Function of DNA and RNA

DNA and RNA are chainlike macromolecules that function in the storage and transfer of genetic information. They are major components of all cells ~15% of the cells dry weight. Just as the amino acids are building blocks of proteins, the nucleotides are the monomeric unit of nucleic acids.

In this section of the course, we will first examine the structure and chemistry of the nucleotides and polynucleotides. Subsequently, we will examine the the non-covalent forces that form help nucleic acid helices and the types of helices that are possible. We will then move into an examination of nucleic acid higher order structure and its interactions with proteins

The Transforming Principle.

Why all the interest in Nucleic Acids? Avery's Experiment!

In 1928 Griffith observed that non-virulent strains of pneumonia causing bacteria became virulent when mixed with their heat killed pathogenic counterparts (Figure 1). Infectious Pnemococcus is a slimy bacteria-encased in a polysaccharide capsule. This capsule is required for human infection. Mutants exist that do not have this outer coat and are, therefore, nonpathogenic. Griffith found that injecting the non-virulent mutant into mice did not affect them, neither did the heat killed virulent strain. When, however, they were mixed together first and then injected, the mice died of pneumonia. Moreover, the pneumococcus isolated from the poor dead mouse were virulent! Something in the extract of heat-killed bacteria transformed the non-virulent to virulent bacteria. This something was later found by Avery and co-workers in 1944 to be nucleic acid. They fractionated the mixture and found that "a nucleic acid of the deoxyribose type is the fundamental unit of the transforming principle of Pneumococcus type III" (Figure 2). Avery's experiments showed that purified DNA has genetic specificity.

The components of Nucleic acids-Sugars and Bases.

The primary structure of a polynucleotide has some analogy with the primary structure of proteins. Like proteins nucleic acid polymers are chains of monomers, in the nucleic acids these monomer units are called nucleotides. In both RNA and DNA, each nucleotide repeating unit consists of three characteristic components:

Polynucleotides are polymers of this basic unit (FIGURE 3). This figure illustrates the massive number of adjustable angles in the backbone and highlights the conformational variablity in the nucleotide chain. We will concern ourselves in this section of the course with the structural bases and physical forces that determine the conformation of polynucleotides. We will then examine how these factors affect the biological function of polynucleotides.

Nucleic Acid Bases

The common nucleotides come in 5 different flavors. Adenine, guanine, cytosine and thymine in DNA, with uracil substituting for thymine in RNA. In strict analogy with the amino acids nucleotide bases have different functional groups and these differences determine the polynucleotide's structure and function (FIGURE 4). Note different shapes, different functional groups. Purines-two ring system, pyrimidines single rings. Both pu and py have prononced aromatic character. Hence the conformation of the bases is fixed! You are responsible for the strucuture of the bases!!! The precise 3D structure of the bases is known with high precision. The pyrimidines are planar molecules; purines are nearly planar with a slight pucker. The angles between the atoms in the rings of the 5 bases are very close to 180^o and do not change. The important differences to note in the functional groups of the bases are their capacities to donate or accept hydrogen bonds and their ability or lack thereof to have hydrophobic interactions.

These common bases can and often are derivatized by the cell for various structureal and functional uses. The most common derivatives are those which are methylated (FIGURE 5). A list of some other types are given in the TABLE 1. Most of these bizarre bases are found in tRNA. We will discuss them later.

One useful property of the nucleic acid bases is their strong UV absorbtion in the range from 250-280nm. Having determined the extinction coefficient(s) [e] of the bases, this property allows one to determine the concentration of the nucleic acid in solution.

OD= c e l

The knowledge of the ratio of the average extinction coefficients of the bases at two UV wavelengths, commonly 260 vs. 280 allows one to assess the purity of the nucleic acid solution. Additionally, the OD of a particular concentration of nucleic acid bases depends on the structure into which they are assembled. Hence absobance can be used as a structural probe of nucleic acids (see below).

Sugars.

Both RNA and DNA nucleotides contain a cylic furanose-type sugar; b-D-ribose in RNA and b-D-deoxyribose in DNA. The difference in chemical structure of the sugar occurs at the C2' position. In DNA the -OH group is replaced by an H. The five-membered sugar ring is generally non-planar. It can be puckered in an envelope (E) form-four of the five atoms lie in a plane and the fifth atom lies out of plane by ~0.5 angstrom; or in a twist (T) conformation where two adjacent atoms are displaced on opposite sides of a plane through the three other atoms (Figure 6 & 7). Conformations in which the atoms are displaced from these three or four atom planes are on the same side as the C5' are called endo and those on the opposite side are called exo.

In an unsubstituted furanose ring, the conformational changes do not proceed via planar intermediate, but the maximum pucker rotates virtually without potential energy barriers, giving rise to a potentially infinite number of conformations. This is illustrated by the pseudorotation cycle (FIGURE 8). This point will become important as we learn about the extraordinary flexibility of the DNA backbone.

Since the furanose ring in polynucleotides is unsymmetrically substituted, they do display a preferred puckering modes conformations. The pseudorotation angles determined from crystallographic investigations are not distributed evenly over the pseudorotation cycle. Instead they cluster in two domains, centered at C3'-endo (found in A-type helices) and C2'-endo (found in B-type helices). These two puckering modes are favored because they best alleviate steric clashes of the substituents of the sugar by placing them in a staggered, not eclipsed, conformation.

The interconversion between these two forms can occur by two paths, via an O4'-endo or O4'-exo intermediate. The route taken by most interconversions in via the O4'-endo intermediate. Consideration of why this is true illustrates the kinds of constraints on the backbone of a polynucleotide (Figure 9).

The O4'-exo intermediate of a substituted furanose ring causes the C5' and N-base to clash, whereas the O4'-endo allows adequate spacing between them. This serves as a paradigm for thinking about restrictions on the mobility of the backbone polynucleotides. THAT IS: even is single stranded polynucleotides there are restrictions on chain conformations!!!

So now we have talked about the bases and the sugar, now lets put them together to form a nucleoside and nucleotide.

Define: Nucleoside- base + sugar

Nucleotide- base + sugar, the sugar is phosphorylated at one of the free sugar hydroxyls.

The most common nucleotides are those which are phosphorylated at the C5', thus, 5' nucleotide monophosphate (Figure 10).

The formation of nucleotides adds another angle that we must be concerned about. This angle is that which the base pivots in relation to the nucleotide. This angle c (chi), defines the glycosyl bond. Relative to the sugar moiety, the base can adopt two main orientations about the C1'-N link, called either syn or anti. In the anti configuration, the bulky part of the base is pointing away from the sugar. The bulky part of the purine base is obvious; the six membered ring is pointed away from the sugar. For the pyrimidines the O2 on the C2 carbon is the bulky group; in anti, this group points away from the base. The definitions of syn and anti are somewhat non-specific and range as shown in FIGURE 11). Figure 12 gives a definition of torsion angles. The gross value of thec angle is correlated with the major observed sugar pucker. That is the steric hinderance in the syn conformation can be somewhat alleviated by placing the sugar in the C2'endo pucker. This is because the base and the C5' are then placed in an equatorial position and moved apart (FIGURE 13). other extremely relevant exception to the suggestion that anti configurations are preferred by the nucleotides is that of guanine. Both as a single nucleotide and in the alternating co-polymer G-C-G-C, G is in the syn configuration. This is probably due to the H-bonding of the N2 amino group of the guanine to its 5' phosphate.

