 port of an Apple I I plus microcomputer (229). A BASIC program for the Apple then scales each voltage-voltage pair to a temperature-cal mol-1 deg"1 pairs, and stores the complete data set on a floppy disk. c. Computer software. Because of the limited working memory of the Apple II+ , not all of the desired operations can be called from a single program resident in core memory. Therefore, a menu is provided for selection of individual interactive Applesoft BASIC language programs for data acquisition (DSCACQ), curve simulation (DSCCURFIT), and plotting (DSCPLOT). During system start-up, the Accelerator I I card is activated and shape tables of alphanumeric characters and symbols are loaded in memory addresses 36946-38400, for later use in labeling of graphs. High memory is then set below 36946. A menu is then installed to direct traffic among the various programs (229). 112 DSCACQ captures digitized MC-1 signals from the A/D converter, and saves the data for late r manipulation. DSCCURFIT performs curve simulation and deconvolution of the acquired data. DSCPLOT plots curves in a format defined by the user. I t can also superimpose experimental, simulated individual component, and/or a sum of simulated component curves, and perform screen dump to an Integral Data Systems Prism p rin te r/p lo tte r. All the supporting subroutines are based on published algorithms, they include First, Marquardt's CURFIT algorithm (196) which has been adapted to f i t experimental DSC data to Equation 4.11, second, Savitzky-Golay seven point data smoothing which is based on a cubic fitting function (197), third, a baseline tilting, in which a line is drawn between the averages of the five left-m ost and five right-most Cp-values in the selected temperature range (see section D .2., Figure 41B). Finally, Simpson's rule with a quadratic fitting function (196) can be used for numerical integration of the DSC curve. d. Extaction of Tm and 4H from experimental DSC curve. Figures 37a to 39a show experimental DSC curves for dilute (<1% wt/vol) aqueous solutions of cytochrome c, ribonuclease A, and an equimolar mixture of the two proteins. The high sensitivity of the MC-1 instrument is apparent from the good signal-to-noise ratio obtained from the raw data for such polymers. This high sensitivity is accomplished at the expense of sample size: each scan requires about 0.7 ml of sample, compared to about 0.04 ml for most other commercial DSC instruments. 113 • 50 60 70 TEMPERATURE r\ o \ HI _i a E _l a o * V a. o 55 60 65 70 75 80 85 90 95 TEMPERATURE <*C) Figure 37. DSC curves for Cytochrome c. (a) Digitized raw data, with uncalibrated y-axis. The scan rate was 71.6°C/hr, with a y-axis sensitivity setting at maximum, (b) Dotted line represents the data which has been scaled, smoothed, and with the baseline flattened. The solid curve is the best-fit to two-state melting process (Eq. 4.11), with the f i t parameters listed in Table 4. 114 Compared to sim ilar curves obtained from direct analog output to an x-y recorder, the baseline of the digitized DSC data is flatter. In analog detection, the x-y recorder time constant is large in order to reduce noise, and the baseline and peak shapes can be distorted during the DSC scan. In contrast, no analog filte rin g is required when the DSC data are routed directly to the microcomputer, because the data can be smoothed digitally after acquisition. The elimination of baseline distortion is a major advantage of the digitized data acquisition process. The baseline of an experimental DSC plot almost always exhibits a non-zero slope. Apart from small differences in specific heat capacity between sample and reference, or between the unmelted and melted polymer, any difference in absolute mass of the sample and reference will require a differential power required to raise the temperature of the two cells by the same amount. It is therefore necessary to t i l t the baseline of experimental DSC curves, using reference temperatures near the low and high temperaure lim its of the scan. In addition, data smoothing is appropriate when (as in these cases) the thermal phase transitions are relatively broad (10-20°C). The dotted curves in Figures 37b-39b show the effect of baseline t iltin g and seven point Savitzky-Golay smoothing on the raw data of Figures 37a-39a. The reduced DSC data of Figures 37b-39b were then subjected to CURFIT analysis based on Eq. 4.11. The results are shown as the solid curves in Figures 37b-39b, and the b e s t-fit parameters are listed in Table 4. zlH obtained from the integrated area under the DSC curve is 115 a 56 a. 35 35 40 45 50 55 60 65 70 75 80 85 90 95100 TEMPERATURE <*C> 1 5 F (j \ Ui o \ V 0. O 45 50 55 60 65 70 75 TEMPERATURE <»C> Figure 38. DSC curves for ribonuclease A. The format is the same as in for Figure 37. The scan rate was 72.0°C/hr. 116 Table 4. Melting midpoint (Tm) and molar enthalpy of melting (4H) were obtained from theoretical best fits of Eq. 4.11 to experimental DSC curves for dilute aqueous protein solutions. Samples Tm <°C> a Ha, kcal/mole 4Hb, kcal/mole Cytochrome c 75.3 99.3 92.7 Ribonuclease A 62.2 116.2 111.9 Equimolar mixture 60.3 and 199.6 107.4 of both proteins 76.3 a. From DSC area. b. From best fit to Eq. 4.11. 117 compared to 4H from a best f i t of the DSC curve to Eq. 4.11, as a test of the two-state melting model. The fitting algorithm typically takes about 2 min per iteration, and reaches a final best-fit (variance constant to within about 1 part in 10^) after four to six iterations. For cytochrome c or ribonuclease A alone, the "true" AH obtained from the DSC area is 4-7% larger than obtained by fittin g the data to a theoretical two-state model. This small but real difference is in the same direction for both proteins, and can be ascribed to a combination of: presence of multiple DSC peaks (due to multi-stage melting); invalidity of Equation 4.11 (due to cooperative melting, scan rate too fast to permit establishment of equilibrium at each temperature, variation of £Hme-|ting temperature); and a difference in heat capacity between unmelted and melted forms (change in baseline slope on passing through the DSC peak). Such problems have been noted in e arlie r calorimetry experiments (156, 199). More important there is an agreement between the Tm and/iH values obtained for the two-protein mixture as for the individual proteins alone. Since Tm and^lH for both proteins are highly pH-dependent (158), the observed discrepancies are most lik e ly due to small differences in pH between the individual protein samples and the mixture. D. DSC analysis of WG 5S RNA. So fa r, several techniques previously tested on tRNA have been applied to 5S RNA: UV, CD (125, 164), 0RD(165, 166), FT-IR(128), 118 1 0 0 F RNase A a _ Cyt \ w 90^ _J a e \ S eof a- . .-si** CJ J______I I______I______L io 40 50 60 70 80 90 100 TEMPERATURE <»C) A o \ UJ o X. \ 0. o 40 45 50 55 60 65 70 75 80 85 90 TEMPERATURE (»C) Figure 39. DSC curves for equimolar mixture of cytochrome c and ribonuclease A. The format is the same as in Figure 37. The scan rate was 71.5°C/hr. The experimental curve has been simulated by the sum of two b e s t-fit components (computed from Eq. 4.11). Raman (69, 70), ESR( 148) as well as 1H, 19F, 31P NMR. Along with th is , DSC studies can add new data to help solve this d iffic u lt problem (159). Each different techniques provides its own original approach to this problem and hence leads to many informative structural features of the 5S RNA molecule. Even with a ll these, the unique DSC measurement can detect all structural energy features whose dissociation requires energy (e.g., base paired hydrogen bonds, base stacking, metal ion mediated intrachain phosphate links, and interaction between backbone and backbone etc.) and thus provides a direct measurement of the total thermodynamic s ta b ility of 5S RNA for comparison of the secondary structural models. 1. Materials and Methods. Wheat germ 5S RNA was isolated and purified as has been discribed in Chapter I I , section B of this work. For DSC sample preparations, four different salt conditions were designed, (A) no Na+ or Mg++ were added; (B) 100 mM NaCl; (C) 10 mM MgCl 2 ; (D) 100 mM NaCl and 10 mM MgC^. Each buffer contains 10 mM sodium cacodylate, with a pH of 7.0. Approximately 30 mg of 5S RNA was divided into four aliquots, each placed in small dialysis tubing (1 ml volume, Spectrapor membrane tubing 2, molecular weight cutoff= 12.000-14,000). Samples with Mg++ were dialyzed against their corresponding buffers. Samples without Mg++ were prepared by dialysis f ir s t against buffer containing 10 mM Na2EDTA and then against buffer A or B. All dialyses were repeated 4 times at a volume ratio of 1,000:1, for 6 hours 120 periods, stiring at 4°C. Each final dialysis buffer was used as the reference sample for calorimetry measurements. DSC samples of 5S RNA were 98% pure according to slab gel electrophoresis, and gel scan techniques. RNA concentrations were determined as described in Chapter I I , section D. The four sample concentration were found to be; (A), 1.59 x lO'4 M, (B), 1.72 x 10-4 M, (C), 1.14 x 10-4 M, (D), 1.67 x 10-4 M. DSC data acquisition and analysis were as described in section C of this Chapter. For each scan beginning at 15°C, i t is necessary to bring the whole system f ir s t to 0°c. Since the instrument requires a relatively long response time, reliable information in low temperature region can be lost. This matter is illustrated in Figure 40. In the experiment, a sample of water against water as reference was used. The scanning rate was 60°C/hr. Before one pushes the "RUN" botton, every item was checked according to the operating manual. I t can be clearly seen that the temperature range of 10°C-20°C (corresponding to 10 min of scan time) is the response period of the DSC system. If better baseline and low temperature (0°C-2°C) measurements are desired, this factor certainly must be taken into account in order to avoid data acquisition during the equilibration period. POWER (mCAL/MIN) 20 25 - - — iue 0 Ilsrto o te S sseai rsos time. response DSC systematic the of Illustration 40. Figure 10 20 30 EPRTR (°C) TEMPERATURE 40 50 60 70 80 90 O P H-* 122 2. Results a. DSC profiles of wheat germ 5S RNA. Wheat genm 5S RNA melting profiles for four different salt conditions are shown in Figure 41. These curves represent raw data (i.e., unsmoothed, with no baseline correction). The minor variation in the y-offset from scan to scan is due to differences in baseline offset voltage and does not affect the peak profile. A common feature of all four curves is a decrease in baseline slope on passing through the melting transition, as has been previously observed and described in earlier work (174, 176). The heat capacity of the unmelted form increases with temperature, but the heat capacity of the melted form is independent of temperature. Thus, for each two-state melting process, the baseline slope will change according to the fraction melted, 0, as a function of temperature. The simplest approximation (and, for multiple unresolved transitions, the only practical choice) is a baseline drawn from the left-most to right-most DSC ordinates at which the data begins to deviate detectably from straight-line fits to the low-and high-temperature slopes (see Figure 41B). The four DSC data sets of Figure 41, after 7-point smoothing and baseline tilting, are re-plotted in Figure 42. CURFIT analysis of the reduced data according to a two-state melting model for each component transition gives the overlapping individual peaks shown in Figure 42, and th eir sum is shown as a solid line for comparison to each experimental DSC curve. 