Ribozymes: structure and mechanism
`in RNA catalysis
`
`William G. Scott and Aaron Klug
`The hammerhead
`RNA is a small catalytic RNA found
`in a number of RNA virus
`genomes
`and virus-like RNAs. The
`recently
`determined
`crystal
`structures
`of
`hammerhead
`ribozymes
`reveal how a small RNA motif can
`fold up into a con-
`formation
`suitable
`for mediating RNA cleavage.
`
`has
`in enzymology
`A NEW FIELD
`the
`emerged
`in the past decade with
`discovery
`that RNA can act as an en-
`zyme. First discovered
`in the cellular
`RNA-splicing and processing machinery
`in the form of self-splicing
`group 1 in-
`trons’
`and precursor
`tRNA-processing
`RNase P (Ref. 2) RNA catalytic
`activity
`in a number of smaller RNAs has subse-
`quently
`been
`identified3-5. The small,
`naturally
`occurring
`catalytic RNAs are
`generally
`found
`in the genomes of RNA
`viruses and in virus-related RNAs, which
`are believed
`to replicate
`by a rolling-
`circle mechanism.
`of one
`structure
`Recently,
`the crystal
`of these small catalytic RNAs, the ham-
`merhead
`ribozyme, was elucidated
`by
`two research groups using different ap-
`proaches. The first structure was that of
`the hammerhead
`ribozyme,
`in which
`the catalytic
`or ‘enzyme’ strand was
`composed
`of RNA and
`the RNA sub-
`strate was replaced with a ‘substrate-
`analogue’ strand composed of DNA. The
`
`W. G. Scott and A. Klug are at the MRC
`of Molecular
`Biology,
`Hills Road,
`Laboratory
`Cambridge,
`UK CB2
`2QH.
`
`DNA strand was employed as a competi-
`tive inhibitor
`to prevent catalytic
`cleav-
`age6. The second hammerhead
`ribozyme
`structure
`was
`composed
`entirely
`of
`RNA with a single 2’-methoxyl modifi-
`cation
`at
`the active
`site
`to prevent
`Despite
`superficial
`differ-
`cleavage:.
`ences,
`the
`largely conserved
`catalytic
`core region of both ribozyme
`structures
`is quite similar. The hammerhead
`ribo-
`zyme crystal
`structure,
`in conjunction
`with numerous
`experimental
`biochemi-
`cal results,
`aids our understanding
`of
`RNA catalysis
`and
`its relation
`to RNA
`three-dimensional
`structure.
`Indeed,
`it
`has allowed us to propose
`a testable
`mechanism
`for RNA catalytic
`cleavage
`in the hammerhead
`ribozyme.
`
`RNA enzymes and catalytic mechanisms
`li-
`The group 1 intron
`catalyses
`the
`gation of adjoining exons, using the 3’-OH
`of a guanine
`co-factor as a nucleophile
`to mediate
`trans-esterification,
`and
`RNase P
`(an RNA-protein
`complex
`whose RNA subunit possesses
`the cata-
`lytic
`activity
`of
`the enzyme)
`simply
`hydrolyses
`the phosphodiester
`back-
`bone of precursor
`tRNA. Other RNAs,
`
`Table I. Some characteristics
`
`of naturally occurring catalytic RNAs
`
`Ribozyme
`
`species
`
`Nucleophile
`
`Reactron
`
`products
`
`Group I mtron
`
`Y-OH of guanosine
`
`RNase P
`Group
`II intron
`
`H,D
`2’.OH of adenosine
`
`Hammerhead
`ribozyme
`Harrpin nbozyme
`
`delta
`Hepatitis
`virus
`ribozyme
`tRNAPhe
`
`220
`
`divalent metal hydroxide,
`e.g. fMg(H,o),(oH)I+
`divalent metal hydroxide,
`e.g. fMg(H,o),(oH)l+
`divalent metal hydroxide,
`e.g. fMg(H,o),(oH)I+
`divalent
`lead hydroxrde,
`e.g. fPblH,o)~IoH)I+
`
`and
`
`exons and
`
`intron with 5’guanosine
`
`5’ to 3’jorned
`Y-OH
`and 3’-OH
`5’.phosphate
`jorned
`lariat
`intron wrth 2’-3’
`exons and
`5’ to 3’joined
`at A and 3’.OH tail. Also acts as a DNA endonuclease
`when bound
`to a protein
`5’.OH and 2’,3’ cyclic phosphatase
`
`5’.OH and 2’.3’ cyclic phosphatase
`
`5’.OH and 2’,3’ cyclic phosphatase
`
`5’.OH and 2’.3’ cyclic phosphatase
`
`C 1996, Elsevier Science Ltd
`
`TIBS 21-JUNE1996
`
`such as the group 11 intron’, as well as
`several
`smaller
`self-cleaving RNAs, de-
`rived
`from RNA viruses
`and virus-like
`RNAs”-“, also possess
`catalytic
`activity.