Polynucleotides.

The condensation reaction between multiple 5' mononucleotides at the 3'-OH group forms a chain. In analogy with protein primary structure, 5'-3' chains have direction, a polarity, they have distinct and different ends. The features of the backbone repeat in sequence 5'-P, C5', C4', attached to sugar and base, C3', O3' and so on (FIGURE 3). As you might expect from the rest of this lecture, this linkage of nucleotides adds more angles to be concerned about. We will not discuss them in detail now, but will cover them later in double helical structure of DNA.

Before we begin describing the structural features of the oligonucleotides and double helices, a few remarks about the forces that govern base-base interactions are required. Two different interactions exist: a) those in the plane of the bases (horizontal); most commonly hydrogen bonds; b) those perpendicular to the base planes or "base stacking" effects; these are stabilized by London dispersion forces and the hydrophobic effect.

1) Hydrogen bonds.

Hydrogen bonds are electrostatic in character. In general a hydrogen bond

X - H ^... Y

is formed if a hydrogen atom connects two atoms of higher electronegativity. Since these bonds are electrostatic, their strength depends on the partial charges located on the component atoms in the bond. Under the influence of a hydrogen bond, the H becomes more electropositive and X,Y becoming more negative. This affect increases the affinity of X,Y for H and strengthens the interaction. If Y is the oxygen of the -OH group, the hydrogen attached to it is more positive and hence becomes a better donor.

As you should already know, the strength of an H bond is 20-30X weaker than a covalent bond. This weakness is reflected both in the bond's greater length and is relatively weak directionality (TABLE 2). This weakness should not be mistaken for insignificance!!!! Moreover, the "slop" in directionality is not limitless. In fact, the tolerable distortion level in the angle of a hydrogen bond measured between the vector of the bond and the angle of the X - H bond is less than 20^o. That is to say that the most favorable hydrogen bond angles are 180^o (FIGURE 14). Some bonds in the figure are distorted 27^o and hence is weaker than the others distorted less than 2^o.

2) Base Stacking.

Bases in solution pile up like coins in a roll. In aqueous solution, the bases in a single stranded oligonucleotide are stacked such that the base planes are separated by their van der Waals distance, 3.4 A, parallel to one another. Base stacking is the least understood but, undoubtedly most important force stabilizing helices.

Stacking is a diffusion controlled, additive, and stabilized by weak forces. The enthalpies associated with base stacking are favorable, while the entropy associated with the stacking of the bases is strongly unfavorable. The stacking reaction is overall favorable, however, since the entropy and enthalpy of the solvent are both strongly favorable. Stacking is made-up of two separate forces: hydrophobic effect and London dispersion forces.

Hydrophobic interactions.

If a hydrophobic base is dissolved in water, the water molecules cluster around it in an order fashion. This is caused by the fact that they cannot form H-bonds with the non-polar base and adopt an ordered "clathrate" structure to maximize H-bonding with itself. This ordering is a very unfavorable entropy change for the water. Burying this hydrophobic base in the stack, releases this water and results in an overall entropy gain for water.

The importance of the hydrophobic effect in helix formation is seen by considering the effect on the energetics of solvent interactions upon folding the non-polar bases into the helical structure. Folding the polar backbone atoms, into a regular structure has slightly unfavorable DH and slightly unfavorable TDS. The effect of burying the non-polar side chains is dramatic, the TDS of the solvent is incredibly favorable, and the DH is also favorable (owing to the now completely satisfied H-bonding potential of the solvent).

The bases stack upon one another at their van der Waals distance. It is at this distance where two molecules have an attraction for one another. This attraction, termed a van der Waals interactions is a gravitational forces. At too close a distance the electron of the two approaching molecules overlap, causing repulsion. At any given instant, the electronic charge distribution within atomic groups is asymmetric due to electron fluctuation. Therefore, dipoles created in on group of atoms polarize the electronic system of the neighboring atoms or molecules, thus inducing parallel dipoles that attract each other.

- A + ^... - B +

These forces are additive and are extremely distance dependent, falling off with the sixth power of distance (r⁶).

Since London dispersion attraction depends on formation of induced dipoles, the polarizable p electron cloud of the aromatic bases is extremely important. Therefore, stacking requires aromaticity of the bases-nonaromatic bases do not display stacking interactions. The strength of stacking interactions depends on how polarizable the p electron cloud of a base is. This in turn depends on the e- withdrawing or donating potential of the base substituents These determine the primary basis for base stacking which is the electron structure of the bases. Since all bases have different substituents, the stacking potential of the bases are all different. Additionally, the electronic structure of a base can be modified by chemical modification-e.g., alkylation, halogenation.

Stacking is BOTH base composition and base sequence dependent. In general, the stacking interactions of base paired nucleotide dimers containing G+C base pairs are more stable than those containing A+T base pairs (TABLE 3). Another generalization that can be made is that

3'pyrimidine^.purine5'

5'purine^.pyrimidine3'

is more stable than

3'purine^.pyrimidine5'

5'pyrimidine^.purine3'

For example (5'G-C3')₂ is 5kcal/mol more stable than (5'C-G3')₂. The origin of this sequence-energy correlation is seen in FIGURE 15. In the alternating purine,pyrimidine sequences, overlap between adjacent base pairs in a stack is much greater in B-DNA the pyrimidine, purine alternation, note the polar groups placed over the center of the p electron cloud, than the purine-purine stack.

Helical Formation

Let us now turn to how both base stacking and H-bond formation are involved in helix formation. The association of double stranded polynucleotide helices is a cooperative process. One can think of helix formation as analogous to closing a zipper. The first step is association of the a single base pair with a stability constant expressed as the product of a nucleation parameter, b, which essentially represents the unfavorable entropy of bringing together two ends of separate chains and a chain growth parameter, s, which represents the favorable aspects of hydrogen bonding and hydrophobic interactions (FIGURE 16). The addition of a second stacked base pair is not influenced by the nucleation parameter because the proximity effect is taken care of in the first association. This is the basis of the cooperativity of helix formation, base pair formation and stacking are influenced by the nearest neighbors, except for the first base pair formed. Because of the overall unfavorability of the nucleation constant, b, the energy of helix formation is unfavorable until after about three base pairs have formed. From then on, growth of the double helix is spontaneous, due mainly to the geometrical constraints of the sugar phosphate backbone, as implied by the sterochemistry of the nucleotide unit, whose preferred configurations is preset to form a double helix.