123 Figure 41. Differential scanning calorimetry plots for wheat germ 5S RNA, in 10 mM sodium cacodylate buffer, pH 7.0, for four differen t salt conditions: (A) no added NaCl or MgCl2 ; (B) 100 mM NaCl [see text for treatment of baseline]; (C) 10 mM MgCl2 ; (D) 100 mM NaCl and 10 mM MgCl2 * The calorimeter output has been conditioned and digitized. The y-offset is arbitrary. Scan rates varied from 70.1 to 72.8 K/hr. The data has not been filte re d or smoothed in any way. O (KCAL/MOLE/°C) Cp (KCAL/MOLE/°C) Q. 200F 180 0 0 3 0 6 1 200 0 4 1 200 120 0 5 1 0 0 50 60 70 80 90 100 1 20 30 40 50 60 70 80 90 100 0 9 0 8 0 7 0 6 0 5 0 4 0 3 0 2 10 0 0 1 0 9 0 8 0 7 0 6 0 5 0 4 0 3 0 2 10 » « E P R T R (C TEMPERATURE (°C) TEMPERATURE (°C) ___ I ___ I l ___ — _ L 1 I40| 0 4 I J 1 i iue 41. Figure I50f 140 200 130 120 220 100 110 180 - 0 6 1 0 100 0 9 0 8 0 7 0 6 0 5 0 4 0 3 0 2 10 ____ | ____ i i i i i i i i i i i 124 125 Figure 42. The data of Figure 41 are re-plotted after 7-point smoothing and linear baseline correction over the transition region. (A) - (D) are as in Fig. 41. In each plot, the thin lines represent the mininum number of two-state component transitions required to fit the experimental curve. The composite dark line is the sum of the component transition peaks. O (KCAL/MOLE/°C) Cp (KCAL/MOLE/°C) . Q 0 7 2 5 4 3 -2 100 0 9 0 8 0 7 0 6 0 5 0 4 0 3 0 2 E P R T R (C TEMPERATURE (°C) TEMPERATURE (°C) iue 42. Figure 0 9 221 10 - 0 0 0 1 0 9 0 8 0 7 0 6 0 5 0 4 0 3 0 2 10 0 9 ro cn 127 The melting midpoint (Tm) and enthalpy of melting (^H) for each simulated component peak are listed in Table 5, for each of the four salt conditions. Provided that the temperature range for simulation was wide enough to include the fu ll DSC composite peak envelope, Tm and AH for the simulated component peaks were not especially sensitive to the low and high-temperature cutoffs used to define the DSC peak. The low-temperature cutoff defines the native state of 5S RNA in solution, and the high-temperature cutoff defines the fu lly denatured state; beyond either lim it, the baseline is nearly straight. Attempts (not shown) to simulate the curve with fewer than four component transitions gave poor fits ; using more than five components gave no significant improvement of the result. b. Reproducibility and reversibility DSC scans for two separate aliquots from the same 5S RNA solution are analyzed in the top and bottom rows of Table 6. The excellent reproducibility of the results is seen from the good agreement (to within 1% in all but one value) between the Tm and AH values for the two scans. Reversibility of the experiment was tested as follows. After aliquot A was scanned from low to high temperature, i t was allowed to re-cool at 0.2 K/min ( i . e . , five times slower than the DSC scan rate), and then scanned again. Comparison of the top and middle rows of Table 6 shows that the scan is indeed reversible, except for a 10% reduction in Tm andAH for the lowest-temperature transition for the re-scan, as seen in Figure 43. Similar results were seen for 5S RNA with no added salt. 128 Table 5. Thermodynamic parameters for multi-stage melting of wheat germ 5S RNA. ^Hevp is the enthalpy of melting obtained directly from the total area under the composite DSC peak. AHSjm is the total enthalpy of melting obtained by adding up the enthalpies (AH) for each of the simulated component DSC peaks. Tm for each simulated DSC component is the melting midpoint for that two-state transition. Also listed are the component (AG298) an(j total (Total AG298) free energies of melting. a G298 = AH(Tm - 298)/ Tm. Sample Na+ Mg++ Peak Tm-273.2 AH ^ Hsim/^8exp AG298 Total AG298 =Ratio 1 26.5 60.5 (kcal/mol) 0.30(kcal/mol) 2 34.7 78.2 2.47 A None None 1.02 10.38 3 39.5 72.6 3.37 4 42.1 78.1 4.24 1 43.1 65.7 3.76 B 0.1 M None 2 50.2 103.1 0.99 8.04 31.48 3 55.3 101.2 9.34 4 61.6 94.5 10.34 C None 10 1 52.3 58.0 4.87 2 64.3 99.3 1.01 11.57 60.09 3 70.3 238.8 31.51 4 78.2 85.8 13.00 D 0.1 10 1 59.3 58.4 6.03 2 65.8 96.5 0.98 11.62 51.09 3 72.1 130.8 17.85 4 77.5 104.1 15.59 iue 3 Rvriiiy f S eprmn fr ha gr 5 RNA germ 5S in wheat for DSC experiment of Reversibility 43. Figure POWER (mCAL/MIN) 20 25 - - xet o te nii mlig ein T 4°) the 40°C), < unfolding (T region reversible a melting l indicating itia superimpose, in curves the for Except process. h peec o 01 NC ad o de M . ( . Mg added no and NaCl 0.1M of presence the rst sa a 6. Kh; ( K/hr; 67.4 at scan t s ir f H 0 2 ------, 30 1 ------ EPRTR (°C) TEMPERATURE 40 1 ------ ------50 1 ------scn sa a 6. K/hr. 67.2 at scan second ) 60 1 ------ 70 1 ------ ------80 ) - h 129 Table 6. Reproducibility and reversibility of DSC scans for wheat germ 5S RNA in the presence of 0.1M NaCl with no added MgCl 2 - Peaks are numbered as in Tables 5. Two aliquots were taken from the same 5S RNA solution. Tm - 273.2 (K) ZlH (kcal/mole) Scan Rate (K/hr) 1 2 3 4 1 2 3 4 Aliquot A Scan 67.4 43.1 50.2 55.3 61.6 65.7 103.1 101.2 96.5 Re-Scan1- 67.2 40.8 50.0 55.2 61.8 59.9 100.9 103.1 93.4 Aliquot B Scan 67.5 43.0 50.4 55.3 61.7 58.3 104.5 99.7 94.4 * The same sample was cooled slowly, and another DSC scan taken. 131 Quite different results were observed for samples containing 10 mM MgClg (with or without 0.1M NaCl). In these cases, a re-scan (not shown) gave a near-flat baseline, indicating that the 5S RNA is irreversibly degraded afte r exposure to Mg++ at high temperature (90°C). Slab gel electrophoresis (not shown) performed on Mg++-containing 5S RNA samples before and a fte r a DSC scan confirmed that the RNA polyribonucleotide chain had been broken in several places. c. Effect of scan rate Figure 44 presents plots of DSC power vs. temperature for three different scanning rates. (The scanning rate was calculated a fte r the data were acquired.) Faster scanning produces larger peaks and thus better signal-to-noise ratio. However, scanning too fast might distort the DSC curve, due either to insufficient transient response from the calorimeter itself, or to insufficient time for equilibrium to be established at each temperature as the temperature changes during the scan. The b e s t-fit analysis of the results for each scan rate are compared in Table 8. The close agreement between the results from different scan rates indicates that the scan rates are sufficiently slow to yield reliable data. 3. Discussion Suppose that a macromolecule consists of n domains, each with different thermodynamic enthalpy of melting, and that those domains iue 44. Figure POWER (m CAL/MIN) 20 25 - - - - 65 /r tp, 74 /r mdl) ad 89 K/hr 48.9 and (middle), K/hr 67.4 (top), K/hr 96.5 hs te rsn rsls r idpnet f cn rate. scan of independent are results present the Thus, S pwr s sa tmeaue fr he sa rates: scan three for temperature, scan DSC vs. power cn ae te uv sae ean te ae se al 7). Table same (see the remains shape curve the rate, scan bto) Atog te inl tegh nrae with increases strength signal the Although (bottom). 30 40 EPRTR (°C)TEMPERATURE 50 + 60 + 70 + 80 + 90 132 133 Table 7. Effect of DSC scan rate upon Tm for component melting transitions of wheat germ 5S RNA in the presence of 0.1 M NaCl with no added MgCl 2 * Peaks are numbered as in Table 5. Melting midpoint, Tm - 273.2 (K) DSC scan rate (K/h) Peak 1 Peak 2 Peak 3 Peak 4 48.9 41.6 50.0 55.3 61.7 67.4 43.1 50.2 55.3 61.6 96.4 41.1 50.1 55.1 61.7 134 melt independently. Further denote a given conformation according to whether a given domain is melted (0) or unmelted (1 ). Next, rank the domains in order of decreasing free energy of melting, with the most stable domain listed f ir s t . For example, a four-domain system for which the two most stable domains are unmelted and the remaining two are melted might be described as ( 1, 1,0 ,0 ), with a total of 2n = 16 separate free energy states. When such a system melts reversibly, all 2n states are in mutual equilibrium, and the system is expected to progress sequentially from lowest to highest energy states (161) ( 1. 1. 1. 1) ( 1, 1,1,0 ) (1, 1,0 , 1) ( 1,0 , 1, 1 ) ...... (0 ,0 ,0 ,0 ) (a) For example, in the (1 ,1 ,1 ,0 ) (1 ,1 ,0 ,1 ) step, the macromolecule has to break a more stable domain and simultaneously re-form a less stable domain. Fortunately, when the domains differ significantly in s ta b ility , only (n+1) = 5 states are significantly populated (159) and the melting process simplifies to: ( 1. 1. 1. 1) (1, 1,1 ,0 ) (1, 1,0 ,0 ) (1 ,0 ,0 ,0 ) (0 ,0 ,0 ,0 ) (b) If the domains do not differ significantly in stability (e.g., Tm values d iffe r by < 3K or so), then the more general mechanism applies. However, in such cases the DSC peaks overlap so closely that they cannot be distinguished reliab ly (161). In other words, reliab le data reduction is possible only when Tm values for successive transitions d iffe r by >5K, so that mechanism (b) applies, and one should not attempt to resolve additional states. I t remains only to decide how best to analyze the DSC scan for such a system into its n component peaks. 135 A method has been proposed (163) that the analysis begin by fitting the highest temperature transition by using the highest temperature portion of that peak. I t is argued that overlap with other peaks is minimal at the highest temperature end of the scan, and that the first transition can thus be fitted most accurately. Assuming that the f ir s t transition f i t is correct, one then proceeds to the next highest-temperature transition, and so on in sequence. Here, the method is used as described in section C.2, in which all of the transitions are fitted simultaneously using a CURFIT algorithm. The CURFIT method has two advantages. F irs t, i t is superior when (as in this case) the lowest-temperature transition is the least reversible. Second, the sequential-fitting method propagates errors which accumulate in proceeding from highest to lowest temperature transitions, whereas the CURFIT approach uses the fu ll DSC curve to judge the quality of the f i t , so that errors are equally distributed among a ll of the peaks. The excellent agreement (+_ 2%) between the total enthalpy of melting (AHeXp) obtained as the area under the experimental DSC envelope and the total enthalpy of melting (iJHs1-m) computed from the sum of the height and width of a Gaussian peak are independently adjustable, whereas the height and width of a true DSC two-state transition are not independent (see section B). For each of the four salt conditions, the DSC baseline slope is positive and linear at temperatures below the RNA transitions, and nearly zero at temperatures above the transitions. The heat capacity of unmelted 5S RNA evidently exhibits a stronger temperature dependence than does melted 5S RNA. A sim ilar effect has been observed for proteins (156) and for tRNA's (155). Four salt conditions a. No added Na+ or Mg++. In the absence of added cations, 5S RNA melts in four stages at re latively low temperature (26°C < Tm < 42°C). The total enthalpy of melting (^Hexp) is small (ca. 285 kcal/mole) as is the thermodynamic s ta b ility (ag298 = 10.4 kcal/mole). In the absence of counterions for the RNA phosphates, intrachain Coulomb repulsions are evidently able to overcome base stacking and base-pair hydrogen-bonding with the help of only a l i t t l e additional heat energy. Tertiary structure may be absent altogether. b. 100 mM NaCl with no Mg++. Following neutralization of the backbone phosphates by 100 mM Na+, the melting process broadens to a much wider (and higher) temperature range (35°C < Tm < 62°C), and resolves into four stages. AHexp increases markedly to 371 kcal/mole, as does the thermodynamic 137 stability (4G298 = 31.5 kcal/mole). The maximum of the composite DSC peak occurs at about 52°C, in good agreement with Tm = 54°C from FT-IR (see Chapter I I I ) . The agreement is important because i t shows that the temperature-induced changes in the optical spectrum (which are due primarily to unstacking of bases) account for most of the energy of unfolding seen by DSC in the absence of Mg++. Therefore, i f most of theZIG 298 observed by DSC is due to base-stacking (as opposed to interactions between base-paired segments, for example), then DSC should provide a good basis for assessing various secondary structural models whose calculated stabilities are based in turn upon optical melting studies of mixtures of complementary homopolymers. c. In 10 mM MgCl2 with no added Na+ . As previously observed for tRNA (155), the addition of Mg++ stabilizes the 5S RNA: melting occurs much more cooperatively, and at significantly higher temperature (ca. 70°C). This result is significant different from the observation that _E. coli 5S RNA does not melt cooperatively under similar conditions (159). In addition to neutralizing the phosphates, Mg++ evidently acts to link single-stranded segments together, forcing the molecule to hold together until essentially all of the segments have melted. ^Hexp reaches its largest value (480 kcal/mole), as does the thermodynamic stability (ZlG298 = 61 kcal/mole). 