`The group 11 intron
`ligates
`adjoining
`exons using
`the 2’-OH of an adenosine
`within
`the intron
`to mediate h-ans-esteri-
`fication,
`and
`in the process
`generates
`a ‘lariat’ product. Although
`the small,
`self-cleaving RNAs have very different
`conserved
`catalytic
`core
`sequences
`(and
`presumably,
`distinctly
`different
`three-dimensional
`structures),
`they have
`in common
`a metal-hydroxide-mediated
`trans-esterification
`that generates
`prod-
`ucts having 5’-OH and 2’.3’-cyclic phos-
`phate
`termini.
`Interestingly,
`tRNAPh’,
`both
`in solution8
`and
`in
`the crystal
`form”,“‘, has been observed
`to cleave
`catalytically
`and highly
`specifically
`in
`the presence
`of Pb”,
`yielding
`these
`same RNA-strand cleavage products.
`The naturally
`occurring
`self-cleaving
`RNAs including
`tRNAPh’ are single RNA
`molecules,
`but can be made
`into
`true
`enzymes
`exhibiting multiple
`substrate
`turnover
`simply by division
`into
`two
`strands of RNA. The nucleophiles
`impli-
`cated
`in
`the mechanisms
`of several
`catalytic RNAs are listed
`in Table 1, to-
`gether with the reaction products, which
`are characteristic
`of each
`type of ribo-
`zyme. A number of artificial
`ribozymes
`have now been produced
`by means of
`in vitro RNA selection methods”. These
`ribozymes
`are believed
`to employ other
`types of catalytic mechanisms.
`
`The hammerhead RNA is small and well
`characterized
`Because
`it is small and has a simple
`cleavage mechanism,
`the hammerhead
`ribozyme
`is perhaps
`the best experi-
`mentally
`characterized
`RNA enzyme,
`and
`therefore,
`is a clear candidate
`for
`structural
`studies. Owing
`to the dedi-
`cated efforts of a number of biochemists,
`a wealth of information
`regarding
`the
`conserved
`base
`requirements
`in
`the
`catalytic
`core of the hammerhead RNA,
`as well as the chemical
`nature of the
`divalent-metal-catalysed
`strand-cleavage
`reaction, has been made availablej,‘“,‘3.
`The hammerhead motif consists
`of
`three base-paired
`stems
`flanking a cen-
`tral core of 15 conserved
`nucleotides.
`(Fig. 1). The conserved
`central
`bases
`are essential
`for ribozyme activity. Most
`of these conserved
`bases cannot
`form
`conventional Watson-Crick
`base pairs,
`but
`instead
`form more complex
`struc-
`tures, which mediate RNA folding and
`catalysis.
`Substitution
`of any of
`the
`conserved
`bases with other naturally
`
`1
`
`MTX1043
`
`

`

`TIBS 21-
`
`JUNE 1996
`
`occurring bases14, or sometimes even
`artificial alteration of their functional
`groups5, results in diminished catalytic
`activity. In addition,
`two sets of base
`pairs in stem III and one pair in stem II
`are conserved; changing these to other
`base pairs either impairs or abolishes
`catalytic function. The crystal structure
`of the hammerhead
`ribozyme provides
`rationalizations
`for several sets of pre-
`vious experimental observations’, al-
`though a few other sets of results are at
`odds with the crystal structures
`(and
`sometimes contradict one another)5.
`The hammerhead RNA, like all other
`naturally occurring ribozymes, is a met-
`aIloenzyme15
`and requires a divalent
`metal ion, such as Mg*‘, to mediate cata-
`lytic cleavage. As with Pb*+-tRNAPhe, the
`divalent metal ion is thought to be hy-
`drated, and becomes active when it binds
`to the RNA and ionizes, i.e. the active
`form is an RNA-bound metal hydroxide
`that acts by abstracting a proton from
`the 2’-OH at the cleavage site. The rate
`of divalent-metal-ion-assisted
`catalytic
`cleavage generally
`increases with de
`creasing pKa of the metal hydroxide (see
`Table II), strongly suggesting the active
`species is indeed a metal hydroxide1*J3.
`However, exceptions
`to this rule (such
`as Pb2+, which
`specifically
`cleaves
`tRNAPhe but not the hammerhead RNA)
`do exist, indicating that other factors,
`such as ionic radius and ‘hardness’ of
`the metal ion might also play a role in
`determining catalytic activity*.
`Finally, replacement of the pro-R (but
`not the pro-S) phosphate oxygen (see
`Fig. 2) at the active site with a sulphur
`reduces hammerhead catalytic activity
`in the presence of Mg2’; this activity can
`be rescued partially by the addition of
`softer (hence more thiophilic) divalent
`metal ions such as Mn2+ and Cd2+ (Ref.