Also, the summation of these weak forces, over a number of nucleotides provide cooperative energy. This is illustrated schematically in FIGURE 17 and quantitatively expressed by K, the helical stability constant.

K = (b*s) + (n*s)

where b<1 (~10^-3 l M^-1), s= 10 @ 0^oC and 1 @ T_m and n=number of base pairs

Helical Breakdown- DENATURATION

The denaturation of a double helix is also cooperative, for much the same reasons as the formation is. In unzipping a helix, a bulge is formed due to input of energy. These bases are unable to H-bond with solvent and the solvent is order around the aromatic ring. In order to relieve this unfavorable situation, a nucleus of stacked single-stranded polynucleotide is formed which has favorable vader Waals interactions and has reduced hydrophobic surface in contact with the solvent. Then on the same ideas for unzipping apply as for zipping

Stability and Base Composition

In general, helical stability is linearly related to fractional G+C base pair content in DNA. As G+C increases so does stability. An empirical formula for calculating the melting temperature of a particular helix is given as

T_m (^oC) = 69.3 + 41 * fG/C

This expression quantifies the observed result that there is a linear relation between T_m and G+C content. This observation argues that the energetic contributions of the bases in the helix to its stability are independent and therefore additive. This finding implies that stabilization energies are sequence independent. That is the base pairs are all contributing equally and independently a constant amount of stacking energy, independent of the neighbors.

This line of argument is, however, an oversimplification. In complex DNA, melting occurrs in domains (FIGURE 18 & 19). This realization gives rise to a calculated stability matrix, for stacked paired dinucleotides in B-DNA configuration (TABLE 3). This stability matrix gives T_m values to the doublets under standard conditions. From these data, the melting of any DNA can be calculated. And the answers are surprisingly accurate (Table 4).

If the observed and calculated T_m's are plotted against the stacking energies we saw before, a linear correlation is observed (FIGURE 20). This indicates that stacking has a role in determining the T_m of DNA. Since stacking is a sequence dependent phenomena, then T_m is sequence dependent.

Helical Stability and Salt

It has been long observed that multiple stranded polynucleotide helices are stabilized by increasing monovalent cation concentration. In fact the T_m of a given DNA is linearly dependent on the log of the monovalent cation concentration.

We will not spend a lot of time on the polyelectrolyte behavior of nucleic acids, but instead we will simplify the treatments and take an empirical and thermodynamic approach. This approach can only be used to analyze the effects of salt concentration on nucleic acid stability, the polyelectrolyte approach is required to analyze the effects of salt protein-nucleic acid interactions.

The DNA phosphate backbone is negatively charged. In salt solutions, cations are associated with it. When DNA is denatured fewer total cations are associated with the separated strands than with the nucleic acid helix in its native state. This is because the charge density on double stranded DNA is higher than single strand nucleic acids. This creates a larger electrostatic potential, that more effectively attracts counterions.

Thus, in the denaturation reaction, the mass action equation can be written.

DNA_(Helix)^. M_rh Û DNA_(Coil)^.M_rc + M_(rh-rc)

rh= # of ions bound/base pair in a helix

rc= # of ions bound/base in a coil

rh-rc= net gain in free cations due to denaturation

Therefore, the denaturation reaction equilibrium can be shifted by adjusting the cation concentration.

We have already discussed that effect of temperature on helix®coil transition. The two effects can be balanced at particular conditions. That is if the salt is raised, increases helix potential, can increase temperature to denature.

We have only discussed here the effect of monovalent cations on structure. The effects of divalent cations are much more complex due to their multiple interactions with the DNA phosphate backbone-each M²⁺ can potentially bind one or two DNA phosphates and the binding is likely to be cooperative. Hence, the T_m dependence on divalent cation concentration is decidedly non-linear.

Patterns of base-base hydrogen bonds-Characteristics of the base pairs

Interactions between like and unlike bases have been observed in crystal strucutres individual nucleotides. There are 28 different base pairs which can be formed that have at least two H-bonds between them (FIGURE 21), and whose H-bonds are reinforced in strength by virtue of the fact that they are cyclic (Figure 22). These base pairs are symmetric at least at the level of the C1'-N bond. That is transformations about the central axis between the two base pairs results in the exact placement of one link on top of the other. In some of the cases delineated in the figure, an exact symmetry axis exists, the entire base can be placed on top of the other by rotating thru the dyad axis (SEE FOR EXAMPLE I-IV). Base pairs which have a dyad axis that is perpendicular to the plane of the bases most readily form parallel stranded double-helices, whereas base pairs whose dyad axis is parallel to the plane of the bases most readily for anti-parallel double helices (Figure 23).

Not many of these base pairs in Figure 21 are found in naturally occurring double stranded polynucleotides, but more of them exist than we would have expected at first from the Watson-Crick (W/C) DNA structure.

Let's examine the most common base pairs present in DNA and RNA, the Watson Crick base pair. This base pair stoichiometry of the W/C base pair satifies the most striking observation of the composition of DNA and RNA double helicies. That is A=T and G=C, indicating that these A+T and G+C occur in pairs. This observation is known as Chargoff's rule. The bonding pattern of the bases gives a partial symmetry, at least at the level of the C1' carbons, a feature that was key to discerning the bonding pattern in DNA.

The Hoogsteen base pairs also can satisfy the stoichiometry requirements of G=C and A=T, but failed to make the symmetry test. In no way can these be rotated to give any dyad symmetry, pseudo or otherwise.

Several other type of non-Watson-Crick base pairs exist, most notably, the wobble pairings (FIGURE 25) surmised from the complementarity of the tRNA with mRNA, and accounts for the observation that despite the fact that there are 64 possible codons and all are usable, there are not 64 different tRNA anticodon stem sequences. Thus, mispairings must occur. Example of these are shown. Note that all are missing a dyad axis, rotation about an axis can not be done to give C1'-C1' superposition.

These wobble base pairings do not only occur in tRNA-codon interactions but have been observed in crystal structures B-DNA containing mispairings.

How are double helices assembled??

Let us first examine the angular characteristics of base pairs. Figure 26 diagrams a base pair. The bases in a base pair are usually not coplanar, instead they are twisted about the hydrogen bonds that connect them, like the blades of a propeller. In fact, the dihedral angle that defines the non-coplanarity is called the propeller twist angle. To view the angle, hold the base pair such that you are looking down its long axis, the angle is defined as positive when the nearer base rotates clockwise.

If the base pair is imbedded in a helix, then there are several more angular attributes of the base pair that we must consider:

1) D-displacement from the helix axis. By virtue of the symmetry axis we discussed above, in a double stranded nucleic acid, their exists a helix axis which is defined by the average symmetry axes of the base pairs. In some cases, the base pair "slips" from this axis, and this displacment from the axis is measure as distance from the helix axis.

2) Base pair tilt. Despite the propellor twist of a base pair, an average mean plane of the base pair is defined (GRAY AREA IN FIGURE 26). The rotation of this plane about the pseudo-dyad axis defined above is the base pair tilt.