138 d. 100 mM NaCl with 10 mM MgC^. This data set shows that addition of Na+ actually reduces the total enthalpy of melting produced by Mg++ alone: i4Hexp = 398 kcal/mole and/iG298 = 51 kcal/mole. Again, a similar effect is seen for tRNA's (157). In particular, it is interesting to note thataH and 4G298 f or component transitions 1, 2 , and 4 are relatively unaffected by Na+: the principal effect of Na+ is to destabilize component 3. The DSC composite peak maximum fa lls at about 72°C, somewhat higher than Tm = 69°C from UV hyperchromism and Tm = 57-65°C from FT-IR (see Chapter III). One might at first suppose that the additional enthalpy and higher Tm produced by addition of MgCl 2 to the 5S RNA are due to increased s ta b ility of the macromolecular conformation due to strong attachment of Mg++ to the backbone phosphates. This might be true. However, because Mg++ at high temperature produces irreversible cleavage of the polyribonucleotide chain (see Results) due to breakage of phosphodiester bonds, the results for 0.1M NaCl without MgCl2 probably give the best measure of secondary and te rtia ry conformational enthalpy for wheat germ 5S RNA. Number of base-pairs. In this work, the enthalpies of melting for A*u (9.6 kacl per mole of base pairs), G-c (13.6 kcal/mole) and G-U (7.2 kcal/mole) base pairs are taken from (177). TheirziH for A*u melting is in turn based on optical (178) and calorimetric (179) experiments on 139 poly-rA:poly-rU. for G*C and G'U melting cannot be determined directly: poly-rG:poly-rC does not completely dissociate even at >100°C (188), and poly-rG with poly-rU do not spontaneously form a duplex polymer in the f ir s t place. In these cases, AH is estimated from indirect methods (189) based on data from complementary polyribonucleotide sequences containing known relative numbers of A*U, G*C and G*U pairs. The numbers of A*U, G‘C, and G:U base pairs for wheat germ 5S RNA under salt condition B (100 mM NaCl, no Mg++), have been determined by FT-IR (see Chapter I I I ) , from which an enthalpy of melting is readily computed: 14 A*U + 17 G'C + 5 G*U = 36 total base pairs (14 x 9.6 ) + (17 x 13.6 ) + (5 x 7.2) = 401.6 kcal/mole The DSC value fo r^ H eXp (see Table 4.2) under these conditions is 371 kcal/mole. The consistency between the DSC and FT-IR values greatly strengthens the c re d ib ility of the FT-IR technique for determination of base-pairing in RNA. Since the FT-IR method assumes that all stacked bases are paired (128), both secondary and te rtia ry base pairs are included in the FT-IR base-pair numbers. However, tertiary base pairs exhibit only about half the enthalpy of melting of secondary base pairs (157). Thus, the difference in melting enthalpy between the DSC and FT-IR provides a crude estimate of ((401.6 -371 / ( (9.6+13.6)/2))) x 2 = 6 te rtia ry base pairs in wheat germ 5S RNA. 140 The proton NMR spectrum of wheat germ 5S RNA is highly resolved throughout the melting range (see Chapter V), suggesting that the molecule possesses high conformational homogeneity. Thus, i t is likely that each DSC scan represents a series of transitions for similar molecules, rather than a sum of transitions from a mixture of in it ia lly differen t conformers. Secondary base-pairing models. Figure 12, 13, 14 shows three representative secondary base-pairing models for wheat germ 5S RNA. It is tempting, but not necessarily correct, to identify a particular DSC peak with the melting of a particular base-paired segment of the secondary structure. For example, a single DSC peak might well involve the cooperative unfolding of portion of two or more base-paired segments in the secondary/tertiary structure. The present analysis will thus be limited to comparing total experimental free energy of melting with that predicted from each of the three models. The total experimental free energy of melting of wheat genm 5S RNA is listed in Table 5 for all four salt conditions. Rules have been proposed (151, 153) (see Table 8) for estimating the free energy and enthalpy of an RNA duplex of known sequence compared to its corresponding single strands at 25°C in neutral buffer at an ionic strength corresponding most closely to salt condition B. Figure 45 shows the free energy of the three proposed secondary structures of WG 5S RNA. The experimental total enthalpy of melting from DSC (Table 9) matches most closely to the cloverleaf model (69, 70). The DSC free energy of melting most closely matches the Fox & Woese model, because a Studnicka/Nishikawa model would lead to more secondary base-pairs than are detected via DSC. In Chapter V, homonuclear proton nuclear Overhauser experiments make i t possible to identify particular base pairs, whose melting profiles may then be correlated to the DSC results. 142 Table 8. The base-pairing energies. 3' nucleotide 5' Nucleotide A C G U A 1.2 2.1 2.1 1.8 C 2.1 4.8 3.0 2.1 G 2.1 4.3 4.8 2.1 U 1.8 2.1 2.1 1.2 The shown numbers are negative. The energy for stacking of GC and AU base pair are given. For example, the paired sequence -C-C-and -G-G-are both shown as 4.8 kcal, which is appropriate since both indicate the energy of two successive GC pairs with G's on the same strand. In addition, the energies stacking for the internal GU pairs were computed as follows: GU next to GC = 1.3 Kcal; GU next to AU or GU = 0.3 Kcal. Terminal GU pairs are assumed to be equal to AU pairs. Table 9. Total enthalpy and free energy of melting for wheat germ 5S RNA. AH (kcal/mole) ^G (kcal/mole) Cloverleaf (Luoma & Marshall) 360.0 -23.2 Fox & Woese 326.4 -29.5 Studnicka/Ni shikawa 396.8 -32.3 Experimental (Table 5, Condition B) 371.0 -31.5 143 Luoma & Marshall CCAUCCCAUCAUACCACCACUAAACCACCCCAUCCCAUCACAACI/CCCfcACUUXACCCUC CUUCGGCGAGAGUAGUACUACGAUGCGUGACCUCCUGGGAAGUCCUCCUGUUGCAUUCCC • PAIRED BASES: (33) cc l : { \ l ZZ l i l .1,7 ZX a3:.1.1? MSS |S IVW II SSIS2* gg 52151 ss is; 1 §5 : II 37:74 UA *0:73 GU *1:72 CC ■ 2 • » 4 GC 7 8 - 7 * CC 7 7 : 7 7 UG 8 0 : 7 * AU 8 1 : 7 5 CC 8 2 . 7 4 GC 83:73 AU 83:72 GC 8*:71 GC •HAIRPINS: <3> „ v.„ 3 3 - - > 4 4 * 4 ------> * 7 8 7 — — > 7 0 •BULGES: (3> 4 f >30 3 8 ------> 3 8 8 ------4 > 84 •INTERNAL LOOPS: Fox & Woese GCAUGCGAUCAUACCAGC ACU A AACC ACCCGAUCCC AUC ACAACUCCG AACUUAACCCUG • P A I R E D BASES < 2 7 > 3 : 1 1 4 GC 3 : 1 1 * AU 4 . 1 1 3 UA 1 : 1 1 8 CC 2 1 1 7 GU 1* : *2 AU GU 8 : 1 1 1 AU 7 : 1 1 0 UG 1:111 CG 7 1 1 2 CG 2 7 : 4 8 CG I 8 : * 0 CC 1 7 : 3 7 AU 2 0 : 3 8 1 7 : 4 1 CC * 8:107 GU * 7 : 1 0 8 CG GC 3 1 : 4 * CC 3 2 : 4 3 AU 3 0 : 4 7 GC 7 8 : 7 8 CC 77 : 77 UG «S : 107 GC * 7 10* AU 7 0 : 1 0 3 8 2 : 7 4 CC 8 3:73 AU 8 0 : 7 * AU 8 1 73 GC •HAIRPINS (23 3 3------> 44 8 4------> 72 •B U L G E S ( 0 > NONE •INTERNAL LOOPS: <*> -- 4 « . _ > 3 7 7 1 ------>7 7 TT - > I 0 4 •CRUDE STABILITY- -33.2 KCAL/MOLE • MULTI-BRANCHED LOOPS: < * * LOOP SIZE: 7 LOOP TYPE. MIXED •ADJUSTED STABILITY- -27 3 KCAL/MOLE Studnicka & Nishikawa • P A I R E D BASES <34* UA 3 : 1 1 4 GC 2 1 1 7 GU 3 : 1 * AU 4 113 1:118 GC UG 1 4 *3 CG CC 7 GU 8 : 1 1 AU 7 1 1 0 * : 1 1 3 1 1 2 CG 1 7 5 7 AU CC 1 * * 2 AU 1 7 * 1 GC 1 a * 0 13:84 4 7 GC 3 1 4 * CC 2 0 : 3 8 CG 2 7 4 8 CG 3 2 4 3 AU 3 o * 7 AU 70 1 0 3 GC 2 7 : 3 2 AU * 7 1 0 8 CG * 8 10 7 GC 10* 7 3 CC 82 7 4 CC 7 8 : 7 8 CC 77 7 7 UC 8 0 7 < AU 8 1 3 1 CG 8 3 : 7 3 AU 83 7 2 GC 8 * 7 1 GC 2 8 • H A I R P I N S C 2 > 3 3 — — > 4 4 17 — > 7 0 •BULGES « 3 > * 3 — — >8 3 4 7 - - > 3 0 8 4------> 84 * IN TERNAL LOOPS: < 2 > 2 1 - — > 2 * 3 3----->37 7 1 -----> 77 77-->l04 •CRUDE STABILITY- -33.* KCAL/MOLE •MULTI-BRANCHED LOOPS: (1> LOOP SIZE: * LOOP TYPE: MIXED ■ADJUSTED STABILITY- -32.3 KCAL/MOLE Figure 45. Free energies of the three proposed secondary structural models of WG 5S RNA. CHAPTER V IDENTIFICATIONS AND ASSIGNMENTS OF BASE PAIRS IN 5S RNA SECONDARY STRUCTURE FROM WHEAT GERM BY *H 500 MHZ FT-NMR A. Introduction Since Purcell and Bloch independently demonstrated the nuclear magnetic resonances (NMR) effect in water and hydrocarbons in 1946, NMR techniques have become indispensable for the modern chemical laboratory. In the early 1970's, NMR methods were applied to biological macromolecules and invaluable information was added to the field of biochemistry. In 1971, tRNA in a water solution exhibited a large number of hydrogen-bonded imino resonances in the 10-15ppm downfield region of *H NMR spectrum (200). About a year later, sim ilar resonances were observed for £ . coli 5S RNA (137). However, highly effective water nulling excitation methods (131, 136) and homonuclear Overhauser Enhancement (NOE) (201) have not made i t possible to identify the types and sequences of base pairs, and to assign these resonances to their correct primary sequences positions in the secondary and te rtia ry structure of tRNA. These two developed methods ( i. e . , water suppression and NOE enhancement) were widely used in the elucidation of tRNA solution strucure (142, 202-205). Similar 144 145 techniques have recently been applied to the detection of secondary base-pairing in ribosomal 5S RNA from E^. coli (117-119). Since 5S RNA is approximately 1.5 times larger than tRNA, the proton spectrum for the whole molecule is much more crowded in the region of 10-15 ppm than in the corresponding region for tRNA. Attempts have been made to examine RNase A-resistant fragment (in order to avoid the ambiguities resulting from extensive peak overlapping from imino-hydrogen bond protons (119)), from which the molecular stem and prokaryotic loop of proposed universal base-pairing schemes for 5S RNA (68) were p a rtia lly assigned. However, the regions which d iffe r between different proposed base-pairing schemes (e.g. helix A and A' in Figure 12 and helix II and helix III in Figure 13) have not yet been detected by NMR . Furthermore, since helix I I I in commonly proposed models has been suggested and argued to be involved in the interactions with tRNA as well as having molecular transitions between the A form and the B form (58). Therefore, its existance and detection will provide valuable evidence on the true structure of 5S RNA. In this chapter, a variety of NMR techniques is discribed that have been used to study the secondary structure of wheat germ 5S RNA. B. Water suppression techniques. Although the incorporation of Fourier Transform methods into NMR spectroscopy (206, 207) gave a landmark improvement of two orders of magnitude in the spectral signal-to-noise ratio S/N, the special problem of observing weak signals in the presence of an interfering 146 strong signal (usually from solvent) remained one of the most difficult experiments. In many biological studies, water cannot be replaced with D 2O, as in the detection of exchangable protons (-NH) in proteins and nucleic acids. I f not su fficiently suppressed, the strong H2O signal can exceed the dynamic range of either the am plifier or digitizer of the spectrometer, creating intense spectral artifacts and rendering weak resonances undetectable. 1. Dynamic range problem. Dynamic range is important not only for seeing small signals in the presence of large ones, but also for detemining the S/N in a given experiment. It affects spectral resolution as well as the overall quality of the transformed spectrum. For optimal operation of the analog-to-digital coverter (ADC), the input signals must be adjusted to fill as much of the ADC word length as possible: thus the presence of large signals ratioing down the weak signals from the same sample thus preventing the dynamic range of the ADC from being used e ffic ie n tly . This aspect of the dynamic range problem can be solved only by supressing the large signal prior to d ig itiza tio n . The other aspect of the dynamic range problem is related to the limited word length of the computer memory. To understand this problem, two principles of time-averaging may be recalled, (a), coherent signals grow lin e arly with the number of scans (N) while incoherent noise grows only with JIT; (b), when the noise level exceeds 147 the digitizer threshold (DT), then signals which by themselves are too small to activate the d ig itiz e r can be also digitized. When free induction decays (FID) are accumulated in the time domain, no overflow of data can be tolerated at any stage in the collection process. For example, the Bruker AM-500 spectrometer is equipped with a 16-b it ADC and 24 bit computer word length for the signals, theoretically one can accumulate 256 fu ll-scale transients since data accumulation can continue only until the computer word is filled. Obviously, 256 scans are not enough for the very d ilu te and weak signals, and many more scans must be accumulated. In order to prevent overflow of data and truncation of any part of the FID, data has to be scaled down when the accumulated signals are approaching the contents of the appropriate least significant b it of the ADC, thus effectively reducing its word length. However, reducing the ADC resolution in this way eventually causes the S/N to decrease and continued averaging under these conditions w ill then not improve the overall S/N (136). 2. Water suppression techniques. Increasing the computer word length has many benefits. For example, compared to a 16-b it computer, a 24-bit computer offers at least three advantages. F irs t, even before scaling the data becomes necessary, the number of scans and smallest possible signal detectable are both improved by a factor of 256. Second, the dynamic range is also improved by a factor of 256. Third, the maximum S/N gain possible is improved by a factor of 16. 148 There is a large volume of lite ra tu re in which the methods for overcoming the dynamic range problem have been discribed. These methods are summerized below: (a ), rapid-scan correlation spectroscopy (208), (b). deuterated samples, (c ). data manipulations (data s h ift accumulation (DSA), and alternate delay accumulation (ADA) techniques)(136), ( d ). selective saturation of the solvent protons (209), (e). water eliminated Fourier transform (WEFT) method which takes advantage of the differences in relaxation times between the solvent and solute protons (135), ( f ) . Redfield 21412 selective pulse sequences used to eliminate the excitation power near the water NMR frequency. This method is an improvement over the long (or soft) pulse method (Figure 46B) (131, 132), (g ). time shared hard pulse method (e .g ., 1331 pulse sequence) (Figure 46C)(133). Among these methods, only (c ), ( f ) , and (g) are suitable in nucleic acid research for studying secondary structure. Since guanine N^H and uracil N3H exchange rapidly with water protons, saturation of water protons may give rise to the disappearance of the imino proton resonances via saturation transfer. In this work, only the selective pulse method and data manipulation technique are used and discussed. Figure 46 shows the frequency domain representation of: a time-domain long pulse, a Redfield 214 pulse, and a time shared hard 149 pulse (1331), to demonstrate the principles of these methods; Figure 47 shows an experimental comparison of these methods. B riefly, the long pulse method is based on the different flip angles for H 2O and the protons of interest during the relatively long excitation pulse (Figure 46A). During time interval, t j , the water proton magnetization precesses by 360° around the effective magnetic fie ld , Hef f , whereas the magnetization of a peak of interest precesses around Hi (exciting r.f. magnetic field) by only 90°. To satisfy this relationship, the following equations were derived. J( a v i) 2 + (Hi/r)2 * ti = 2-rc (5.1) (H i/r) t i (5.2) where aw is frequency difference between the water resonance position and the resonance position of the proton of interest , If is the magnetogyric ra tio , and t j is the excitation pulse length. The solution of this equation is as follows: few) t i =JTE K/Z = 1.936 1Z (5.3) This experiment thus provides a good way to observe (fo r example) an A*U base pair proton, for which the peak is located 10 ppm away from water peak (4.78 ppm); at 500 MHz, the 90° pulse width, t]_, is 193.6 us. Since the excitation is longer than the commonly used 90° "hard" pulse (6.8 us), Hi must be made smaller in amplitude. Thus it is refered to as a "soft" pulse. 