`18). This latter result indicates that Mg2+
`(a relatively hard Lewis acid) binds di-
`rectly to the pro-R oxygen at the cleav-
`age site (see also Fig. 2b).
`
`Crystal structure of the hammerhead
`ribozyme
`Despite differences in nucleotide com-
`position, phosphate backbone connect-
`ivity, crystallization
`conditions
`and
`crystal-packing
`interactions,
`the three-
`dimensional structure of the catalytic
`
`ions, such as Mg2+, are
`‘hard’ metal
`*So-called
`Lewis acrds that
`interact with ‘hard’ Lewis bases
`such as phosphate oxygens and H,O
`in preference
`to ‘softer’ Lewis bases such as the exocyclic func-
`tional groups of nucleotide bases. ‘Hard’ interactions
`are predominantly
`electrostatic, whereas
`‘soft’
`in-
`teractions are dominated by orbital interactionsi6,17.
`
`‘..,:A
`
`REVIEWS
`
`ribo-
`core of the all-RNA hammerhead
`zymer is almost identical to that of the
`hammerhead ribozyme in complex with
`a DNA substrate
`inhibitor6, suggesting
`that both structures
`represent
`the cor-
`rect fold for an active hammerhead ribo-
`zyme in solution. Chemical crosslinking
`experiments
`also demonstrate
`that a
`hammerhead ribozyme, restrained to the
`crystal structure
`fold, has unaltered
`cleavage activitylg, further suggesting
`that the crystal structure
`indeed repre-
`sents the correct fold of the hammer-
`head ribozyme.
`The global conformation of the all-
`RNA hammerhead ribozyme is depicted
`in Fig. 3a as a roughly y-shaped fold.
`Stem II and stem III are approximately
`co-axial, with stem 1 and the catalytic
`pocket branching away from this axis.
`Stem 11, augmented by two GA, reversed-
`Hoogsteen base pairs and an unusual AU
`base pair, stacks directly upon stem 111,
`forming one pseudo-continuous
`helix.
`The helix is not actually continuous, be-
`cause
`it incorporates
`a three-strand
`junction where the active site cytosine
`(C-17) is squeezed out of the helix and
`forced into the four-nucleotide catalytic
`pocket, which is formed by a sharp turn
`in the hammerhead
`enzyme
`strand.
`This turn is identical in sequence and
`structure
`to the uridine turn found in
`the anticodon
`loop of tRNAPhe (Ref. 6).
`The phosphate backbone strands, which
`diverge at the
`three-strand
`junction,
`subsequently
`reunite
`to form stem 1.
`These structural features are illustrated
`schematically in Fig. 3b, which is colour-
`coded to complement Fig. 3a.
`
`Structural details and a proposed
`mechanism for RNA catalysis
`The uridine turn in the hammerhead-
`RNA smoothly connects stem I to the
`augmented
`stem II helix by bending
`the enzyme strand of the ribozyme
`molecule, forming a highly structured
`pocket into which the cleavage base is
`positioned
`suggestively. The
`tRNAPhe
`uridine turn binds divalent metal ions
`
`Stem
`
`III
`
`Scissile
`
`UA
`
`A
`
`G
`C-G
`U-A
`15.2 c_~'6.2
`151 A__Ul6.1
`
`Stem
`
`II
`
`A,;‘4
`
`5’G G C C,Ai2
`I/II
`& C G G1y9
`
`676G
`GUA
`
`17
`
`’
`“A
`
`CCAC3
`
`2.1
`3C
`54u
`
`Stem I
`
`Figure 1
`structure
`RNA secondary
`Hammerhead
`and cleavage site. The secondary struc-
`ture of the all-RNA hammerhead
`ribozyme
`used for structural determination, consist-
`ing of a 16nucleotide
`enzyme strand and
`a 25nucleotide
`substrate strand. The con-
`served bases shown as red
`letters are
`required for catalytic activity. The cleavage
`site is indicated.
`
`such as Mg2+ and Pb2+, suggesting that
`the catalytic pocket in the hammerhead
`ribozyme is also capable of binding the
`catalytically active divalent metal ion.
`Difference Fourier analyses of the all-
`RNA hammerhead ribozyme crystals re-
`veal a number of peaks of different elec-
`tron density, which we have assigned as
`Mg(H,O),*+ complex ions, based on dis-
`tance geometry criteria. Included is a
`single peak found near
`the catalytic
`pocket corresponding
`to a Mg(H,O),‘+
`complex
`ion, which can make hydro-
`gen-bonding contacts with the exocyclic
`amines on C-3 in the catalytic pocket,
`and on C-17, the cleavage-site nucleo-
`tide. The cytosine, corresponding
`to C-3
`in the hammerhead RNA, makes similar
`contacts with hydrated metal ions in
`the tRNAPhe uridine turn.