3) Base pair roll. This parameter measures the degree of departure of the mean plane of the base pairs from the perpendicular helix axis on the short axis of the base pairs.

4) Helix twist. Defines the orientation of a base pair with respect to the helix axis. That is how big an arc the base pair traces as it measured from one base pair to the next.

What do these geometric considerations mean to polynucleotide structure

Propeller twist, base pair roll and displacement are extremely important components in maintaining the stacking interaction of DNA. Figure 27 shows that in standard Watson Crick B-DNA, the bases are coplanar, no propellor twist and the bases stack readily upon each other. In this type of DNA, the helical twist is 36^o, meaning that there are 10.0 base pair/1 turn of helix. However, if any of these parameters are varied the stacking of the base pairs is changed dramatically. For example we lower the temperature or bind a protein. If this change in environment induces an overwinding of the DNA to give 9.33 base pairs per 1 turn of helix (helical twist of 38.6^o) and there were no change in either propellor twist, tilt and roll occurred, stacking would be disrupted or worse, the base pairs would crash into one another. These considerations are known as Calladine's rules.

More importantly, and as many things are in nature, more subtly, base pair tilt, roll, helical twist and propellor twist are SEQUENCE dependent, both from the influence of stacking interaction energies and the van der Waals constraints imposed by different base pairs.

For example, take the sequence G-C-G-C in the B-form. The propeller twist of these base pairs is close to zero. Moreover, the helical twist of this imaginary helix is close to the true B-form of 36^o. Lets examine the reasons for these properties. The G-C base pair has three hydrogen bonds holding it together. This property causes it to prefer low propeller twist values, presumably because keeping these H-bonds planar is energetically favorable. The value of twist is close to 36 degrees because a higher or lower value would cause the N2 amino group to clash into the guanine ring on the opposite strand above or below it.

The structure of A-T-A-T is also a B-form variant. The propeller twist of these base pairs is quite high ~20^o The helical twist of these base is also higher, more like the B-DNA of W and C, and possibly higher. The propeller twists of these bases is so high because these bases can form a 3-centered H-bond, giving a dinucleotide unit five H-bonds as opposed to four in normal DNA. The overwinding of this molecule occurs to relieve potential clashes that result from the high, positive propeller twist.

Table 5 shows the conformational variations in DNA, dependent both on composition and sequence of the DNA as well as the composition of the surroundings. For example, native DNA can be inter converted between two different families of DNA, B and A, just by changing the humidity. The compositional isomers serve to illustrate the sequence dependence; poly (dA-dT)₂ exists A, and B DNA, but poly (dA-A-T).dT-T-A) can exist only as a member of the B-family. These are only a few of the sequence dependent things that can occur in DNA, lets now examine A, B and Z-DNA structure in detail.

Information content of DNA.

As we discussed in the first lecture, the main function of a nucleic acid is the transfer of genetic information. In the case of DNA, this means not only inter-generational information or passing of the genetic blueprint as well as coding for the general body plan, but it also must contain regulatory information, to help the cell decide when to transcribe a particular gene and when to replicate. These "read-outs" of information are usually made by specific DNA binding proteins. These proteins recognize and bind to DNA sequences that are present in only one or a few copies per genome. Since the DNA of a simple organism like a bacteria contains millions of base pairs, how does the protein recognize a specific sequence--well it "reads" the pattern of H-bond donor and acceptor groups present on a sequence of DNA. The pattern of this information readout must be unambiguous.

The base pairs contain two different surfaces which, when they are contained in double helical DNA are displayed on opposite sides of the molecule. By convention, the sides are defined based on which side is facing where in B-DNA (Figure 29). MAJOR GROOVE-MINOR GROOVE. The information content of the major groove is unambiguous, the minor groove is ambiguous.

Comparison of A and B type DNA.

How different are A and B DNA. Grossly, these two types of DNA are quite different (SEE FIGURES OF 27 AND 28. A-DNA is a short stubby helix, while B-helices are rather thin. A-DNA is underwound with respect to B-DNA, having 11 residues/turn of the helix, while B-form has 10-10.5 (Table 5). The most obvious difference is evident if one looks down the helix axis of the DNA. The pseudo dyad axis of B-form base pairs lies almost exactly on the helix axis, therefore the end-on view of B-DNA shows the center of the "cylinder" of B-DNA filled with the base pairs, and the sugar phosphate backbone meanders around the outside of it. By contrast the end on view of A-DNA shows that the helix axis is "hollow". The polynucleotide chains wrap around the axis like a ribbon. Both the base pairs and the sugar-phosphate backbone are driven out towards the periphery of the double helix. Moreover, the groove sizes and widths of A and B DNA are dramatically different.

How are these difference generated by the microscopic parameters we discussed earlier?

The primary distinction that can be made between A and B type DNA is their differences in preferred sugar pucker and the degrees variability of the backbone torsion angles allowed in each type of DNA. A-DNA contains exclusively C3'-endo type sugar puckers, while B-DNA tolerates the C2'endo family of puckers which includes the lower right quandrant of the mutatrotation cycle, (FIGURE 8) from c3'exo to O4' endo. This difference in sugar puckerings causes a variation in the distance between the adjacent phosphates in the same polynucleotide chain; ranging from 5.9 A in C3'endo, to 7.0 A for C2'-endo configurations (FIGURE 30). The decrease P-P distance causes the helical rotation of A-DNA to be less than that of B-DNA. This difference has a consequence in determining the helical arrangements of B and A- DNA, such that these helices are MACROSCOPICALLY different.

The sugar puckers characteristic of each type of DNA cause the base pairs in the helices to assume different tilt angles with respect to the helix axis. In A-DNA, this tilt, defined as the angle formed between normals to the base pairs and the helix axis is positive in A-DNA, but negative in B-DNA. Looking from the minor groove sides of the bases, tilt in the clockwise direction is positive, and counterclockwise is negative. In B-DNA, the tilt parameter is variable, but is usually quite small, average of about -6^o. In A-DNA, tilt is much larger and less variable, +20^o. This sugar pucker determined angle has large consequences in terms of the base stacking overlap between adjacent bases in A and B DNA. In B-DNA, base stacking is largely limited to intrastrand interactions-interactions between bases in the same DNA chain. In A-DNA however, stacking involves both interactions between bases on the same strand and on the complimentary strand (SEE FIGURE 15). The reasons for this difference in stacking interactions is two-fold. First, the rotation in A-type DNA is less, ~31^o as opposed to 34-45^o in B-DNA. This favors interstrand overlap in A-DNA. Second the positive tilt in A-DNA also contributes to interstrand overlaps. Hence a change in the sugar puckers drives the base pairs to do fundamentally different things in their interactions with their brothers.