150 Experimentally, however, the long pulse produces only a very narrow null area at the H 2O peak, as seen in Figure 46B (solid curve). In order to produce a broad null (see Figure 46B dotted curve), the Redfield 21412 pulse sequence was employed (see Figure 47C). This pulse can be viewed as a simple long pulse of amplitude Hj minus two short pulse of amplitude 2Hi. A very effective result for water suppression was obtained as illustrated in Figure 47C, which provides a much better spectrum than in the case of long pulse method. Even with this powerful method, however, the water peak is only suppressed 500-1000 times depending upon operator s k ill. Under these circumstances, the Alternate Delay Accumulation (ADA) method is then imposed on the Redfield pulse sequence. The idea behind the ADA method is that if an FID is phase shifted by 1/2 of one cycle of the detected H 2O frequency by introducing a suitable delay prior to data collection, and the delay is introduced on the alternate scans, then the shifted and unshifted FIDs can be accumulated alternately. The water peak can then be cancelled by selecting the delay time. This NMR program is listed in Appendix under the name of REDADA.AU. Results from combining the Redfield 214 pulse and the ADA technique is show in Figure 47D, in which the suppression factor for water has improved about 100 times. This method is often used to acquire normal spectrum. Since the Redfield pulse sequence requires fine adjustment of experimental parameters (pulse lengths, phase sh ifts, delays and transmitter frequency), a new method for water suppression for proton NMR spectra of aqueous solutions was designed with a short, strong 151 pulses as illu strated in Figure 47E. This pulse, called the 1331, can be Fourier transformed as shown in Figure 46C, from which i t is clear that this pulse sequence produces a broad f la t region of near-zero excitation power near the solvent frequency and ye t appreciably excites the relatively distant resonance frequencies. The major advantage of the 1331 method is its ease of implementation, since the centering of the r . f . carrier freqency is no longer c r itic a l. This method with the ADA technique was used to obtain the wheat germ 5S RNA spectrum shown in Figure 47E. The program corresponding to this spectrum is listed in Appendix as P1331ADA.AU Both the REDADA.AU and P1331ADA.AU methods gave nearly the same water suppression e ffe ct. One important aspect that should be mentioned is that, in theory the pulse can be a hard pulse, but i f the normal 90° pulse length is 6.8 us as in the Bruker AM 500 spectrometer, then the "1" pulse would be only 0.85 us, and the pulse rise and fa ll times are significant; therefore to achieve better results, the transmitter power is reduced by a factor of 100 with an in-line 20 dB attenuator, thus giving a "1" pulse length of nearly 6.8 us. In conclusion, both the REDADA.AU and P1331ADA.AU are effective and successful for water suppression and both were used. C. Nuclear Overhauser Effect (NOE) for determination^ RNA structure. 1. The general description of NOE. In general, NOE is the process in which strong, selective irradiation of nuclear resonance affects the intensity of the 152 A B f Figure 46. (A). The water peak rotates magnetization precession for long pulse centered off-resonance from H 2O by ZK around Heff and the imino proton peak rotates by nr/2 around Hi (132). (B). Freqency-domaln spectrum of a long pulse of length, ( ------) , spectrum of Redfield 214 pulse ( ------) (131). (C). Power spectrum resulting from the 1331 pulse sequence as a function of frequency offset from the transmitter (133). 153 Figure 47. Effect of various excitation pulse sequences on *H NMR spectra of wheat germ 5S RNA. A. Single hard (i.e ., short, high-amplitude) pulse. The presence of the huge (ca. 100 M) water peak effectively makes detection of the dilute (ca. 1 mM) RNA base-pair protons impossible. B. Single soft ( i . e . , long, 1ow-amp!itude) pulse, centered at 15 ppm, and designed to produce very low amplitude excitation at H 2O. Base-pair protons are v is ib le , but weak. C. Redfield 2-1-4-1-2 soft pulse sequence. RNA peaks are seen, but the large residual water peak affects the baseline and makes phasing d iffic u lt. D. Redfield soft pulse sequence with alternate delayed aquisition. The residual water peak is reduced by about another factor of about 64. E. 1-3-3-1 hard pulse sequence with alternate delayed acquisition. Performance is sim ilar to case D, but requires less optimization of pulse widths and center frequency. Wheat germ 5S RNA A* Hard Pulse 0 # Long Pulse __|~ C« Redfield j~ n Q . Red ADA E. ADA & Hard 1331 ~ i------1------1------1------1------1 14 12 10 8 6 4 < J l PPM •t* Figure 47. 155 resonance signals from other nuclei (143). The change in intensity is the result of cross relaxation between the observed and the irradiated nuclei. In this way, the relative spatial proximity of various sets of nuclei in the molecule can be detected. The efficiency of this cross saturation fa lls o ff as the sixth power of distance, and the size of the NOE for a particular proton depends on how many other protons are also close enough to be dipole-coupled to the irradiated proton (144). For a molecule the size of 5S RNA, proton-proton NOEs are detectable only out to distances of around 4A, beyond which the NOE fa lls to about 1% and becomes v irtu a lly undetectable. Normally less than the maximum 50% enhancement is observed in the case of the homonuclear NOE as shown in Figure 48, because other relaxation processes exist. Figure 48 shows the relationship between fft(B) and logu?rc, where fa(B) is the change of the intensity of spin-A under uO conditions of total saturation of the B-spin,Ais the Larmor frequency, H c is the correlation time of random motion of the molecule. Figure 48 is obtained under the assumption that a ll of the relaxation of the spin-A is by magnetic dipole interaction with irradiated spin-B, so no distance information is implied. In the NMR time scale, as the molecular weight of solute increases , or as the viscosity of solvent increases, correltion time ~Cc becomes longer. When a higher static magnetic fie ld strength is used in an NMR spectrometer, the Larmor frequency u )o f the observed nucleus is larger. The product oftDTJ; determines whether the NOE is positive or negative. As shown in Figure 48 the change from positive to negative effect occurs atu)*£ = 1.118. As long asu )T c < 0.1, the NOE is close 156 -0.5 - 1.0 2 log tSTc Figure 48. Plot of the homonuclear NOE in a two-spin system, f/\(B) versus logd is Larmor frequency,Xc is the correlation time (the time taken by a molecule to rotate through one radian) of the molecule (ca. 3 x 10"® sec for 5S RNA). The product o f^ ^c is 15 (a t Bruker AM-500). See text for detail s i 157 to the maximum value, + 50%, whereas = 10 or more, NOE is then near -100% which corresponds to disappearance of the signal from spin A. The present application are based on negative NOE's between base pair protons because the product of is bigger than 1.118 (ca. 15 ). 2. Application of NOE to the nucleic acid research. The homonuclear NOE technique for nucleic acid research has been extensively reviewed (210-211). Here, only the application of NOE's for base pair sequencing is presented. There are two aspects of base pair sequencing. F irs t, the base pair types have to be distinguished (e .g ., A-U, G-C, or G-U). Second, the relation between those identified base pairs has to be established. For both of these two aspects one relies on the NOE technique. Since presaturation of an A-U base pair imino proton gives a sharp, strong NOE in the aromatic region, which is in contrast to the broad NOEs in the same region produced by a G-C base pair, whereas the two protons of a G-U wobble base pair gives strong mutual NOE's (20-40%). Thus, these three normal base pair types can be easily distinguished from each other (see Figure 51). On the other hand, because of the saturation between base pair planes, a given imino o proton is usually at least 3.8 A away from any other imino proton and only weak NOE values (2-4%) between the imino protons of adjacent Watson-Crick pairs can be observed in the low -field region. From these weak NOEs and their identified base types, the NOE connectivity can then be detected, from which a base-pair sequence may be infered. 158 Such techniques have been used extensively in tRNA solution structure determination. The same method can now be applied to the determination of secondary structure of wheat germ 5S RNA. D. Materials and Methods. 1. NMR sample preparation. Wheat germ 5S RNA, fragment a, fragment Bj, and B 2 were isolated and purified as discribed in Chapter II.E . The NMR samples were prepared as follows: a. Intact wheat germ 5S RNA NMR. Samples of intact WG 5S RNA were prepared in three d ifferent buffers. First, all samples were dissolved in a solution containing 10 mM EDTA, 10 mM sodium cacodylate at pH 7.0, and dialyzed twice against the same buffer at a volume ratio of 1:500 for 4 hr. The samples were then dialyzed three times against water at a volume ratio of 1:1000 for 3 hr, and lyophilized. The lyophilized sample was dissolved in the three different buffers which all contain 10 mM sodium cacodylate, 95%/5fc H2O/D2O, pH 7.0 to which 100 mM NaCl, lOmM MgCl2 (buffer A), 100 mM NaCl only (buffer B), 10 MgCl 2 only (buffer C), or no NaCl or MgCl2 (buffer D) were added. The concentration of each sample was determined as described in Chapter II.D . RNA concentrations were: (A) 30 mg/ml, (B) 35 mg/ml, (C) 26 mg/ml, and (D) 24 mg/ml. 159 b. NMR samples of RNase T1 cleavage fragments. NMR samples of RNase T1 cleavage fragments A, Bl, and B2 (B1 is called the "core" of 5S RNA, and B2 is called "common arm") were prepared as described above, except that the fragments lyophilized powders were dissolved in buffer B. In order to renature these fragments, dissolved samples were heated in a water bath at 58°C for 5 min and cooled slowly (30 min). The amount of RNase T l-resistan t fragment A was 148 A 26O in 0.4 ml • An NMR sample of the common arm fragment (Bl) was prepared as above, as 105 A 26O units in 0.4 ml; the core fragment of the molecule (Bl) was 450 A 26O units in 0.4 ml. 2. NMR spectroscopy. All spectra were obtained at 11.75 tesla (500 MHz for *H) with a Bruker AM-500 spectrometer unless otherwise specified. The NMR signals were produced by a modified Redfield 21412 pulse sequence (131) with Alternate Delay Aquisition (ADA) (136) in order to suppress the H2O signal (factor of ca. 35,000) and give flatter baseline. 4K data sets (16 bit/word per transient; 24-bit/word in the accumulated transient) were acquired in about 0.17 s. Total acquisition cycle time was 0.8 sec, based upon independent determination (not shown) the longest T^ relaxation time as 70 ms. The NMR programs for acquiring normal spectra are listed in Appendix as REDADA.AU and/or P1331ADA.AU; minor baseline corrections were also performed. Sample temperature was calibrated from independent measurements of the chemical shift of methanol and ethylene glycol (212, 213). Downfield 160 shifts are defined as positive, and chemical shifts are defined re lative to an H 2O chemical s h ift of 4.78 ppm. Proton homonuclear Overhauser enhancement (NOE) experiments were performed by 0.3 s pre-irradiation of the resonance of interest with phase cycling procedures described elsewhere. Each displayed NOE (difference) spectrum was based on time-domain subtraction of 8000 on- from 8000 off-resonance scans. The decoupler power was within a few dB of 30 dB below 0.2 watt, according to the of a given peak, the degree of overlap with nearby peaks, and the degree of diffusion of spin polarization from one base pair proton to another. The NMR program used in the NOE experiments is listed in Appendix as MULNOE.AU. E. Results and Discussion. 1. Number of base pairs in intact wheat germ 5S RNA. The downfield 500 MHz iH FT/NMR spectrum of wheat germ 5S RNA is shown in Figure 49 Because the hydrogen-bond base pair imino protons give FT/NMR resonances located 10-15 ppm downfield from DSS, one might expect to be able to determine the total number of base pairs in an RNA by simple integration of the 1H NMR intensity in that spectral region. Such estimates are made difficult by several factors. First, even with the Redfield 21412 excitation and alternate delay acquisition, it is difficult to obtain a flat baseline because of the residual signal from H2O. When, as in the present data, the baseline curvature is not very severe, it is permissible to flatten the 161 K U V W PQi ST , 1 1 1 1 1 14 12 10 PPM Figure 49. 