`a mechanism,
`We have proposed
`based on the position of the Mg(l-l,0)62+
`complex ion near the catalytic pocket
`of the hammerhead
`ribozyme, as well
`as on the similarly situated metal-bind-
`ing sites in the uridine turn of tRNAPhe,
`in which the Mg(H,0),2+ complex
`ion
`
`Table II. Relative cleavage
`
`rates for hammerhead
`
`ribozyme with various divalent metalsa
`
`Metal
`
`Ca2+
`Mg2+
`Mn2+
`co*+
`Cd2+
`Pb*+
`
`pKab
`
`Relative rate
`
`Relative
`
`[M*+(OH)-]
`
`Hardness
`
`Pauling’s ionic radius
`
`12.9
`11.4
`10.6
`10.2
`9.6
`7.7
`
`l/16
`1.0
`-10
`-10
`-6-10
`No cleavage
`
`l/32
`1.0
`6.3
`15.9
`63.1
`5013
`
`Hard
`Hard
`soft
`Borderline
`soft
`Borderline
`
`0.99A
`0.65A
`o.aoA
`0.72A
`0.97 A
`1.21A
`
`“See Refs 12, 13, 16, 17 and 31.
`bpKa = pH - log([A-]/[AH]).
`CRelative metal hydroxide concentration as compared
`
`to [Mg2+(OH)-] at pH7.0
`
`free
`
`in solution.
`
`221
`
`2
`
`

`

`(a)
`
`CYt
`I
`
`(b)
`
`CYt
`I
`
`(c)
`
`CYt
`I
`
`Figure 2
`and
`(a) ‘in-line’
`showing
`Illustration
`mechanisms.
`cleavage
`of hammerhead
`Chemistry
`for
`(c) Two possible
`sites
`cleavage.
`of phosphodiester-strand
`(b) ‘adjacent’ mechanisms
`ProR and pros
`non-bridging
`magnesium-mediated
`RNA-strand
`cleavage.
`phosphate
`oxygens
`are
`indicated.
`Note
`that one Mg*+ can
`fulfil both the
`roll of binding
`to the proR oxygen and
`of abstracting
`the 2’-proton by a metal-bound
`hydroxide,
`and
`that
`there
`is now experimen-
`tal evidence28
`against
`the existence
`of a second metal
`ion acting as a Lewis acid by bind-
`ing directly
`to the bridging 5’Yeaving oxygen.
`
`pocket by
`first ‘docks’ in the catalytic
`interacting with C-3 and C-17 as noted
`above (see Fig. 4a). Independent
`experi-
`mental corroboration
`for this
`initial
`in-
`teraction has recently emerged;
`removal
`of the exocyclic
`amine
`from either C-3
`or C-17 causes
`the dissociation
`con-
`stant
`for the catalytic Mg(H,0),2+ to in-
`crease by almost one order of magni-
`tude’“. Although
`both
`hammerhead
`RNA crystal
`structures
`have C at pos-
`ition 17, this C can be replaced with A
`or U (A-17 works almost as well as C-17,
`but the activity
`for U-17 is somewhat
`re-
`duced”). Essentially,
`the same ‘docking’
`interaction
`could still take place with A,
`where
`the Mg(H,0)62+ complex
`ion now
`interacts with the exocyclic
`amines on
`A-17 and C-3, but
`the analogous
`inter-
`action would be weaker
`in the case of
`U-17, which
`lacks an exocyclic
`amine,
`
`the observation
`and thus could explain
`that A replaces C at position
`17 more
`effectively
`than does U.
`We propose
`that
`the metal complex
`ion is then drawn
`in towards
`the cleav-
`age site 2’-OH group until
`it is within
`striking distance’.
`(The
`trajectory
`and
`final position
`of the complex
`ion are
`both
`inferred
`from the metal positions
`in the uridine
`turn of tRNAPhe.) As the
`metal
`is positioned,
`one of the six H,O
`molecules bound
`to the metal
`ion is dis-
`placed by the pro-R phosphate
`oxygen
`at the cleavage
`site, and that direct co-
`ordination with this phosphate
`oxygen
`assists
`in orienting
`and perhaps
`in ion-
`izing one of the
`remaining H,O mol-
`ecules which
`is now close
`to the 2’-OH
`group, i.e. binding
`the phosphate oxygen
`might lower the effective pKa of the hy-
`drated magnesium
`ion, thus activating
`
`Figure 3
`of
`structure
`three-dimensional
`(a) The
`ribozyme.