Outside of the sugar puckering modes and its affect on base pair tilt sense and base stacking, the most important feature distinguishing A from B-DNA is the dislocation of the base pairs (D), from the helix axis. As we have already described, the base pair in B-DNA is astride the helix axis, displacement of ~0.2A. In A-DNA, however, the helix axis is pushed far out into the major groove side of the base pairs with D amounting to 4.4-4.9A (FIGURE 31). This gives A-DNA its hollow appearence from the end-on. In A-DNA the displacement of the helix axis into the major groove gives rise to a very deep and narrow major groove and a reduced and relatively shallow minor groove. The shallow major groove of A-DNA is only accessible to small molecules, water and metals. In B-DNA the major groove is freely accessible. Since the unambiguous information of a double helix is present in the major groove face of the bases, this limits the utility of A-DNA in storage of regulatory information.

Z-DNA

The formation of Z-DNA is extremely composition and base sequence dependent. In general, only alternating co-polymers can form Z-DNA (G-C)₂, the compositional isomer, dG^.dC, does not form Z-DNA at all.

The formation of Z-DNA is solvent dependent. In oligonucleotides, it forms only in the presence of alcohol and either very high or very low salt. Moreover, as we shall see, the salt concentration affects the sugar puckering assumed.

The main characteristics of Z-DNA is that it is a left-handed double helix. Compare FIGURE 27 AND 31. It contains ~12-13 base pairs/turn of helix so it is a relatively loosely wound helix. It is called Z-DNA because of the zig-zag pattern of the phosphate back bone. In fact, this zig-zag forms the repeating unit of the Z-form DNA- UNLIKE A AND B-DNA, THE REPEATING UNIT OF Z-DNA IS A DINUCLEOTIDE. We will illustrate why that is in a moment.

This zig zag pattern of the backbone is a result of an alternating c angle configuration in the backbone and with the correlated differences in the sugar pucker. The pucker of the C, or pyrimidine is C2'endo, and the base is in the anti configuration. The G, or purine residue is in the syn configuration. The sugar attached to the G residue is in the C3'endo configuration (WHY??->equatorial conf of C1'-N and C4'-C5' bond.)

As a consequence of this alternating backbone conformation, the base stacking patterns in Z-DNA are quite bizzare (FIGURE 32). In the GpC step, relatively normal intrastrand base stacking interactions occur, while in the CpG step, there is a very extensive INTERSTRAND stacking of the C residues. The G residues are not stacked with the bases at all, but are instead stacked over the O4 of the furanose of the adjacent cytidines. This pattern repeats its way down the chain.

These things; the alternating sugar pucker and associated anti/syn conformation together and the alternating base stacking patterns serve to place the phosphates associated with the G and C bases in chemically non-equivalent environments. They in fact are different distances from the helix axis CpG=6.2 and GpC=7.6. Summing these factors together should convince you that the repeating unit of the Z-DNA is the dinucleotide.

The helical projections of Z-DNA are also unique. Z-DNA has only a minor groove; the G-C base pairs are not symmetrically related to the helix axis. Instead, they are shifted, with the C5 of the cytosine and the N7 and C8 atoms projecting into the major groove. This gives the Z-DNA helix its bizzare apprearance, the major "groove" is convex as opposed to the normal concave. Because of this shift, the minor groove is extremely deep and very narrow, so the information content of this DNA structure is limited.

As I stated before, solvent has a remarkable affect on Z-DNA structure at the microscopic level. In low salt, the N2 of guanine is hydrogen bonded via a H2O molecule to the PO4 on the 3' side of the base. In this configuration, the sugar is in the C3' endo configuration. In high salt, this water is replaced by a Cl- ion. The H-bond is no longer made and the negative Cl-, repels the PO4 and pushes the sugar into the C1'-exo conformation (Figure 33).

Polymorphism of B-DNA

The fine structure of B-form DNA is very polymorphic-assuming different values for the structural parameters we have discussed; sugar pucker varies, c varies, propeller twist can be large or small as can be base pair twist or roll. In B-DNA, these parameters and the variations in them are SEQUENCE DEPENDENT!!! This sequence dependence can be understood from the viewpoint that the base pairs wish to maximize their stacking interactions while avoiding steric clashes.

The purpose of propeller twisting is to maximize stacking interactions. The degree of propeller twist is somewhat correlated with helical twist. At values close to 36^o, propeller twists are near to zero because satisfactory stacking interaction can occur. Deviations from this ideal result in decreased stacking compensated for by propeller twists. Propeller twists, however, can cause steric clashes in alternating py-pu sequences. These clashes occur between G O6 and A N6 in the major groove for pu3'-py5' sequences and in the minor groove between G N3 and N2 or A N3 atoms for py3'-py5' sequences. The potential for overlap is about twice as severe in the minor groove as compared with that in the major groove. All these clashes can be relieved by the use of one or all of the following strategies:(FIGURE 34).

These escape strategies are known as Calladine's rules- derived from mechanical engineering principles.

The DoDecamer-An example.

Point to ponder- are these variations required for pattern recognition-these all change the spatial relation of one H-bond donor to the next.

Kinematics of DNA-"Ramachandran Plots" of DNA.

Just like the Ramachandran plots for amino acids which show that particular types of secondary structure have particular values of f and y angles. Particular pattern variations in sugar torsion angle d with c are characteristic of a type of DNA. In a double helix, consider first a C1' endo configuration. Viewed down the C4'-C3' bond, there is an 82^o d torsion angle between the C5' and O3' and the c angle is 146^o. As one opens-up the d , by "pushing-up" (anthropomorphically speaking) on O3', this move the C2' carbon leftward, closing down c (See Figures 33-36).

A-DNA d never varies so c is set

Z-DNA two ranges, one for each base C and G

CONFORMATIONAL STATISTICS

The packaging of 3 meters of DNA in a eukaryotic cell obviously involves the folding of the molecule back on itself. How the polynucleotide bends is thus important to placement in the cell. Moreover, many regulatory functions involve bringing together protein binding sites that are well separated from each other-a process which is thought to require DNA bending. Because of what we have already learned about the sequence specific factors that influence DNA conformation of DNA, we would also like to know something about how base sequence may affect the folding of DNA upon itself.

Before we get into the effect of base sequence on DNA deformability, we have to have an understanding of how we would begin to measure flexibility and how we would characterize the movements of a DNA chain.

Native DNAs from organisms have very high molecular weights-the MW of the E coli chromosome is ~3 X10⁹. As we know, DNA is made of segments, in the simplest case these segments can be thought of as the base paired nucleotide. Since these segments have finite size, it is clear that many conformations of the chain in solution will be excluded, because of the prohibition of physical overlap of two segments in the same volume of space. Outside of these rather obvious considerations, nucleic acid polymers can be modeled as existing between two limiting case of chain deformability-the completely rigid rod and the completely flexible random coil. In fact DNA can be modeled in either of these two configurations, depending on the number of segments that one includes in the analysis of DNA deformation. As we shall see, short stretches of DNA can behave as rigid rods while enormous DNA lengths can assume the characteristics of completely flexible chains. In actual fact, DNA of intermediate number of segments (length) is best modeled between these two limiting cases, as the so-called worm-like coil, which is of intermediate flexibility.

Average dimensions.