500 MHz low fie ld NMR spectrum of wheat germ 5S RNA (35 mg/ml) at 23°C in buffer B (10 rrM sodium cacodylate, 0.1 M NaCl, 1.0 nW EDTA, pH 7.0, in 95:5 HeOiDzO, after complete removal of Mg2+ ion). The spectrum has been resolution-enhanced by Lorentz-to-Gauss conversion. The peaks are labeled from A to W for convenience in di scussion. 162 residual curvature by fitting the baseline to a polynomial and subtracting out that curve. Second, the large number of base pairs in 5S RNA leads to severe overlap in this spectral region (see peak K in Figure 49). The best integration is therefore achieved by simulating the spectrum by superposing the minimum number of Lorentzian (or Gaussian, depending on the experimental line shape) peaks of variable position but equal width and height required to match the experimental spectrum. When applied to the spectrum in Figure 49, such a simulation gives a total base pair number of about 35. Although such a procedure corrects for peak overlap, i t can s t ill give incorrect base pair estimates because the relative peak areas may not be proportional to the numbers of base pair imino protons, due to variation in exchange rate between d ifferen t base pair protons and H2O. Yet another problem is that some resonances in this region may arise from non-Watson-Crick base pairs: e.g. ring NH protons (as for G*U pairs) or exocyclic amino protons involving in hydrogen binding of other non-Watson-Crick base pairs (as for guanosine, cytosine, and adenine). In support of this point, Hasnoot et a l . (214) have shown that some protons in the molecular in te rio r (e .g ., non-paired G's and U's) may be buried so deep that they cannot exchange with water molecules. Variation in observed peak intensities arising from detection cycle period £ Tj can be eliminated (as in this case) by selecting a repetition period that is long compared to Tj. A final problem is that i f 5S RNA in solution can exhibit two or more conformations with interconversion lifetim es longer than a few seconds, then the same base pair proton from two different conformers 163 may display more than one chemical s h ift. We are thus led to seek methods which can establish the number of distinct base pair protons for an envelope of overlapping resonances. 2. Resolution of overlapping peaks via change in salt concentration. Magnesium ions bound to tRNA (157) or wheat germ 5S RNA (see Chapter IV) are known to stablize the RNA secondary and te rtia ry structure significantly. Figure 50 shows that wheat germ 5S RNA exhibits different configurations under different salt conditions, as evidenced by major shifts in the positions of the *H NMR peaks corresponding to hydrogen-bond base pair imino protons. In particular, peaks D and E, unresolved in the presence of Mg2+ (Fig. 50, middle) or Mgz+ and Na+ (Fig. 50, top), split into two easily resolved peaks on removal of Mg2+ and Na+ (Fig. 50, bottom). Sim ilarly the resonances at position K, which are unresolved in the absence of Na+ and Mg2+ (Fig. 50, bottom), s p lit into two major peaks on addition of either Mg2+ (Fig. 50, middle) or Mg2+ and Na+ (Fig. 50, top). Additional Mg2+-induced changes are seen for the peaks located between 10-12 ppm. It is clear that variation in Na+ and/or Mgz+ concentration produces differential chemical shifts for the hydrogen-bond base pair imino protons, thereby furnishing a means for separating overlapping resonances. Once these resonances are identified and assigned, such spectra w ill report the site(s) of specific conformational changes induced by changes in salt concentrations. 164 100 mM N a * mM Mg Wheat germ 5S RNA 10 raM Mg++ DE G No Mg 15 14 13 12 11 10 PPM Figure 50. Use of salt-induced shifts to resolve overlapping resonances in the downfield NMR spectrum of intact wheat germ 5S RNA. Top: 30 mg/ml RNA in buffer A. Middle: 26 mg/ml RNA in buffer C. Bottom: 24 mg/ml RNA in buffer D. Removal of Na+ and Mg2+ (bottom spectrum) resolves peaks D and E. Addition of either Mg2+ (middle spectrum) or Na+ and Mg2+ (top spectrum) resolves peak K into two peaks. Numerous salt-induced changes in the 10-12 ppm region suggest that the solution conformation of wheat germ 5S RNA may change significantly according to Na+ and Mg2+ concentrations. 165 3. Peak identification (A*U, G*C, G*U) from NOE pattern. One of the most useful NMR methods for identifying the number and type (e .g ., A*U, G*C, G.U) is based upon the homonuclear Overhauser enhancement (NOE) technique. As discribed in Section C of this Chapter. In particular, presaturation of an A*U base pair proton (ca. 13.5-14.3 ppm) produces a characteristically sharp and strong NOE in the aromatic region (ca. 7.5 ppm) arising from coupling to the C2-proton of adenine in that base pair (142). Plots B to D of Figure 51 show that the three left-most peaks in the downfield NMR spectrum of wheat germ 5S RNA arise from A*U hydrogen-bond base pair imino protons, as might have been guessed from their chemical s h ifts. The identification of peaks D,E offers a good example of the need for multiple experiments to test NMR-based inferences. F irs t, the resonances at D,E appear more intense than those at A, B, and C (Fig. 49), suggesting the presence of two base pair protons. Accordingly, removal of Na+ and Mg2+ (compare Figs. 50A or 50B with Fig. 50C) shows that peaks D,E indeed represent two d ifferent base pair protons whose chemical shifts happen to overlap when salt is present. F inally, the NOE experiment (Fig. 51, plot E) further reveals that the two peaks at D,E (Fig. 51, plot A) are both A*U's, as seen from the two sharp NOE's generated in the aromatic region on irradiation of peaks D,E. The G*U wobble base pair was the f ir s t to be identified (201, 215) from the strong mutual intra-base NOE's produced between the 166 Figure 51. Identification of base pair types from intact wheat germ 5S RNA via proton homonuclear Overhauser enhancement difference spectra. Presaturation of peaks A-E (spectra B to E) gives sharp NOE's between 7.0-7.5 ppm (see arrows), identifying resonances A-E as A*U base pair hydrogen-bond imino protons. Note that the two distinct NOE's in plot E reveal that there are two A*U's at the positions labeled D and E in the normal *H NMR spectrum (plot A). Presaturation of either peak R or V (spectra F and G) gave a strong mutual NOE to the other, identifying R and V as the two hydrogen-bond imino protons from a G*U base pair. Various smaller NOE's (e.g., K and L in plot B) represent NOE connectivity to nearest-neighbor base pairs (see te x t). 167 K V W KL —,------[------1------1------1 14 12 10 8 6 PPM Figure 51. 168 spatially close (2.5 Angstrom) GN1H and UN3H. A typical G*U NOE pattern is observed for peaks R and V of Figure 49 (see Figure 51, plots F and G), providing immediate identification of R and V as the two hydrogen-bond imino protons of a G*U base pair. 4. Base-paired segments determined by NOE connectivity. The small (ca. 2-4%) NOE difference peaks arising from population transfer from an irradiated base-pair proton to a proton of the base pair immediately above or below i t in an RNA helix can be used to identify adjacent base pairs. For example, Figure 51 (plot D) shows that resonance C (an A*U base pair proton) is NOE-connected to resonance G (a G*C base pair proton— see below). Sim ilarly, resonances R and V (G*U base pair protons from the same base pair) are NOE-connected to resonance G (Fig. 51, plots F and G). A series of such two-fold connections can be used to sequence longer helical base-paired segments of tRNA's and DNA's in solution ( 202-205, 216-217). However, extension of the method to 5S RNA's presents new problems. F irs t, the secondary and the te rtia ry structure of tRNA were already known (218-219) in advance of the NMR work, whereas the 5S secondary structure (if a unique structure exists) is still debatable (43, 68-70, 73, 75, 76, 220) ; te rtia ry 5S RNA models are basically guesswork (58). Second, because 5S RNA is 1.5 times bigger than tRNA, there will be more peaks of greater individual line width, exacerbating the problem of peak overlap, and the same wt/vol concentration will yield 50% lower signal-to-noise ratio for 5S RNA compared to tRNA. Suppose a base pair sequence from NOE connectivities is going to be constructed. Resonance C (Fig. 52, top) is an A*U base pair proton, and is NOE-connected to resonances G and K. The chemical s h ift and broad aromatic NOE produced by irrad iation of peak G (Fig. 52, middle) or peak K (not shown) identify both as G*C's. Figure 52 shows that peak G is NOE-connected to peaks C and R,V (and Q and possibly K). Finally, peaks R and V (Fig. 52, bottom) are clearly from the same G*U, as argued previously. We have thus established NOE connectivity from peak K to peak C to peak G, and from peak G to peaks R, V corresponding to base pair sequences, G*C-A*U-G'C and G*C-G*U. [The polarity of each base p a ir--e .g ., G*C vs. C*G— cannot be determined from these measurements.] One might be tempted to propose the completely connected sequence, K-C-G-R,V, corresponding to G-C-A-U-G-C-G-U. However, the strong intensity of peak G suggests that i t may contain more than one base pair proton. Thus, we cannot be sure that the G*C (peak G) adjacent to an A*U (peak C) is the same G*C that is adjacent to a G-U (peaks R,V). This point is readily resolved from the temperature-dependence of the *H NMR spectrum, as discussed next. 5. Temperature-induced shifts and melting as an aid in base pair sequenci ng Temperature-induced changes in the ! h NMR spectrum of 0.7 5 mM wheat germ 5S RNA in buffer A are shown in Figure 53, obtained with a General Electric NT-500 FT/NMR spectrometer. Since two NOE-connected 170 A-U G * C G * C VS. G-C G*U G* U U G IVheat germ 5S RNA G-C G * U , 1 1 1 1 14 12 10 8 6 PPM Figure 52. Ambiguity of NOE connectivity when overlapping resonances occur. Peaks C, G, and R are identified as A-U, G*C, and G*U from their characteristic primary NOE patterns. The top two spectra show that A*U is adjacent to G*C, and the bottom two spectra show that G*C is adjacent to G*U. However, since peak G represents two resonances (see 35°C spectrum in Figure 53), we cannot be sure that the G*C connected to the A*U is the same G*C that is connected to G-U. 171 base pair proton peaks presumably occupy adjacent rungs of an RNA A-helix, both base pairs should "melt" ( i . e . , the inverse lifetim e for chemical exchange between the base pair proton and H 2O becomes faster than about 30 sec- *) at about the same temperature. For example, peaks A and L (whose NOE connectivity is shown in Fig. 51B) melt together at about 50°C. In considering the problem of the overlapped G*C resonances a t position G, two resonances, Gi and Gg, are clearly resolved when the temperature is increased to 35°C. Peak G 2 melts together with peak C (an A*U pair) at about 50°C. However, peaks R and V (a G*U pair) remain stable up to 55°C, along with peak Gj_. Thus, i t is now clear that the G*C (peak G 2) which is adjacent to an A*U (peak C) is different from the G*C (peak Gj) which is adjacent to a G*U (peaks R,V). The corresponding base paired segments, G*C-A*U-G*C (peaks K-C-G2) and G’C-G*U (peaks Gi-R,V), lik e ly occur in different base paired helical stems. G*U base pairs have generally been regarded as much weaker than A*U or G*C pairs (151). I t is thus somewhat remarkable to find a G*U base pair (peaks R,V) that is as stable at high temperature as the remaining G*C pairs near 12.5 ppm (see 55°C spectrum in Figure 53). The stem containing the G*C-G*U segment must be stabilized by the presence of other secondary G*C pairs or additional te rtia ry base pairing. Finally, peaks B (A*U pair) and 0, whose NOE connectivity is shown in Figure 5.6C, are also both stable at high temperature (see 50°C spectrum of Figure 53). In contrast, the A*U base pair protons 172 Figure 53. Use of temperature-induced melting and shifts to identify and confirm base pair sequences in intact wheat germ 5S RNA (buffer A). Peaks A and L melt together, in agreement with their NOE connectivity (Fig. 51, plot B). Note that peaks G splits into two peaks as the temperature is increased from 25°c to 35°. Moreover, peak G2 melts together with its NOE-connected peak C. On the other hand, peak Gj is still intact at 55°C, as are peaks R and V from a G*U base pair (see Fig. 51, plots F and G). Peaks B and 0 are more stable than peaks A or C. Temperature-variation experiments thus verify NOE connectivities and help to distinguish different NOE-connected segments. Wheat germ 5S RNA 2 5°C T T T I ...... IS 1 4 13 1 2 11 io PPM Figure 53. 174 corresponding to peaks A and C are relatively la b ile , and lik e ly occur in less stable regions (e.g., helix II or helix III of the Fox/Woese model (see Figure 13). 