`of the hammerhead
`structure
`The crystal
`the substrate
`the all-RNA hammerhead
`ribozyme,
`showing
`the enzyme
`strand
`in red and
`strand
`in yellow. The cleavage-site
`base
`(C-17)
`is highlighted
`in green. Difference
`electron
`density
`interpreted
`as Mg(H20)c2+
`sites
`is shown as purple peaks
`containing
`blue spheres
`corresponding
`to the complex
`ion center of mass.
`(b) A corresponding
`schematic
`diagram
`II
`indicating
`the
`location
`of stems
`I, II and
`Ill, the catalytic
`pocket,
`the augmented
`stem
`helix and
`the
`tetraloop.
`The colour-coding
`is preserved
`and
`the essential
`nucleotides
`are
`shown as shadow
`letters.
`The universal
`numbering
`scheme
`is indicated.
`
`222
`
`TIBS 21- JUNE 1996
`
`a nucleo-
`it. Loss of a proton generates
`philic metal hydroxide, which
`in turn
`acts by abstracting
`the 2’-OH proton
`from the ribose
`in the cleavage site, ini-
`tiating nucleophilic
`attack at the phos-
`phorus and formation of the penta-coor-
`dinated 2’,3’-cyclic phosphate
`transition
`state or intermediate
`shown
`in Fig. 4b.
`
`In-line or adjacent nucleophilic attack?
`Both
`hammerhead
`crystal
`struc-
`tures”,’ reveal
`that their respective
`sub-
`strate
`analogues
`(e.g. DNA, and RNA
`with a 2’-0-methyl-cytosine,
`each incor-
`porated
`to prevent cleavage
`in the crys-
`tal) are not in a conformation
`that would
`support
`an ‘in-line’ mechanism
`of RNA-
`strand
`cleavageti. This might be due to
`the absence
`of an unmodified
`2’-OH at
`the cleavage
`site, but it is interesting
`to
`note
`that
`the same
`is true
`for tRNAPhe
`crystals with Pb2+ bound at the cleavage
`site, where
`the active 2’-OH is unpro-
`tected, except
`for being
`in a low-pH en-
`vironment”. However, experimental
`evi-
`dence obtained by three research groups
`clearly demonstrates
`that
`the hammer-
`head-RNA cleavage
`reaction
`proceeds
`by an in-line mechanism”-‘3
`(see Fig. 2a).
`The reaction
`requires
`a change
`in the
`conformation
`of the phosphate
`back-
`bone”,‘,” and results
`in inversion
`of con-
`figuration of the reaction product. This
`was demonstrated
`using thio-substituted
`phosphate
`oxygens. These experiments
`also demonstrated
`that
`the catalytic
`metal
`interacts with
`the pro-R, but not
`the pro-S phosphate
`oxygens.
`(like
`A simple ‘adjacent’ mechanism
`the one originally
`proposed
`for Pb2’-
`bound
`tRNAPh” crystals”‘), which would
`not
`require
`a
`rearrangement
`of
`the
`phosphate
`backbone
`conformation,
`but
`which would require pseudo-rotation
`of
`a penta-coordinated
`phosphate
`inter-
`mediate10,“4, leading
`to retention
`of con-
`figuration
`in the productZ4, is ruled out.
`
`One metal ion or more?
`opportuni-
`There are three potential
`ties
`for catalysts
`to accelerate
`the
`hammerhead
`self-cleavage
`reaction. The
`first is for a base catalyst
`to abstract
`the
`proton
`at the cleavage
`site 2’-OH; this
`role appears
`to be fulfilled by a divalent
`metal hydroxide,
`as discussed
`above.
`The second
`is for this same metal
`ion,
`or possibly another,
`to polarize the phos-
`phate by binding
`to the pro-R oxygen
`directly. The geometric
`constraints
`per-
`mit both roles to be fulfilled by a single
`metal
`ion, and the single metal mecha-
`nism possesses
`the added
`advantage
`of allowing
`the Mg(H,O),“+ to become
`
`3
`
`

`

`TIBS 21-
`
`JUNE 1996
`
`activated by lowering the effective pKa
`upon binding
`the pro-R oxygen, as
`noted
`above. Both of
`these well-
`established
`interactions are illustrated
`in Fig. 2b. The third, and more con-
`tentious, opportunity
`for metal cataly-
`sis is for an acid to stabilize
`the 5’-
`bridging-oxygen
`leaving group as the
`scissile bond breaks. This can, in prin-
`ciple, be accomplished either by proton-
`ation of
`the 5’-oxygen as negative
`charge begins to accumulate
`(general
`acid catalysis) or by direct coordination
`of the 5’-oxygen with a divalent metal
`ion such as Mg*+ (Lewis acid catalysis).