A solution of flexible polymers contains molecules that may have a different conformation at any given instant. Further, due to the random bombardment of the polymer molecules by solvent (Brownian motion), the conformation of a particular polymer molecule will change with time. This suggests that we cannot specify a single dimensional parameter to characterize flexible polymers in solution, but must instead deal with average dimensions. One of the most commonly used average dimensions is that of the mean square end-to-end distance <L²>. We have a polymer of N repeating units each connected by bond vectors of length b. A vector L is drawn between the beginning and end of the chain, which equals the sum of the individual bond vectors b_i. Therefore

_N

ⁱ⁼¹

Thus, _{N N}

^{i=1 j=1}

This all means that the root mean square distance between point i and point j is the sum of the vectors that lie between them.

Now that we have defined our measurements somewhat, we shall proceed to apply these considerations to particular models of polymer chains.

Rigid Rod.

This is relatively easy to understand, since the rod is inflexible, no averaging over internal conformations is required. Thus, the distance between any two points is equal to

L_ij= b*i-j

where all bonds have constant length b and the distance between two points at the ends of the chain is

L=b*N

where N is the number of segments and b is the length of the bond that connects the segments.

Flexible Linear Chains

At the opposite extreme from the rigid rod, let us imagine a completely flexible chain containing N segments each separated by bonds of length b. We can imagine that the bonds are connected by universal joints that allow completely free rotation. The conformation of the polymer is therefore that of a random walk in which successive steps are completely uncorrelated in direction. The quantity which must be ascertained for the polymer is its distribution function; that is the probability W(L,N)dL that after N steps the end of the chain will be at distance L and L+dL from the origin.

Chains that conform to this distribution function are often called gaussian chains because of the gaussian character of the population of chain conformations. Substituting the distribution function into this integral gives the mean end to end distance

<L²>=ô _{o ì} L2 W(L,N)dL

<L²>=Nb²

Thus the mean square end-to-end distance of a completely flexible chain is proportional to the first power of N as opposed to the N² for the rigid rod.

The concept of a freely jointed chain is obviously an inaccurate representation of a natural polymer, in this case a nucleic acid chain because the joints connecting the segments are not universal joints because of steric considerations and there are fixed bond angles (e.g. geometry imposed by tetrahedral carbons). Thus, both N and b for a real polymer are modified to effective N and b. The values for N_e are less than actual N and b_e smaller than b. The distribution function for the effective chain is just that for the freely jointed chain with N replaced by N_e and b with b_e. The mean square end to end distance of this chain is thus,

<L²> = b_e² N_e

Worm-like chains.

Even with the concept of statistically equivalent chains, the gaussian distribution requirements are such that N_e must be substantially greater than 1, although N_e << N. For sufficiently short chains or relatively stiff chains, this requirement may not be satisfied. In fact native DNA is stiff enough that gaussian statistics are inapplicable for many purposes. This situation thus demands an even more elaborate treatment. Since stiff chains can be envisioned to bend only gradually and smoothly, somewhat like a worm, hence the term wormlike chain

In order to characterize the stiffness of a polymer chain, we imagine that the polymer is laying along the y-axis of a 3-D coordinate reference frame. The chain consists of N segments, connected by bonds with length b. We imagine that the first step is taken along the positive z axis. We then ask, what is the average projection <z> of the chain at N steps along chain at the z-axis. It is obvious that if the rod were completely stiff, <z> would have its maximum value of N*b; for a freely jointed chain, <z>=b since setps beyond the first the first would have equal probability to be taken in the (+) or (-) z direction. Thus, polymers of intermediate stiffness will have intermediate values of <z>. For a chain with constant bond angle q , the average projection must consider the projections of all succeding bonds on the first, thus,

<z>= b+ b<cos q _1,2> + b <cos q _1,3> + ... b<cos q _1,N>

for the chain with free rotation

<z> = b* 1-cos^N q / 1- cos q

if the cos q is close to zero (i.e. angle of 90^o; cos 90=0) <z> = b/(1-cos q ), however if q is 0^o, true for very stiff chains, cos q = 1 then <z> becomes very large. There is, therefore a continuum, in which both segment length b and the bond angle q are taken to the infinitesmal limit in such a way that the quantity

a = b/1- cos q

remains finite. The quantity a is called the persistence length of the worm like chain and represents the average extension along the z-axis of an indefinite length polymer. In other words the persistence length is that number of segments N where the chain behaves as a rigid rod. In fact for a wormlike chain, the effective segment length b_e is twice its persistence length. Thus for a short wormlike chain, the limiting behavior is a rigid rod.

Thus for considering DNA, the persistence length in units of distance is a measure of its stiffness. The longer the persistence length, the more stiff the chain.

With all of this stuff in mind let's take a look at how to apply it to real DNA

In the Shore et al, they wished to determine experimentally the degree of DNA flexibility and whether the model of a worm like coil held true for DNA of all lengths or were other considerations needed. To do this, they measure the rate of circularizing several linear pieces of DNA of different lengths and compare the results to that predicted from the model. The ring closure probability can be understood as the effective concentration of one end in the vicinity of the other. That is if they are sooooo far away the concentration of one in near the other is low and closure probability decreases, as they are brought together, the effective concentration increases, as does the closure rate. In all cases they compare the rate of closure of the linear DNA into a circle, and normalize this to the probability of joining two ends on separate pieces of DNA (j-factor). The relative increase is the benefit from being on the same piece of DNA.

In fact the effective concentration of ends (which is proportional to the average mean square distance <L²> and thus to the number of base pairs) is not the only thing that limits joining of the two ends. If the DNA is particularly short, the relative orientation of the two ends are correlated. For joining to occur, the number of bases has to be close to 10, i.e., the number of bases/10.5 should be an integer.

The data shown in FIGURE 38 is in fact in quite good agreement with the worm like coil representation of DNA >500 base pairs and end to end correlations do not have an effect until below this value. The fact they do indicates that the twisting motions of DNA are energetically restricted. This leaves open the possibility of using this technique to measure the energetics of twisting.

The linear double stranded DNAs that we have discussed so far in this class exists in a topologically relaxed state. The active DNA inside a cell is not relaxed DNA, but instead is supercoiled. This topological state of supercoiling is the next highest order level of DNA organization after the linear relaxed state.

In order for a DNA molecule to be supercoiled, its must be in a closed loop, with the ends fixed. Closure is most often achieved by joining the ends of a each DNA strand to form a circle. This, however is not the only way to close a loop; in a eukaryotic chromosome, for example loops are formed by holding the ends on a protein scaffold or by simply restricting the freedom of rotation of the ends of a long DNA molecule.

What is super coiling.