6. Use of RNase T1 cleavage fragments for assignment of base paired segments. Although the resonances corresponding to five A’ U and one G*U base pairs have been identified, and their NOE connectivities established (Figs. 51-53), it remains to assign particular resonances to particular primary sequence positions in wheat germ 5S RNA. For example, the base pair sequence, G*C-A*U-G*C, deduced from NOE connectivity between resonances K-C-G 2 could be assigned to two possible segments in the Fox and Woese model: namely, C18G60-A19>J59-C20G58 and C105G70-U106A69"C107G68* A simple method for resolving such an issue is to cleave the intact 5S RNA molecule in such a way as to eliminate one possible helical segment while preserving another. Figure 54 shows *H NMR spectra and corresponding secondary structures (Fox/Woese model) of intact wheat germ 5S RNA, an RNase Tl-resistant cleavage fragment A, a "core" of the molecule (B i) and a smaller fragment corresponding to the "common arm" that is present in v irtu a lly a ll proposed secondary structures for 5S RNA. Figure 54 shows that peaks B (A*U), D (A-U), E (A*U), and R,V (G-U) are missing in the RNase T l-resistant cleavage fragment a, and thus appear to arise from segments I , IV, and/or V (and/or from 175 Figure 54. Proton 500 MHz NMR spectra (le f t ) and proposed secondary structures (right) of intact wheat germ 5S RNA (top) and Three purified fragments produced by RNase T1 cleavage. All samples are in buffer B. The most obvious results of RNase T1 cleavage are the elimination of three A*U base pairs (peaks B, D, and E) and one G*U (peaks R and Y) in the fragment a, common arm fragment (B 2) is a part of the fragment a, and the common arm fragment (B 2) and the core of the molecule (B^) are complementary with each other. Wheat germ 5S RNA r r PC . £iK C.u . A,u I N ative : * c ill iv;;, < * lu i « *cccc »* /* c»cS»-cf cccuccuc " euccc,6 'V c^ 4^ c«u «“ u * 6 « Cc C l « A C A * _ .^ A A C Uu iC» c C e iu i „ -. Cc 6 * ~-C »• W N PC . c • H: RNase T1 IU u « c c c ^LcCC »•riSUA<^cufsc*^c to r r Core Domain *c.r6 *U A »U id ! C'O t:a v-!.,> :r vuu.r IV *f CCCCCU^C cce*u.?'V5e 4 ^cilcc^X #6euceue J; Common Arm III ,'u/* eCc'-. U»$“C I »c ’5 . c « ccc. *• °** —J— 14 12 10 a>'Nj PPM Figure 54. 177 te rtia ry base pairs). For the same reason, since peaks A and C remain afte r RNase T1 cleavage, we discover that peaks A (A-U) and C (A*U) and their NOE-connected partners are found in helices I I and I I I . Further cleavage of the RNase T l-resistan t fragment was used to produce the "common arm" segment shown in Figure 54 (bottom). This common arm can also be obtained together with the core of the molecule by cutting the intact molecule using relative strong dose of RNase T1 and higher reaction temperature which is discribed in Chapter II.E . Since peak A remains in the spectrum of the common arm fragment, we have detected at least one A*U base pair from the common arm. Sim ilarly, since peak C is missing from the common arm segment, we conclude that peak C lies in helix I I and is assigned to A 19 U59 . Therefore, NOE-connected resonances K-C-G 2 can be assigned to c18g60“a19u59"c20g58> and constitute the f ir s t direct observation of the "tuned" helix I I segment of the Fox/Woese secondary structural model for 5S RNA. For fragment B^, i t is a complementary fragment with common arm. I t remains the main features of the molecule as clearly seeing in Figure 54. The remains of the peak C in this fragment indicate that the tuned helix I I segment is involved in this structure. 7. Ring current-induced chemical shifts confirm assignment of C18 g60-a19 g59-C20 g58- Further corroboration of the assignment of peaks K-C-Gg as g18 g60"a19 g59~C20g58 1S afforded from ring current calculations. 178 Table 10. Comparison of experimental ^-H NMR chemical shifts with those computed from ring-currents for a proposed base pair sequence. Peak position3 Base pair Calculated shiftb Experimental shift K C18-G60 12.49 ppm 12.53 ppm C A19-U59 13.80 13.64 g2 C20-G58 13.24 13.11 3 See Figure 49. b Computed from Arter & Schmidt (221) for RNA A-helix, fo r assumed intrinsic {unshifted) positions of 14.35 ppm for A-U and 13.45 ppm for G-C (228). 179 Table 10 compares the experimental chemical shifts for peaks G 2 , C, and K with those computed from Arter & Schmidt (221) for nearest- and next-nearest base pairs in the "tuned" helix I I of the Fox/Woese model. The calculated values are clearly consistent with the proposed assignment. In conclusion, a combination of proton homonuclear NOE experiments, variation of salt conditions, temperature-induced melting, and NMR of RNase T1 enzymatic cleavage fragments yields a re la tiv e ly d efin itiv e assignment of a segment of base pairs that can be assigned to the previously unobserved "tuned" helix I I region of the Fox/Woese model of the secondary structure of wheat germ 5S RNA. 8. Are the conformations of the fragments the same as in the intact molecule? I t is important to establish that the enzymatic cleavage fragments d iffe r from the intact 5S RNA molecule only in the absence of some structural features and not by the formation of some new, non-native structure. Since the principal difference between the fragments and the intact molecule is that several resonances are absent in the fragments, i t seems lik e ly that the base paired regions of the fragments are similar to those in the intact 5S RNA. A better test is to compare the NOE behavior of the fragment and the intact 5S RNA. Figure 55 shows that irradiation of peak C in the RNase T l-resistant cleavage fragment produces an NOE pattern near-identical to that from the intact 5S RNA (Fig. 51D or Fig. 52, top spectrum). As further evidence, Figure 56 shows NOE patterns for irradiation of 180 K ~ i ------1------1------1------1 14 12 10 8 6 PPM Figure 55. Presaturation of peak C on the RNase Tl-resistant fragment shown in the middle of Figure 54 gave the same NOE pattern (bottom spectrum) as for intact molecule (compare spectrum 0 in Figure 51). The spectrum of the fragment (top diagram) is included for reference). This result supports the assumption that the fragment retains at least some of the features of the intact 5S RNA molecule. 181 A*U NOE A Wheat germ 5S RNA N ative A RNAse Tl-Resistant Fragment A Common Arm -i------1------1------r 14 12 10 8 PPM Figure 56. The same NOE pattern is produced by irradiation of the A*U resonance at peak A, whether from intact 5S RNA (top), the RNase Tl-resistant fragment (middle), or the common arm fragment (bottom), indicating that this segment of the molecule survives intact afte r RNase T1 cleavage. 182 A * U Wheat germ 5S RNA N ative a -u RNAse T l- R e s is tant Fragment T T 14 12 10 8 PPM Figure 57. The same NOE pattern is produced by irrad iation of the A*U resonance at peak B, whether from intact 5S RNA or in the fragment Bj { the core of the molecule), indicating that this fragment remains many features of the intact molecule. 183 peak A from the intact molecule (bottom spectrum), RNase T l-res istan t fragment (middle spectrum) and common arm fragment (top spectrum). The near-identical NOE pattern for all three RNA's offers strong evidence that the conformation of the fragments is sim ilar to the conformation of the corresponding segments of the intact wheat germ 5S RNA. [The slight sh ift in position of peak A in the common arm fragment compared to the other fragment and the intact 5S RNA is due to the lower temperature (2°C vs. 23°C) at which the NOE of the common arm fragment was conducted]. Similar evidences are shown in Figure 57 for confirming the fragment Bj also retains its native features. When peak B was irradiated, Peak K and 0 were responsed as NOE signals no matter in the intact 5S RNA or in this fragment. 9. 5S RNA secondary structure. The secondary structures of ribosomal 5S RNA (58, 220, 222-224) and 5.8S RNA (225) have been repeatedly and extensively reviewed. Although there are minor differences between recent models with respect to length of helical segments and the disposition of single stranded loops or bulges, most are characterized by a three-stem arrangement (including four helical segments for prokaryotes and five for eukaryotes), except for the cloverleaf model (69, 70) which has four stems. The two model types have been adapted to the primary base sequence of wheat germ 5S RNA in Figures 12, and 13. The major difference between the two models is the base pairing arrangement in the middle of the primary sequence: i . e . , helices I I 184 and I I I in the three-stem model, or the single helix AA1 in the four-stem cloverleaf model. As previously noted, the present results require a G*C-A*U-G*C base pair sequence in both the intact 5S RNA and in the RNase Tl-resistant cleavage fragment. However, no such base pair arrangement can be found in the cloverleaf model. Moreover, stem A of the cloverleaf model includes three G*U pairs, but no G*U base pair protons are NOE-observable in the corresponding RNase T l-res istan t cleavage fragment. On the other hand, NOE-connectivity and ring current calculations support the Ci8G60- Al9U59-C2()G58 assignment to the "tuned" helix I I of the three-stem model. The alternative Gg8Cio7-A69u106“270Cl05 segment in helix V is ruled out by the continued presence of the signals in the RNase T l-resistan t cleavage fragment, in which bases 90-120 have been removed. The experimentally absence of G*U pairs in the RNase Tl-resistant cleavage fragment further supports the existence of helix I I , in which no G*U's are present. The present results thus provide the most direct evidence to date for the existence of the "tuned" helix I I , and thus strongly support a three-stem secondary structural model for 5S RNA. The exact number of base pairs in helix I I remains to be determined. First, RNase Sj-cleavage of positions 8-18 (172) indicates that helix II may not contain as many as five consecutive base pairs. Second, we do not detect an A*U base pair corresponding to AigUg2 in the NMR spectrum of either the intact molecule or the RNase Tl-resistant cleavage fragment. Furthermore, a few resonances remain after subtraction of the common arm resonances and the assigned peaks G2 , C, and K from the RNase T l-resistant fragment 185 spectrum, suggesting that the revised Fox/Woese model is incomplete. Although there are no less than five G*U base pairs in the secondary structure of intact wheat germ 5S RNA, only one or two G*U's are observed by *H NOE experiments. This anomaly may be explained by: overlap of peaks at positions R, V, and possibly W; absence of G*U pairs from the actual secondary structure; or high la b ility for chemical exchange for the remaining G*U pairs, so that their NMR signals are too broad for us to see, even at 2°C. 10. Limitations of the NOE method, and implications for future work. Base pair sequences constructed from NOE connectivities have been quite accurate for tRNA solution structures 142, 202-205) and for some small DNA's (226, 227). The principal lim itation of the method occurs when two or more peaks overlap, as in peak K in Figure 49. Peak K contains approximately 8-10 resonances, as judged from its relative intensity and from scrutiny of the spectra from temperature-induced melting in Figure 53. Thus, a base paired segment inferred from NOE-connectivity ends when one of the resonances in peak K is involved (see Figure 51B-E); irradiation at peak K gives too many possible neighbor base pairs to choose the next neighbor in the segment. For example, i f peak K contains 8 resonances, then there could be as many as 16 NOE-connections from base pairs above and below the base pairs whose protons a ll share the same chemical s h ift at K, making base pair sequence assignments very ambiguous. Moreover, because peak K is so .186 large, even a small degree of spillover of decoupler power from irradiation of a nearby peak can contribute an NOE-difference peak at K, whether the K protons are dipole-coupled to the irradiated peak or not— Fig. 51E could present such a case. From the appearance of two aromatic peaks in the NOE difference spectrum, and by variation of buffer conditions, we conclude that the overlapping resonances labeled D, E must represent hydrogen-bond imino protons from two different A*U base pairs. Based upon their disappearance from the spectrum of the RNase T l-resistant cleavage fragment, D,E might be A 3U116 an(* ^4^115* Although the two resonances are slig htly separated when Na+ and Mg2+ are removed (Fig. 50, bottom), direct irradiation of either peak will produce a small difference peak at the other resonance position due to spillover of the presaturation power. A possible solution to this problem would be to employ a two-dimensional NOE experiment (NOESY) (227 ), although such experiments become quite d iffic u lt for spectra with peaks having 50 Hz line widths. F inally, i t is important to recognize that NOE experiments detect protons that are spatially close to the proton whose resonance is irradiated. In particular, secondary base pairs cannot be distinguished from tertiary base pairs. Thus, it is possible to infer an incorrect secondary base pair sequence, i f one of the base pairs happens to be a tertia ry rather than a secondary pair. Because of the above problems, NOE experiments must be supported by other procedures which sh ift or simplify the *H NMR spectrum: use of enzymatic cleavage fragments (117-119; this work), variation in 187 temperature, change in salt conditions, site-specific spin-labeling (Lee & Marshall, unpublished), comparison of 5S RNA spectra from from