`The latter mechanism has been pro-
`posed based on molecular orbital calcu-
`lations of a model compound in the gas
`phasez5. Evidence for the absence of a
`kinetic isotope effect in hammerhead-
`ribozyme phosphodiester
`cleavage has
`recently been obtained,
`indicating the
`non-existence of a proton-transfer pro-
`cess
`in the rate-limiting step of the
`cleavage reactionz6. This result was in-
`terpreted
`to suggest that the 5’-oxygen
`is not protonated
`by a general acid
`catalyst, but rather is bound directly by
`a Lewis acid catalyst such as a second
`Mg*+ ion, as shown in Fig. 2b. However,
`any mechanism in which the rate-limit-
`ing step of the reaction does not in-
`volve proton abstraction or transfer
`is
`equally consistent with the data. For
`example, a mechanism
`in which re-
`arrangement of the phosphate back-
`bone into a conformation
`suitable for
`in-line attack
`is rate-limiting” should
`also show a lack of the kinetic isotope
`effect. Indeed, substitution of the leav-
`ing oxygen with a sulphur yields a
`hammerhead
`substrate whose
`leaving
`group now should be stabilized by soft
`divalent metal
`ions
`relative
`to
`the
`harder Mg*+, if the proposed Lewis acid
`catalyst exists. However, unlike the case
`of the group I intror?‘, no such reaction-
`rate acceleration was observed,
`sug-
`gesting that the leaving oxygen is not
`stabilized by a divalent metal ion bind-
`ing directly
`to it**. These findings are
`likely to be generalized
`for the other
`small
`self-cleaving RNAs
`including
`#NAPhe. In the case of tRNAPhe, as in the
`case of the hammerhead ribozyme, it is
`interesting to note that only one metal
`ion can be found at the cleavage site in
`the crystal structure.
`
`Concluding remarks
`The crystal structure of the hammer-
`head ribozyme, like that of tRNAphe eluci-
`dated 21 years before2g,30, has revealed
`much information about RNA structure
`
`Figure 4
`Metal binding and catalysis. (a) A potentially catalytic Mg(H,0),2+
`site is located adjacent
`to the exocyclic amines of C-17 and C-3 in the hammerhead
`ribozyme catalytic pocket.
`(b) A possible structure of the hammerhead RNA transition-state. The implications of the
`experimental biochemical
`results for the mechanism of hammerhead RNA catalytic cleav-
`age are: (1) that a hydrated metal binds directly to the pr@R phosphate oxygen at the
`cleavage site as (or possibly after) one of its chelated water molecules
`ionizes to form a
`metal hydroxide, and this nucleophile abstracts
`the
`labile proton from the 2’-OH of the
`cleavage-site base; (2) the reaction proceeds by an in-line, S,2(P) mechanism; and (3) the
`phosphate backbone must undergo a conformational change before or during cleavage
`to
`make
`in-line attack possible. Our proposed mechanism, based on the hammerhead RNA
`structure as well as comparisons with metal binding in the uridine turn of tRNAPhe, adheres
`to these three conditions. Substitution of C-17 with adenosine or uridine at the active site
`maintains a functional
`ribozyme, although U-17 functions
`less well than A-17 and C-17
`(Ref. 12). This fact is accounted for in both stages of the proposed mechanism.
`In the first
`step, the interaction with the exocyclic amine could still take place with A, but the analo-
`gous interaction would be somewhat weaker
`in the case of U-17, which lacks an exocyclic
`amine.
`In the second step, the base itself of the cleavage-site nucleotide stacks on A-6 in
`the catalytic pocket. Such a stabilization
`interaction may also take place with adenosine or
`uridine substituting for cytidine at the cleavage site.
`
`and function. In addition, the hammer-
`head RNA structure allows many new
`insights into how the three-dimensional
`structure mediates catalytic cleavage,
`including how the cleavage-site base is
`positioned
`in the uridine turn or cata-
`lytic pocket of the molecule, and how
`this pocket might bind and position the
`hydrated magnesium
`ion responsible
`for catalysing the first step of the ham-
`merhead cleavage
`reaction. However,
`both hammerhead RNAs used for eluci-
`dating the crystal structure are essen-
`tially ribozymes bound
`to substrate-
`inhibitor
`analogues
`(either with a
`2’-hydrogen or a 2’-methoxyl replacing
`the 2’-OH at the active site), and this, by
`necessity, gives only partial information
`about
`the cleavage-reaction mecha-
`nism. What remains to be elucidated,
`through the use of other modified bases
`and by time-resolved crystallographic
`techniques,
`is the structure of the reac-
`tion intermediate(s)
`complete with all
`catalytic metals bound unambiguously.
`
`Acknowledgements
`We thank J. Finch, M. Gait, D. Brown,
`A. Kirby, S. Price, K. Nagai, G. Varani,
`B. Stoddard, K. Flaherty, D. McKay,
`0. Uhlenbeck and two anonymous refer-
`ees for helpful advice. W. G. S. thanks the
`American Cancer Society for a postdoc-
`toral fellowship (grant number PF-3970).