Picture a 260 bp piece of DNA duplex in the B-form. The number of helical turns in this linear DNA is 25 (260/10.4). Now we shall join the ends of this helix. This circular DNA is said to be relaxed (FIGURE 39). Suppose now instead, we unwind the duplex by two turns before we join the ends. The resulting DNA can now fold into one of two different structures; one that contains 23 turns of B-helix and an unwound loop or alternatively, it can adopt a structure with 25 turns of B-helix and two turns of superhelix. The unwinding is taken up by allowing the previously unwound region to adopt the B-form and twisting the circle into a superhelical form. This is the supercoiled form. This structure is energetically favored over the one containing the unwound loop (WHY??). Notice that the superhelical form produced from unwinding the DNA is right-handed. This unwind produced, right-handed superhelix is called negatively supercoiled. On the other hand, superhelices formed by overwinding are left-handed and are said to be positively supercoiled. MOST DNA IN ORGANISMS IS FOUND TO BE NEGATIVELY SUPERCOILED.

Supercoiling play an important role on many life processes. For example, both the transcription of genes and replication of DNA require it to be unwound. Thus, negatively supercoiled molecules (right handed superhelices that result from joining underwound strands) are poised for these processes. In fact several antibiotics have been developed that kill by inhibiting the enzymes that modulate the degree of supercoiling in infectious organisms. So an understanding of this process must occur before we blah, blah, blah. We will briefly discuss these enzymes later.

DNA Topology

In order to better understand the relationship between the degree of supercoiling, and twist, we turn again to the mathematicians. An equation has been worked out to describe supercoiled DNA ribbons. The key topological property of a circular DNA is its LINKING NUMBER L_k . THis quantity is defined as the number of times one strand of DNA winds around the other in the right-handed direction (since we are taking as our reference B-DNA). For the relaxed DNA shown in FIGURE 39), the linking number is 25, for the one in C, the linking number is 23. The circles corresponding to the joining the ends of these two molecules is 25 and 23 respectively. NOTICE, THE SIZE OF THE DNA DID NOT CHANGE, HOWEVER THE LINKING NUMBER DID-CIRCLES OF THE SAME SIZE CAN HAVE DIFFERENT L_k. Molecules which are identical in length, and differ only in linking number are topological isomers or topoisomers. Moreover, since the ends of a circle are fixed-THE LINKING NUMBER OF A CIRCULAR DNA CANNOT BE CHANGED-IT IS A TOPOLOGICAL QUANTITY-A PROPERTY OF THE CIRCULAR STATE. Since we are dealing with closed circles, the only way to form a circle is to match up the ends. THEREFORE, THE LINKING NUMBER OF A CIRCLE MUST BE AN INTEGER. Moreover, the only way to change the linking number of a circle (convert one topoisomer into another) is to CUT the strands. The linking number of a topoisomer is constant.

In a relaxed unstrained circle shown in FIGURE 39b, the linking number equals the twist (T_w). The twist is defined as the number of 360 degree turns the ribbon makes as it goes around the circle. The twist of a given topoisomer can differ, and can be non-integral, however, the linking number as we have said is constant for a particular topoisomer and, moreover IT MUST BE AN INTERGER. Twist is a metrical property and for DNA is limited within narrow bounds because B-form DNA does not stray very far from having 10.4 bp/turn. Thus for DNA, Tw is usually very close the the # of bp/10.4. Since topoisomers can have linking #'s that vary at least by two integers, as we have seen, these linking differences can be taken up by the unwinding DNA or, superhelix formation. Because of the structural demands of DNA, the various values of Lk a circle can have results not in changes in twist, but formation of superhelix. This super helix formation is decribed by the term writhe (W_r). These three terms are obviously physically related as we have described, and are related mathematically by the equation:

L_k = T_w + W_r

As we mentioned before, L must be and integer once the DNA is closed and thus Tw and Wr can adopt any value, dependent upon the three dimensional shape of the DNA. For DNA, right handed twisting is positive as is righthanded rotations for Lk, but for right handed superhelices, the sign of Wr is negative.

Let's demonstrate how the concept of writhing (or superhelix formation) prevents the changes in DNA twist from causing alterations in the structure of the double helix. If circular DNA is confined to a planar arrangement, the changes in linking #

DL_k = DT_w + DW_r

can only be compensated for by equivalent changes in twist, because W_r is held at zero. In this case, DT_w can be compensated for by smoothly modifying all helical turns, or since B-DNA likes to remain that way, by disrupting a few base pairs. If, on the other hand T_w is unchanged, then DT_w is zero and thus

DL_k = DW_r

that is the change in linking number is transformed almost entirely into writhe. Thus in the example given in FIGURE 39c & d, the DNA is twisted in a right handed sense (UNWINDING), i.e. REDUCTION of L_k, and therefore the change in DL_k is compensated by right-handed motion superhelix formation, or negative writhe.

As we said before, most DNA in vivo is negatively supercoiled. The degree to which a molecule is supercoiled can be expressed as the specific linking difference or superhelical density. This quantity gives a length independent measure of the number of supercoils per 10 base pairs. The formula:

s = (L - L_o)/L_o

L_o is the linking number of the relaxed circular molecule and L linking number of the supercoiled molecule. For the DNA that is in the Figure, L_o is 25, L=23, thus s= -.08. The superhelix density of DNA in cells is between -.05 to -.09. That means, the DNA is negatively supercoiled (i.e. right handed superhelices resulting from Unwinding).

Effect of DNA structure on Supercoiling.

If in B-DNA one right handed turn is converted to Z-DNA, the T_w changes by two. Why, because Z-DNA is lefthanded in twist, therefore, from 1->0->-1 is 2 twists. Since in a circularly closed DNA molecule L_k remains UNCHANGED, the writhing number has to change by 2. This means that the B_Z transition of only small stretches of DNA can have a dramatic influence on the macroscopic topology of supercoiled DNA.

Supercoiling and Hydrodynamics

The effect of circularizing a molecule effectively halves the effective mean square radius. Thus, the molecule appears smaller by a number of experimental techniques, among them gel electrophoresis and density gradient centrifugation. Supercoiling a molecule further compacts its shape. This compaction is another reason why the DNA is a cell is supercoiled-it compacts it.

With relation to tRNA structure, to be discussed later, we will now cover the phenomena of intercalation. Intercalation means the insertion of a planar molecule (e.g. other bases, drugs) base between two others. This requires the spacing between the two bases to increase. Several dye molecules and antibiotic drugs insert themselves between base pairs. These intercalators are planar, aromatic molecules that bind to the DNA by essentially becoming part of the base pair stack (FIGURE 40). Intercalative binding is, therefore, characterized by high binding affinities.

Intercalation changes the physical properties of the double helix. As the intercalator slides between the bases, their normal stacking interactions are disrupted as they now stack with the drug. This separation of the base pairs forces a severe distortion of the regular helical structure of the sugar phosphate backbone. Since the intercalation behaves nearly as an extra base pair, as the more intercalator is added, the greater is the apparent length of the molecule. The separation of the base pairs causes them to UNWIND, in order to accommodate the insertion.