organisms whose 5S RNA's d iffe r in only a few primary positions (Chen & Marshall, unpublished), etc. For wheat germ 5S RNA, several other RNase T1 fragments designed to isolate other secondary structural segments are presently under investigation. In particular, using denaturing gel filtr a tio n chromatography, fragment can be dessociated into three individual strands (base 1-25, 26-87, or 89, 90-120). Mixture of any two of the three strands can provide the samples to detect the proposed helix I , I I , and/or helix IV and V. CHAPTER VI SUMMARY Ribosomal 5S RNA is a structural and functional component in ribosome which has received intensive investigation. A widely spreade speculation is that there is a universal secondary structure for 5S RNA and hence it has universal function in the protein synthesis. There are many secondary structures proposed for 5S RNA. Thus, use of biophysical chemistry methods to investigate into the PNA secondary structure is significant in understanding the protein synthesis at molecular level . Wheat germ 5S, (species Triticum Aestivum), is chosen to reveal the universal structure of 5S RNA from eukaryotic organisms. A variety of physical and spectroscopic techniques is used to study its secondary structure. From FT-IR experiments, about 36 base pairs were detected which contained 14 A*U, 17 G*C, and 5 G*U base pairs. From DSC experiments, its thermodynamic parameters were measured from which Nishikawa model was ruled out because the theoretical enthalpy and free energy are much higher than experimental ones. From 500 MHz *H NMR experiments, 5 A*U, 1 G*U, and several G*C were discovered. A combination of proton homonuclear NOE experiments, variation of salt 188 189 conditions, temperature-induced melting, and 1 h NMR of RNase T1 enzymatic cleavage fragments yields a rela tiv e ly d e fin itiv e assignment of previously unobserved "tuned" helix I I region of the Fox/Woese model of the secondary structure of wheat germ 5S RNA. Although the three stems model is more favorable, its structural features are not complete. Thus further experiments are required for revealing its structure. Future work may include: crystalizing fragment Bj., detecting helix I , IV, and V, testing the involvement of helix I I I in the A form-B form transition, finding the conformational changes upon the binding to proteins, and comparing with sim ilar 5S RNA from other plant, etc.. REFERENCES 1. Nomura, M. (1970) B acteriol. Rev. 34, 228-277. 2. S to ffle r, G. (1974) In Ribosomes, ed. M. Nomura, A. Tissieres, P. Lengyel, pp. 615-667. Cold Spring Harbor, NY: Cold Spring Harbor Lab. 3. Wittmann, H. G. (1974) see Ref. 2, pp. 93-114. 4. Wittmann, H. G., Wittmann-Liebold, B. (1974) see Ref. 2, pp.115- 140. 5. Nomura, M., Morgan, E. A., Jaskunas, S. R. (1977) Ann. Rev. Genet. 11, 297-347. 6. Kurland, C. G. (1977) In Molecular Mechnism of Protien Synthesis, ed. H. Weissbach, S. Pestka, pp 81-116. NY: Academic. 7. S to ffle r, G. Wittmann, H. G. (1977) see Ref. 6. pp. 117-202. 8. Kurland, C. G. (1977) Ann. Rev. Biochem. 46, 173-200. 9. Brimacombe, R., S to ffle r, G., Wittmann, H. G. (1978) Ann. Rev. Biochem. 47, 217-249. 10. Stryer, L. (1981) in Biochemistry. Second ed. pp. 653. W. H. Freeman and Company, San Francisco. 11. Wool, I . G. (1979) Ann. Rev. Biochem. 48, 719-754. 12. Van Holde, K. E., Hill,W. E. (1974) see Ref. 2, pp. 53-91. 13. Stuhrmann,H. B., Haas, J ., Ib e l, K., De Wolf, B., Koch, M. H. J ., et al. (1976) Proc. Natl. Acad. Sci. USA. 73, 2379-2383. 14. Lake, J. A. (1981) S cien tific American 245, 84-97. 15. Steitz, J. A.,and Jakes, K., (1975) Proc. Natl. Acad. Sci. USA. 72, 4734-4738. 16. Kaziro, Y., (1978) Biochem. Biophys. Acta 505, 95-127. 190 191 17. Gupta, S. L ., Waterson, J ., Sopori, M., Weissman, S. M. and Lengyel, P., (1971) Biochemistry 10, 4410-4421. 18. Shine, J. and Dalgarno, L., (1974) Proc. Natl. Acad. Sci. USA. 71, 1342-1346. 19. Erdmann, V. A. (1976) in Progress in Nucleic Acid Research and Molecular Biology 18, (cohn, W. E ., ed.) pp. 45-90. Academic Press, New York. 20. Erdmann, V. A., Wolters, J ., Huysmans, E ., and De wachter, R. (1985) Nucleic Acids Research, 13, rl0 5 -rl5 3 . 21. Rosset, R., Monier, R. and Julien, J. (1964) Bull. Soc. chim. B io l. 46, 87. 22. Brownlee, G. G., Sanger, F. and Barrel 1, B. G. (1967) Nature (London) 215, 735. 23. Barrell, B. G. and Clark, B. F. C. (1974) "Handbook of Nucleic Acid Sequences." Joynson-Bruvvers Led., Oxford, England. 24. Nomura, M., Erdmann, V. A. (1970) Nature 228, 744-748. 25. Erdmann, V. A., Fahnestock,S., Higo, K., Nomura, M. (1971) Proc. Natl. Acad. Sci. USA. 68. 2932-2936. 26. Wrede, P. and Erdmann V. A. (1973) FEBS-Letters 33, 315-319. 27. Nierhaus, K. H.and Dohme, F. (1974) Proc. Natl. Acad. Sci. USA. 71, 4713-4717. 28. Horne, J. R. and Erdmann, V. A. (1972) Molec. Gen. Genet. 119, 337-344. 29. Gray, P. N., Garrett, R. A., Stoffler, G. and Monier, R. (1972) Eur. J. Biochem. 28, 412-421. 30. Garrett, R.A., Douthwaite, S., and Noller, H. F. (1981) Trends B iol. Sci. 6, 137-139. 31. Douthwaite, S., Christensen, A. and Garrett, R. A. (1982) Biochemistry 27, 2313-2320. 32. Azad, A. A. And Lane, B. G. (1973) Can. J. Biochem. 51, 1669-1672. 33. Nichols, J. L. and Wijesinghe, W. (1978) 56, 760-764. 34. Azad, A. A. (1979) Nucleic Acids Res. 7, 1913-1929. 192 35. Buetow, D. E. and Wood, W. M. (1978) Nature (London) 228, 744-748. 36. Erdmann, V. A., Appel, B., Digweed, M., Kluwe, D., Lorenz, S., Luck, A., Schreiber, A. and Schuster, L. in Genetics and Evlution of RNA Polymerase, tRNA and Ribosomes, ed. by Osawa, S ., Ozeki, H., Uchida, H. and Yura, T. (1980) pp. 553-568. University of Tokyo Press. 37. Horne, J. R. and Erdmann, V. A. (1974) FEBS-Letters 42, 42. 38. Horne, J. R. and Erdmann, V. A. (1973) Proc. N atl. Acad. Sci. USA. 70, 2870. 39. Gaunt-Klopter, M. and Erdmann, V. A. (1975) Biochem. Biophys. Acta 390, 226-233. 40. Grummt, F ., Grummt, E. and Erdmann, V. A. (1974) Eur. J. Biochem. 43, 343. 41. Vazquez, D. (1974) FEBS-Letters 40, 563. 42. Marcus, A. (1970) J. B iol. Chem. 245, 955. 43. Nishikawa, K. and Takemura, S. (1974) J_. Biochem. 76, 935-947. 44. Chladek, S. (1971) Biochem. Biophys. Res. Commun. 45, 695. 45. Pace, B., Matthews, E. A., Johnson, K. D., Cantor, C. R. and Pace, N. R. (1982) Proc. N a tl. Acad. S ci. USA. 79, 36-40. 46. Zagorska, L ., Duin, J. V ., Noller, H. F., Pace, B., Johnson, K. D. and Pace, N. R. (1984) J. B iol. Chem. 259, 2798-2802. 47. Forget, B. G. and Weissman, S. M. (1967) Science 158, 1695-1699. 48. Ofengand, J. and Hences, C. (1969) J_. B io l. Chem. 244, 6241-6253. 49. Erdmann, V. A., Sprinzl, M. and Pongs, 0. (1973) Biochem. Biophys. Res. Commun. 54, 942-948. 50. Shimizu, N., Hayashi, H. and Miura, K. (1970) J. Biochem 67, 373- 387. 51. Forget, B. G. and Weissman, S. M. (1967) Nature 213, 878-882. 52. Weinberg, R. A. and Penman, S. (1968) cL Mol. B io l. 38, 289-304. 53. Kel've, M., Metspalu, A., Lind, A., Saaema, M. and Villems, R. (1978) Molek. Biol. (RUSS) 12, 695-699. 193 54. Aubert, M., Scott, J. F ., Reynier, M. and Monier, R. (1968) Proc. Natl. Acad. Sci. USA. 61, 292-299. 55. Lecandiou, R. and Richards, E. G. (1975) Eur. J. Biochem. 57, 127-133. 56. Kinie, M. J. and Moore, P. B. (1982) Nucleic Acids Res. 10, 4973- 4983. 57. Kao, T. H. and Crothers, D. M. (1980) Proc. Natl. Acad. Sci. USA. 77, 3369-3364. 58. P ieler, T ., Digweed, M. and Erdmann, V. A. (1984) in Gene Expression, Alfred Benzon Symposium 19, ed. by Clark, B. F. C. and Petersen, H. U. pp. 353-376. 59. Peattie, D. A., Douthwaite, S., Garrett, R. A. and Noller, H. F. (1981) Proc. N a tl. Acad. S ci. USA. 78, 7331-7335. 60. Stoffler-Meilicke, M., Stoffler, G., Odom, 0. W., Zinn, A., Kramer, G. and Hardesty, B. (1981) Proc. N atl. Acad. Sci. USA. 78, 5538-5542. 61. Shatsky, I. N., Evstafieva, A. G., Bystrova,T. F., Bogdanov, A. A. and Vasiliev, V. 0. (1980) FEBS-Letters 121, 97-100. 62. Wagner, R. and G arrett, R. A. (1978) Nucleic Acids Res. 5, 4065- 4075. 63. Boublik, M. and Hellmann, W. (1978) Proc. N atl. Acad. Sci. USA. 75, 2829-2833. 64. Tischendorf, G. W., Zeichardt, H. and Stoffler, G. (1974) Mol. Gen. Genet. 134, 187-208. 65. Stoffler, G. and Stoffler-Meil icke, M (1981) in International Cell Biology 1980/81, ed. Schweiger, H. G. (Springer, B erlin), pp. 93-102. 66. Stahl, D. A., Luehrsen, K. R ., Woese, C. R. and Pace, N. R. (1981) Nucleic Acids Res. 9, 6129. 67. De Jong, E. and Pijpers, F. W. (1985) Analytica Chemica Acta 167, 97-109. 68. Fox, G. E. and Woese, C. R. (1975) Nature 256, 505-507. 69. Luoma, G. A. and Marshall, A. G. (1978) J. Mol. B iol. 125, 95-105 70. Luoma, G. A. and Marshall, A. G. (1978) Proc. Natl. Acad. Sci. USA. 75, 4901-4905. 194 71. Nishikawa, K. and Takemura, S. (1974) FEBS-Letters, 40, 106-108. 72. Luehrsen, K. R. and Fox, G. E. (1981) Proc. N atl. Acad. Sci. USA. 78, 2150-2154. 73. Deli has, N. and Andersen, J. (1982) Nucleic Acids Res. 10, 7323- 7344. 74. Mackey, R. M., Spencer, D. F ., Schnare, M. N., D o o little , W. F. and Gray, M. W. (1982) Can. J_. Biochem. 60, 480-489. 75. De Wachter, R., Chen, M.-W. and Vandenberghe, A. (1982) Biocheimie 64, 311-329. 76. Studnicka, G. M., Eigerling, F. A., Lake, J. A. (1981) Nucleic Acids Res. 9,1885-1904. 77. Osterberg, R. Sjoberg, B. and Garrett, R. A. (1976) Eur. J^. Biochem. 68, 481-487. 78. Kearns, D. R. and Wang, Y. P. (1974) J_. Mol. B io l. 87, 755. 79. Holley, R. W., Apgar, J ., Everett, G. A., Madison, J. T ., Marquisee, M., Merrill, S. H., Penswick, J. R. and Zamir, A. (1965) Science 147, 1462-1465. 80. Kim, S. H., Suddath, F. L ., Quigley, G. J ., Mcpherson, A., Sussmann, J. L ., Wang, A. H. J ., Seeman, N. C. and Rich, A. (1974) Science 185, 435-440. 81. P ieler, T. and Erdmann, V. A. (1982) Proc. Natl. Acad. Sci. USA. 79, 4599-4603. 82. Douthwaite, S., Christensen, A. and Garrett, R. A. (1982) Biochemistry 27, 2313-2320. 83. Bear, D. G., Schleich, T., Noller, H. F. and Garrett, R. A. (1977) Nucleic Acids Res. 4, 2511- 2526. 84. Fox J. W. and Wong, K. P. (1978) J^. B io l. Chem. 253, 13-20. 85. Fanning, T. G. and Traut, R. R. (1981) Nucleic Acids Res. 9, 933- 1004. 86. Morikawa, K., Kawakami, M. and Takemura, S. (1982) FEBS-Letters 145, 194-196. 87. Sankoff, D., Morel, C. and Cedergren, R. J. (1973) Nature New B io l. 245, 232-234. 88. Hori, H. (1975) J. Mol. Evol. 7, 75-86. 195 89. Vandenberhe, A., Chen, M.-W., Dams, E ., De Baere, R., De Roeck, E ., Huysmans, E ., and De Wachter, R. (1984) FEBS-Letters 171, 17-23. 90. Hori, H. and Osawa, S. (1979) Proc. N atl. Acad. Sci. USA. 76, 381-385. 91. Aubert, M., Bellemare, G., and Monier, R. (1973) Biochemie 55, 135. 92. Noller, H. F. (1974) Biochem. 13, 4694-4703. 93. Litt, M. (1969) Biochem. 8, 3249-3253. 94. Peattie, D. A. and Gibert, W. (1980) Proc. Natl. Acad. Sci. USA. 77, 4679-4682. 95. Woese, E. R ., Magrum, L. J ., Gupta, R., Siegel, R. B. and Stahl, D. A. (1980) Nucleic Acids Res. 8, 2275-2293. 96. Silberkang, M., RajBhandary, U. L ., Luck, A. and Erdmann, V. A. (1983) Nucleic Acids Res. 11, 605-617. 97. Rabin, D. and Crothers, D. M. (1979) Nucleic Acids Res. 7, 689-703. 98. Wollenzien, P. L., Hearst, J. E., Thammana, P. and Cantor, C. R. (1979) J. Mol_. B io l. 135, 255-269. 99. Thammana, P., Cantor, C. R., Wollenzien, P. L. and Hearst, J. E. (1979) J. MoK B io l. 135, 271-283. 100. Turner, S., Thompson, J. F ., Hearst, J. E. and Noller, H. F. (1982) Nucleic Acids Res. 10, 2839-2849. 101. Turner, S ., Noller and H. F. (1983) Biochem. 22, 4159-4165. 102. Thompson, J. F. and Hearst, J. E. (1983) Cell 32, 1355-1365. 103. Hearst, J. E. (1981) Ann. Rev. Biophys. Bioeng. 10, 69-86. 104. Uhlenbeck, 0. C ., B ailer, J. and Doty, P. (1970) Nature (London) 225, 508. 105. Lewis, J. B. and Doty, P. (1970) Nature (London) 225, 510. 106. Soot, M. B., Saarma, R. L ., Willems, E., Lind, A. J. and Vasilenko, S. K. (1974) Mol. Biol. (USSR) 8, 723. 107. Brownlee, G. G., Sanger, F. and Barrel 1, B. G. (1968) J. Mol. Biol. 34, 379-412. 196 108. Jordan, B. R. (1971) J. Mol. Biol. 55, 423. 109. Vigne, R., Jordan, B. R. and Monier, R. (1973) J. Mol. Biol. 76, 303-311. 110. Ross, A. and Brimacombe, R. (1979) Nature 281, 271-276. 111. Vassilenko, S. K. and Ryte, V. C. (1975) Biokhimya 40, 578-583. 112. Vassilenko, S. K., Carbon, P., Ebel, J. P. and Ehresmann, C. (1981) J. Mol- B io l. 152, 699-721. 113. Douthwaite, S. and G arrett, R. A. (1981) Biochemistry 20, 7301- 7307. 114. Apple, B., Schreiber, A., Schuster, L., Wrede, P., Erdmann, V. A., Stulz, J., Ackermann, T. (1979) XI. International Congress of Biochemistry, Toronto, Canada, Abstract 02-2-H28. 115. Speek, M., Ustav, M. B., Lind, A. J. and Saarma, M. J. (1980) Bioorg. Khim. 6, 1877-1880. 116. Douthwaite, S. and Garrett, R. A. (1981) Biochemistry 20, 7301- 7307. 117. Kime, M. J. and Moore, P. B. (1983a) Biochemistry 22, 2615-2622. 118. Kime, M. J. and Moore, P. B. (1983b) Biochemistry 22, 2622-2629. 119. Kime, M. J ., Gewirth, D. T ., and Moore, P. B. (1984) Biochem. 23, 3559-3568. 120. Fox, J. W. and Wong, K.-P. (1979) J. B iol. Chem. 254, 10139- 10144. 121. Chang, L .-H ., Burkey, K. 0 ., Alben, J. 0. and Marshall, A. G. (1984) Biochemistry 23, 3659-3662. 122. Boedtker, H. and Kelling, D. G. (1967) Biochem. Biophys. Res. Commun. 29, 759-766. 123. Van, N. T ., Nazar, R. N. and S itz, T. 0. (1977) Biochemistry 16, 3754-3759. 