`
`References
`1 Cech, T. R. (1993)
`in The RNA World
`(Gesteland, R. F. and Atkins, J. F., eds),
`pp. 239-269, Cold Spring Harbor Press
`2 Gopalan, V., Talbot, S. J. and Altman, S. (1994)
`in RNA-Protein
`interactions
`(Nagai, K. and
`Mattaj,
`I. W., eds), pp. 103-126,
`IRL Press
`3 Symons, R. H. (1992) Annu. Rev. Biochem. 61,
`641-671
`4 Symons, R. H. (1994) Curr. Opin. Struct. Biol. 4,
`322-330
`5 Tuschl, T., Thomson, J. B. and Eckstein, F.
`(1995) Curr. Opin. Struct. Biof. 5, 296-302
`6 Pley, H. W., Flaherty. K. M. and McKay, D. B.
`(1994) Nature 372, 68-74
`7 Scott, W. G., Finch, J. T. and Klug, A. (1995)
`Cell 81,991-1002
`8 Sampson, J. R. et a/. (1987) Co/d Spring Harbor
`Symp. Quant. Biol. 54, 267-275
`9 Brown, R. S. et a/. (1983) Nature 303,
`543-546
`10 Brown, R. S., Dewan, J. C. and Klug, A. (1985)
`Biochemistry 24,4785-4801
`11 Chapman, K. B. and Szostak, J. W. (1994)
`Curr. Opin. Struct. Biol. 4, 618-622
`12 Dahm, S. C., Derrick, W. B. and Uhlenbeck, 0. C.
`(1993) Biochemistry 32,13040-13045
`13 Pan, T., Long, D. M. and Uhlenbeck, 0. C.
`in The RNA World (Gesteland, R. F. and
`(1993)
`Atkins, J. F., eds). pp. 271-302,
`Cold Spring
`Harbor Press
`14 Ruffner, D. E., Stormo, G. D. and Uhlenbeck, 0. C.
`(1990) Biochemistry 29, 10693-10702
`15 Pyle, A. M. (1993) Science 261, 709-714
`16 Pearson, R. G. (1963) J. Am. Chem. Sot. 85,
`3533-3539
`17 Fratisto da Silva, J. J. R. and Williams, R. J. P.
`in The Biological Chemistry of the
`(1993)
`Elements, pp. 34-36, Clarendon Press
`18 Dahm, S. C. and Uhlenbeck, 0. C. (1991)
`Biochemistry 30, 9464-9469
`19 Sigurdsson, S. Th., Tuschl, T. and Eckstein, F.
`(1995) RNA 1, 575-583
`
`223
`
`4
`
`

`

`COMPUTERCORNER
`
`a:.
`
`20 Murray, J. B., Adams, C. J., Arnold, J. R. P. and
`Stockley, P. G. (1995) Biochem. J. 311, 487-494
`21 van Tol, H. et al. (1990) Nucleic Acids
`Res.
`18,
`1971-1975
`22 Slim, G. and Gait, M. J. (1991) Nucleic Acids
`Res.
`19,1183-1188
`23 Koizumi, M. and Ohtsuka, E. (1991)
`
`Rev. Biochem. 54,
`
`5145-5150
`30,
`Biochemistry
`24 Eckstein, F. (1985)
`Annu.
`367-402
`25 Taira, K. et al.
`3, 691-701
`Eng.
`Protein
`(1990)
`26 Sawata, S., Komiyama, M. and Taira, K. (1995)
`J. Am.
`Chem.
`Sot.
`117,
`2357-2358
`27 Piccirilli, J. A. et al.
`(1993)
`Nature
`
`361,
`
`85-91
`
`TIBS 21- JUNE 1996
`
`28 Kulmelis, R. G. and McLaughlin, L. W. (1995)
`J. Am. Chem. Sot. 117, 11019-11020
`29 Robertus, J. D. et al. (1974) Nature 250,
`546551
`30 Kim, S-H. et al. (1974) Soence 185, 435-440
`31 Burgess, J. (1978) Metal
`in Solution, Ellis
`ions
`Horwood Ltd
`
`taken when studies are performed with
`any
`type of hybridization
`reaction,
`be-
`cause
`isolation
`of DNA or RNA by
`precipitation
`with
`tRNA carrier
`could
`cause
`false positives
`due
`to the carry-
`over of contaminating
`nucleic acids.