If we start out with negatively supercoiled DNA and add intercalator, we further unwind the helix. In the case of ethidium bromide, the degree of unwinding is -26^o/molecule bound. Since we are not changing the linking number of the plasmid, the changes in T_w are solely compensated for by an increase in writhe. Recall that negatively s.c. DNA has negative writhe so that an increase in W_r indicates that W_r-->0 as more intercalator is added. As the W_r-->0, the DNA loses its compact shape and thus appears "bigger" in gel or centrifugation experiments (FIGURE 41). At even higher concentrations of intercalator, the T_w<L_k and thus, the W_r becomes positive, with the opposite screw sense (Positive supercoils [+W_r], left handed screw. This exercise illustrates the interrelation of these three quantities and the constraints imposed on circle-i.e., the L_k never changes.

For a change and much to your relief, the intercalation process does not cause a simple change in the backbone configuration. The sugars of B-DNA appear to remain in the C2'endo family. The major changes in the angles of the backbone occur at the phosphodiester linkages and at the X-angle which now cause the bases to become strictly perpendicular to the sugar.

The enzymology of supercoiling

Inside of all cells there is a class of enzymes first discovered by Gellert and Wang in the early 1970's. These enzymes called topoisomerases regulate the supercoiling of DNA in a cell. They are called this because they can isomerize, one topoisomer (RECALL DEFINITION OF TOPOISOMER-DIFFERENT L_k) into another. There are at least two kinds of topoisomerases, Topo I and Topo II.

Topo I type topoisomerases relieve supercoiling, i.e., they relax supercoiled DNA. These enzymes do so without the input of energy. They act by cutting one strand of the supercoiled double helix, rotating one strand about the other and then reseal the strands. This reaction proceeds fueled only by the energy inherent in supercoiled DNA. More about the inherent energy later, but there is an interesting enzyme mechanisms question here, and that is how is the energy of the cleaved phosphodiester bond saved. Well it turns out that the cleaved strand of the DNA forms a covalent bond to a tyrosine-the same type of high energy linkage found in the phosphodiester backbone of DNA.

Topo II type topoisomerases function to put supercoils into DNA. The enzyme gyrase required for DNA replication is a type II topoisomerase. The introduction of supercoils into DNA requires the hydrolysis of ATP. Topo II enzymes function by making double stranded breaks in the DNA and rotating one strand past the other and rejoining them. SInce supercoiling is an active process, requiring energy in put, this indicates that supercoiled DNA is a high energy form of DNA. In fact the ---- increases with the square of the superhelix density.

Speculations on the role of Supercoiling in action at a distance.....

RNA structure

All double helical RNA molecules are members of the A-family of polynucleotide double helices. RNA molecules are never found in the B or Z configuration-WHY?

The answer to this question believe it or not is still a subject of some dispute. Three reasons for the conformational uniformity of RNA have been proposed.

The two alternative models both suggest that the 2'-OH stabilizes the C3'-endo configuration as opposed forcing the sugar into this configuration

Thus, this important observation remains unexplained and is likely to result from a combination of two or more of the above explanations.

The Structure of tRNA.

The role of tRNA is two fold, bind to and accept the correct amino acid from the appropriate tRNA synthetase and recognize and bind the ribosome-mRNA complex to deliver the amino acid to the growing polypeptide chain. Because of its central role in gene expression (and its relative abundance), tRNA is one of the best, most thoroughly studied biological macromolecules.

tRNAs are generally 75-90 nucleotides in length and contain not only the four normally occurring RNA bases, but also rare or minor variant bases. Some of these bases are given in the (Figure 44). Most characteristic are the products of alkylating adenine guanine and cytosine amino groups; alkylations of the ring N of these bases, glycosylation and saturation of C=C bonds. As we shall see these modifications often restrict the type of base pairing these bases can assume and has interesting consequences for the tRNA tertiary structure.

Secondary structure

Based on base-base complimentarity, the secondary structure of tRNA can be drawn as the familiar cloverleaf structure of 4 stems, each consisting of four to seven Watson-Crick type base pairs. Five regions of the tRNA are not base paired,. the CCA acceptor stem, the D-loop (named for the presence of dihydrouridine), the anti-codon loop which interacts with the mRNA, the "extra arm" and the TYC loop (named for the presence of the pseudouridine base. (FIGURE 45).

Tertiary Structure.

The three dimensional structure of the tRNA molecule is not a cloverleaf, the molecule instead folds back on itself, forming two segments of double helix and most of the upaired bases in the T and D loops form base pairs with each other (FIGURE 46). One of the two segments of double helix are formed by the stacking of bp region of the T and acceptor stems, the other by the stacking of the D-arm base paired region on the base paired part of the anticodon loop. This tertiary structure of the molecule is L-shaped and is held together by base pairs betwen the bases in the upaired regions of the stems and by unique stacking interactions (Figure 47).

Base modifications and base pairings (Figure 48) 

m¹A₅₈-T₅₄ This base pair is a reversed Hoogstein type, with the chains from each base pair running in an antiparallel direction. The m¹A is unable to form a normal WatsonCrick base pair because the N1, normally a W-C acceptor is methylated, rendering it refractory to H-bonding.

m²₂G₂₆A₄₄ is a purine-purine base pair who normally pair in a wobble fashion. However, this configuration is altered due to the presence of the bulky methyl groups on G N2. This pushes this base pair further apart, more than even in the usual wobble situation. This base pair is at the junction of the anticodon and D-stems, and is stacked with both and is thus responsible for the 26^o kink between these two helices

A number of triplet interactions are formed which also help hold the structure together (Figure 48).

Primary sequence conservatism gives all tRNAs the same structure. An examination of the places in tRNA where the base are conserved turns out ot be in the non-base paired loops. As we have juest seen, these are the bases which are involved in stabilizing the tertiary structure of the tRNA molecules (Figure 47 & 48).

Base stacking

The strength and utiliity of base stacking to hold together the structure of a polynucleotide is illustrated by tRNA. Only 42 of the 76 bases in the yeast tRNA^phe are involved in A-RNA double helical structures. Yet, however, 71 of the 76 are involved in stacking interactions. Of the five that are not, one is the terminal A in the acceptor stem and the two others are the modified dihydrouridines in the D-stem, which are non-planar, non-aromatic and hence oppose stacking. The base stacking in the double helical regions resembles that we have seen in A-DNA. A very interesting stacking interaction is seen where the individual bases of one strand are tucked between the bases of an adjacent strand. These interactions occur where three strands meet; inside the L corner and in the T-loop region.

FIGURE 49 shows the case of A₉ intercalated between the bases G₄₅ and m⁷G₄₆. Because the base inserts between the two others, and the base has finite vander Waals distance (this is a stacking interaction after all), the receiving bases must move apart by at least 3.4A. For this to occur, the ribose phosphate backbone must readjust accordingly. The most obvious way for the chain to readjust is by changing its sugar pucker, such that the bases can separate-a change from the A-RNA C3'endo to the more B-DNA-like C2'endo. As you should recall this increases the intrastrand phosphate-phosphate distance from 5.9 to 7.0 (just like in B-DNA).