124. Fresco, J ., Klotz, L. and Richards, E. G. (1963) Cold Spring Harbor Symp. Quant. B io l. 18, 83-90. 125. Luoma, G. A., Burns, P. D., Bruce, R. E. and Marshall, A. G. (1980) Biochemistry 19, 5456-5462. 197 126. Appel, B., Erdmann, V. A., Stulz, J. and Ackermann, Th. (1979) Nucleic Acids Res. 7, 1043- 1057. 127. Bohm, S., Fabian, H., Venyaminovc, S. Y., Matveev, S. V., Lucius, H., Welfle, H. and Filimonov, V. V. (1981) FEBS Letters 132, 357-361. 128. Burkey, K. 0 ., Marshall, A. G. and Alben, J. 0. (1983) Biochemistry 22, 4223-4229. 129. Chen, M. C ., Giege, R., Lord, R. C. and Rich, A. (1978) Biochemistry 17, 3134-3138. 130. Redfield, A. G. (1978) Methods ENzymol. 49, 253-270. 131. Redfield, A. G., Kunz, S. D., and Ralph, E. K. (1975) vK Magn. Resonance 19, 114-117. 132. Redfield, A. G. and Gupta, R. K. (1971) J. Chem. Phys. 54, 539- 542. 133. Hore, P. J. (1983) d_. Magn. Res. 54, 539-542. 134. Plateau, P. and Gueron, M. (1982) J. Am. Chem. Soc. 104, 7310- 7311. 135. Gupta, R. K. (1976) cL Magn. Reson. 24, 461. 136. Roth, K., Kimber, B. J ., and Feeney, 0. (1980) J_. Magn. Resonance 41, 302-309. 137. Wong, Y. P., Kearns, D. R., Reid B. R. and Shulman, R. G. (1972) J. Mol. B iol. 72, 741-749. 138. Kearns, D. R .,P atel, D., Shulman, R. G., and Yamane, T. (1971) J. Mol. Biol. 61, 265-270. 139. Burns, P. D., Luoma, G. A. and Marshall, A. G. (1980) Biochem. Biophys. Res. Commun. 96, 805-811. 140. Schimmel, P. R. and Redfield, A. G. (1980) Ann. Rev. Biophys. Bioeng. 9, 181-221. 141. Reid, B. R. (1981) Ann. Rev. Biochem. 50,969-996. 142. Johnston, P. D. and Redfield, A. G. (1981) Biochemistry 20, 1147 -1156. 143. Noggle, J. H. and Schirmer, R. E. (1971) The Nuclear Overhauser Effect: Chemical Applications, Academic Press, New York. 198 144. Bothner-By, A. A. (1979) in Biological Applications of Magnetic Resonance (Shulman, R. G., ed.) pp. 177. Academic Press, New York. 145. Sanchez, V., Redfield, A. G., Johnston, P. D. and Tropp, J. (1980) Proc. Natl. Acad. Sci. USA. 77, 5659-5662. 146. Johnston, P. D. and Redfield, A. G. (1978) Nucleic Acids Res. 4, 3599-3615. 147. Roy, S., Papastavros, M. Z. and Redfield, A. G. (1982) Biochemi stry 21, 6081-6088. 148. Luoma, G. A., Herring, F. G. and Marshall, A. G. (1982) Biochemistry 24, 6591-6598. 149. Tinoco, I. J ., Uhlenbech, 0. C. and Levine, M. D. (1971) Nature 230, 332-337. 150. Fresco,J. R., Alberts, B. M. and Doty, P. (1960) Nature 188, 98. 151. Tinoco, I. J., Borer, P. N., Dengler, B., Levine, M. D., Uhlenbach, 0. C., Crothers, D. M. and G ralla, J. (1973) Nature New Biology 246, 40-41. 152. Studnicka, G. M., Rahn, G. M., Cummings, I . W. and Salser, W. A. (1978) NUcleic Acids Res. 5, 3365-3387. 153. Salser, W. (1977) Cold Spring Harbor Symp. Quent. B iol. 42, 985- 1002. 154. L ev itt, M. (1969) Nature 224, 759. 155. Privalov, P. L., Filimonov, V. V., Venkstern, T. V. and Bayev, A. A. (1975) J_. Mol. B io l. 97, 279-288. 156. Jackson, W. M. and Brandts, J. F. (1970) Biochemistry9 , 2294- 2301. 157. Privalov, P. L. and Filimonov, V. V. (1978) J. Mol. Biol. 122, 447-464. 158. Privalov, P. L. and Khechinashvili, N. N. (1974) J. Mol. Biol. 86, 665-684. 159. Matveev, S. V ., Filimonov, V. V. and Privalov, P. L. (1982) Mol. B io l. 16, 990-999. 160. Anson, M. L. and Mirsky, A. E. (1934) J_. Gen. Physiol. 17, 399. 161. Freire, E. and Biltonen, R. L. (1978) Biopolymers 17, 463-479. 199 162. Freire, E. and Biltonen, R. L. (1978) Biopolymers 17, 481-496. 163. Filimonov, V. V., Potekhin, S. A., Matveev, S. V. and Privalov, P. L. (1982) Mol. B io l. 16, 435-444. 164. Richards, E. 6 ., Geroch, M. E., Simpkins, M. and Lecanidou, R. (1972) Biopolymers 11, 1031. 165. Cantor, C. R. (1968) Proc. N a tl. Acad. S ci. USA. 59, 478. 166. Gray, P. N. and Saunders, G. F. (1971) Biochem. Biophys. Acta 256, 60. 167. Connors, P. G. and Beeman, W. W. (1972) J_. Mol. Biq1 • 71> 31 • 168. Moore, P. B., Enghelman, D. M., Schoenborn, B. P. (1974) in Ribosome (Nomura, M., Tissieres, A., Lengyel , P., eds.) pp. 601-614, Cold Spring Harbor, New York. 169. Hori, H. and Osawa, S. (1979) Proc. Natl. Acad. Sci. USA. 76, 381-385. 170. Schwartz, R. M. and Dayhoff, M. 0. (1978) in Atlas of Protien Sequence and Structure (Dayhoff, M. 0., ed.) Vol. 5, Suppl. 3, pp. 28-43, National Biomedical Research Foundation, Waashington, D. C .. 171. Payne, P. I . and Dyer, T. A. (1976) Eur. J_. Biochem. 71, 33-38. 172. Barber, C. and Nichols, J. L. (1978) Can. J. Biochem. 56, 357-364. 173. Soave,C., Nucca, R ., Sala, E ., V io tti, A. and Calante, E. (1973) Eur. J. Biochem. 32, 392-400. 174. Mackay, R. M., Spencer, D. F., D o o little , W. F. and Gray, M. W. (1980) Eur. J_. Biochem. 112, 561-576. 175. Privalov, P. L. (1974) FEBS Letters 40, S140-S153. 176. Krishnan, K. S. and Brandts, J. F. (1978) Methods Enzymol. 49, 3-14. 177. Schott, F. J ., Grubert, M., Wangler, W. and Ackermann, T. (1981) Biophys. Chem. 14, 25-30. 178. Neumann, E. and Ackermann, Th. (1969) vK Phys. Chem. 73, 2170. 179. Filimonov, V. V. and Privalov, P. L. (1978) J. mol. B iol. 122, 465-470. 200 180. Rubin, G. (1975) Methods Cel 1. B io l. 12, 45-64. 181. Salemink, P. J. M., Raue, H. A., Heerschap, A., Plant, R. J. and Hilbers, C. W. (1981) Biochemistry 20, 265-272. 182. Marshall, A. G. and Smith, J. L. (1980) Biochemistry 19, 5955- 5959. 183. Burkey, K. 0. and Marshall, A. G. (1983) unpublished. 184. Wolfe, F. H. and Kay, C. M. (1967) Biochemistry 6, 2853-2861. 185. Soave, C., Nucca, R., Sala, E., Viotti, A. and Galante, E. (1973) Eur. J_. Biochem. 32, 392-400. 186. Lau, R. Y ., Kennedy, T. D. and Lane, B. G. (1974) Can. J^. Biochem. 52, 1110-1123. 187. Luoma, G. A. (1980) Ph. D. Thesis, Univ. of B ritish Columbia, Vancouver. 188. Ornstein, R. L. and Fresco, J. R. (1983) Biopolymers 22, 2001- 2016. 189. Gralla, J. and Crothers, D. M. (1973) J_. Mol. Biol. 73, 497. 190. L i, S .-J ., Burkey, K. 0 ., Louma, G. A., Alben, J. 0 ., and Marshall, A. G. (1984) Biochemistry 23, 3652-3658. 191. Thomas, G. J. Jr (1969) Biopolymers 7, 325-334. 192. Griffiths, P. R., Foskett, C. T., and Curbello, R. (1972) Appl. Spectros. Rev. 6, 31-78. 193. Fellget, P. (1958) J_. Phys. Radium 19, 187. 194. Schernau, U. and Ackermann, Th. (1977) Biopolymers 16, 1735-1745. 195. Tsuboi, M. (1969) Appl. Spectros. Rev. 3, 45-90. 196. Bevington, P. R. "Data Reduction and Error Analysis for the Physical Sciences", McGraw-Hill: N.Y., 1969: pp. 235-242. 197. Savitzky, A. and Golay, M. J. E. (1964) Anal. Chem. 36, 1627-1639. 198. Pvivalov, P. L ., Khechinashvili, N. N. and Atanasov, B. P. (1971) Biopolymers 10, 1865-1890. 201 199. Krishan, K. S. and Brandts, J. F. (1978) Methods in Enzymol. 49, 3-14. 200. Kearns, D. R ., Patel, D., and Shulman, R. G. (1971) Nature (London) 229, 338-339. 201. Johnston, P. D. and Redfield, A. G. (1978) Nucl. Acids Res. 5, 3913-3927. 202. Hare, D. R. and Reid, B. R. (1982) Biochemistry 21, 5129-5135. 203. Heerschap, A., Haasnoot, C. A. G., and Hilbers, C. W. (1982) Nucl. Acids Res. 10, 6981-7000. 204. Heerschap, A., Haasnoot, C. A. G., and Hilbers, C. W. (1983a) Nucl. Acids Res. 11, 4483-4499. 205. Heerschap, A., Haasnoot, C. A. G., and Hilbers, C. W. (1983b) Nucl. Acids Res. 11, 4501-4520. 206. Ernst, R. R. and Anderson, W. A. (1966) Rev. Sci. Instrum. 37, 93. 207. Farrar, T. C. and Becker, E. D. (1971) "Pulse and Fourier Transform NMR", Academic Press, New York. 208. Gupta, R. K., F e rre tti, J. A., and Becker, E. D. (1974) J^. Magn. Resonance 16, 505. 209. Bleich, H. E. and Glasel, J. A. (1975) J. Magn. Resonance 18, 401. 210. Redfield, A. G. and Kunz, S. D. (1979) in "NMR and Biochemistry" (Opella, S. J. and Lu, P., Eds.) Marcel Dekker, N.Y., pp. 225-239. 211. Reid, B. R. (1981) Ann. Rev. Biochem. 50, 969-996. 212. Raiford, D. S., Fisk, C. L. & Becker, E. D. (1979) Anal. Chem. 51, 2050-2051. 213. van Geet, A. L. (1970) Anal. Chem. 42, 679-680. 214. Haasnoot, C. A. G., Den Hartoy, J. H. J ., De Rooy, F. M., Van Boom, J. H., and Altona, C. (1980) Nucl. Acids Res. 8, 169-181. 215. Roy, S., and Redfield, A. G. (1981) Nucl. Acids Res. 9, 7073-7083. 216. Roy, S., and Redfield, A. G. (1983) Biochemistry 22, 1386-1390. 202 217. Scheek, R. M., Boelens, R ., Russo, N., van Boom, J. H., and Kaptein, R. (1984) Biochemistry 23, 1371-1376. 218. Ladner, J. E., Jack, A., Robertus, J. D., Brown, R. S ., Rhodes, D., Clark, B. F. C., 4 Klug, A. (1975) Proc. Natl. Acad. Sci. U.S.A. 72, 4414-4418. 219. Sussman, J. L ., & Kim, S.-H. (1976) Biochem. Biophys. Res. Commun. 68, 89-96. 220. Delihas, N., Andersen, J ., & Singhal, R. P. (1984) Prog. Nucleic Acids Res. 31, 161-190. 221. A rter, D. B. and Schmidt, P. G. (1976) Nucl. Acids. Res. 3, 1437-1447. 222. Eigen, M., Lindemann, B., Winkler-Oswatitsch, R., & Clarke, C. H. (1985) Proc. Natl. Acad. Sci. U.S.A. 82, 2437-2441. 223. Curtiss, W. C. & Vournakis, J. N. (1984) J^. Mol. Evol. 20, 351-361. 224. Ohama, T ., Kumazaki, T ., Hori, H., & Osawa, S. (1984) Nucleic Acids Res. 12, 5101-5108. 225. Vaughn, J. C ., Sperbeck, S. J ., Ramsey, W. J ., & Lawrence, C. B. (1984) Nucleic Acids Res. 12, 7479-7502. 226. Hare, D. R ., Wemmer, D. E ., Chou, S.-H ., Drobny, G., and Reid, B. R. (1983) J. Mol_. B io l. 171, 319-336. 227. Haasnoot, C. A. G., Heerschap, A., and Hilbers, C. W. (1983) J_. Am. Chem. Soc. 105, 5483-5484. 228. Reid, B. R., McCollum, L., Ribeiro, N. S., Abbate, J., and Hurd, R. E. (1979) Biochemistry 18, 3996-4005. 229. Chang, L.-H., S.-J. Li, Ricca, T. L., Marshall, A. G. (1984) Anal. Chem. 56, 1502-1507. 230. Krishan, K. S. and Brandts, J. F. (1978) Methods in Enzymol. 49, 3-14. 231. Chang, L.-H. (1985) Ph. D. Thesis. The Ohio State Univ. 232. Alben, J. 0. (1982) in "Physical Techniques in Biological Research", (Rousseau, D. L ., ed.) Academic Press, New York. 233. Bauer, H. H., Christien, G. D., and O 'R eilly, J. E. (1978) in Instrumental Analysis. Allyn and Bacon, Inc. Boston, pp. 505- 509. APPENDIX NMR PULSE SEQUENCE PROGRAMS 203 204 The NMR programs RED.AU and REDADA.AU are used fo r acquiring normal *H NMR spectra. The pulse sequence programs in Bruker AM 500 are listed as follows. RED.AU ;Redfield pulse sequence for water suppression ; using low-power transmitter with precision attenuator. ;The excitation pulse lengths in the japproximate ratio 2 -1 -4 -1 -2 ', where individual adjustment ;of each pulse length is made to optimize water suppression. ;The asymmetry introduced by the pulse 2' improves the suppression. 1 ZE ;zero memory 2 D1 ; relaxation delay 3 P2 PHI ; fir s t pulse length (P2) with phase PHI 4 PI PHI A2 ;pulse length 1, 180 ‘ phase s h ift (A2) 5 P4 PHI ;pulse length 4 6 PI PHI A2 jpulse length 1, 180° phase sh ift 7 P3 PHI ;pulse length 2' (asymmetry) 8 G0=2 PH2 ;acquire FID 9 EXIT pu 1 = Afl AO A? A? A1 A1 AT AT (i.e ., with transmitter phase 0o,0\180o,180o,90o,90o,180M80°) PH2 = RO RO R2 R2 R1 R1 R3 R3 (i.e ., with receiver phase 0°,0°,180°,180°,90°,90°,270°,270°) ; Start sequence with pulses set for nomial 2-1-4-1-2 -.Adjust 01 for best suppression, then adjust P4, P2, P3 for ; further improvement, small changes in PI may also help. ;RD = PW = 0 ;P1 = 0.1/(frequency difference between water resonance and 01) REDADA.AU ;Redfield 214 pulse sequence coupled with ;the alternate delayed acquisition technique. 1 ZE 2 D1 3 P2 PHI 205 4 PI PHI A2 5 P4 PHI 6 PI PHI A2 7 P3 PHI 8 60=2 PH2 9 D1 10 P2 PHI 11 PI PHI A2 12 P4 PHI 13 PI PHI A2 14 P3 PHI 15 D3 ;half-cycle delay for ADA 16 G0=9 PH2 17 LO TO 2 TIMES C 18 EXIT PHI = AO AO A2 A2 A1 A1 A3 A3 PH2 = RO RO R2 R2 R1 R1 R3 R3 ;Parameters are the same as that of RED.AU ; except D3 = (2*P1 + P2 + P3 + P4)/2. -.Total scans = 2*C*NS where NS must be multiple of 8. Typical parameters are: carrier frequency offset 01 = 14212 Hz, spectral width SW = 23.97 ppm, acquisition time AQ = 0.175 sec, D1 = 0.3 sec, PI = 19.5 ^Jsec P2 = 38.5 >Usec, P3 = 38.0 Msec, P4 = 80.0 Ji sec, D3 = 97.5 jusec. P1331ADA.AU ;Program PI331 with alternate delay acquisition. ;0 ff resonance water suppression. ;90° hard pulse length is divided by 8. Set D2 = 1/2 * offset. 1 ZE 2 D1 ; relaxation time 3 PI PHI ;"1" pulse length 4 D2 5 P3 PHI ;"3" pulse length 6 D2 7 P3 PHI ;"3" pulse length 8 D2 9 PI PHI ;"1" pulse length 10 G0=2 PH2 11 D1 relaxatio n time for next delayed FID acquisition. 206 12 PI PHI 13 D2 14 P3 PHI 15 D2 16 P3 PHI 17 D2 18 PI PHI 19 D3 ; del ay time for FID acquisition 20 G0=11 PH2 21 LO TO 2 TIMES C 22 WR #1 23 EXIT PH1=A0 A2 PH2=R0 R2 Typical parameters for P1331ADA.AU are: With an in -lin e 20 dB attenuator (hard pulse), PI = 8.5>wsec, P3 = 25.5j4sec, D2 = 50.3U(sec, 01 = 14212, SW = 23.9ppm. For NOE experiments, program MULNOE.AU is used. MULNOE.AU -.Multiple Redfield NOE difference files ;Set up a FQLIST for each irradiation resonance. 1 ZE 2 WR #1 ;save date file 3 IF #1 ; increase data f ile number 4 LO TO 2 TIMES X 5 RF #1.001 ;start from file #1 6 RE #1 ;load data file 7 FL #8 ;decoupler frequency l i s t 8 IF #8 ; increase frequency l i s t f ile number 9 D3 ;set decoupler 02 from FL l i s t D3=0.1 sec. 10 D1 02 -.relaxation time 11 D2 HG ; decoupler on, homonuclear gated decoupling 12 D5 DO ; short w ait, decoupler off 13 P2 PHI 14 PI PHI A2 15 P4 PHI 16 PI PHI A2 17 P3 PHI 18 G0=10 PH2 19 LO TO 9 TIMES 2 20 WR #1 21 LO TO 9 TIMES C 22 WR #1 23 IF #1 24 IN=6 25 EXIT PHI = AO AO A2 A2 A1 A1 A3 A3 PH2 = RO R2 R2 RO R1 R3 R3 R1 ;Set NS to a multiple of 8 ;Set DP 30L-40L ;Total scans = 2*C*NS ;Set DS = 2-4 ;Set NE equals to number of frequency lis ts .