`Linear polyacrylamide (LPA)
`can
`carrier
`a 5%
`easily be made by polymerizing
`w/v acrylamide
`solution
`in 40m~ Tris,
`20mM
`sodium
`acetate,
`1 mM EDTA
`(pH7.8),
`together with 0.01~01s of 10%
`ammonium
`persulphate
`and 0.001 vols
`TEMED. When
`the
`solution
`becomes
`viscous
`(15-30mins)
`the polymer
`is
`precipitated with 2.5~01s of ethanol and
`centrifuged
`for 5mins when
`it forms
`into a clot. The pelleted LPA is dried
`and 20~01s of sterile water added,
`then
`left overnight
`to swell. Afterwards,
`the
`LPA stock is mixed by pipetting.
`10 ~1 of
`a 1 x LPA (0.25%) or 2 ~1 of a 5 x (1.25%)
`LPA stock
`is added
`to DNA samples
`in
`100 to 400~1 and 2.5~01s of ethanol
`added for precipitation’.
`Netters
`generally
`use 1~1 of 0.25%
`LPA and 0.1 vols of 3 M sodium
`acetate
`or 0.20-0.25~01s
`of 10 M ammonium
`acetate
`and
`2-2.5~01s
`of
`absolute
`ethanol
`for volumes of DNA solution up
`to 50 ~1. They also advise
`that the chill-
`ing step
`is unnecessary
`and that
`it can
`generally
`be disregarded. One person
`wrote
`that routinely
`placing
`the mixed
`samples on ice for 15-30mins,
`followed
`by centrifugation
`for 30mins
`at 4°C is
`more than sufficient
`in most cases, and
`if more
`than 100 ng $’
`of DNA is pres-
`ent, then
`there
`is no need for chilling at
`all and
`the precipitated
`DNA can be
`held at room temperature”.
`David A. Johnston
`(daj@nhm.ac.uk)
`wrote that he tested various amounts of
`plasmid DNA ranging from 600 ng to 1 ng,
`and determined
`by comparison
`on an
`ethidium
`bromide-stained
`agarose
`gel
`that there was no apparent
`loss of DNA
`when as little as 4ng were used for pre-
`cipitation. As this amount of DNA is very
`close
`to the minimum
`that can be de-
`tected using ethidium
`bromide
`agarose
`gel electrophoresis,
`it is likely that even
`smaller
`amounts
`can be recovered.
`In
`a study
`using more
`sensitive
`radio-
`active
`labelling,
`it has been
`reported
`that 20 bp of DNA in the 20 pg range can
`
`Methods and reagents
`
`Carriers for precipitating nucleic acids
`in the
`reagents
`is a unique monthly
`column
`that highlights
`current
`discussions
`and
`Methods
`bionet.molbio.methds-reagnts,
`available
`on the
`Internet.
`This month’s
`column
`pro-
`newsgroup
`tips
`for the precipitation
`of DNA and RNA samples.
`For details on how to partake
`in
`vides some
`the newsgroup,
`see
`the accompanying
`box.
`
`A recurring question on methds-reagnts
`concerns
`the best method
`of precipi-
`tating DNA or RNA when
`preparing
`them for enzymatic
`reactions. The most
`common
`technique
`for precipitation
`of
`DNA is with
`the addition
`of 0.1~01s of
`3M sodium
`acetate
`(pH5.5) and either
`2-2.5~01s of ethanol
`or 0.8-1.0~01s of
`isopropanol.
`The mixture
`is placed
`at
`-70°C
`for 15mins
`to
`several
`hours
`before being centrifuged
`at top speed
`in an eppendorf
`table
`top centrifuge
`for lo-15 mins at 4°C (Ref. 1).
`
`Oh pellet, sweet pellet
`Even though claims of 100% recovery
`are sometimes made, pessimistic
`netters
`feel this
`is over-estimated
`and
`in prac-
`tice they
`typically
`expect
`to lose up to
`50% of their DNA upon precipitation,
`es-
`pecially
`if the DNA is less
`than 200 bp
`long or of low concentration.
`They there-
`fore feel the necessity
`of doing every-
`thing possible
`to prevent
`such losses.
`
`New WWW service from BIONET
`
`the bionet
`to
`posted
`latest messages
`The
`as well as all past archived messages
`are
`located
`at net.bio.net
`and all you will need
`to do
`in order
`to read and/or
`post
`to any of
`the newsgroups
`is point your World Wide
`Web browser
`to the URL http://www.bio.net
`and
`then click on the ‘Access
`the BIOSCI/
`bionet Newsgroups’
`hyperlink.
`
`now gives
`system
`archiving
`A new hypermail
`of USENET without
`requir-
`you the advantages
`ing a local news server.
`The message
`head-
`ers are
`threaded
`by default,
`but messages
`can also be displayed chronologically or sorted
`by author or subject
`line. This capability gives
`you,
`in effect,
`a threaded
`newsreader
`through
`the Web.
`If you have
`any questions
`or en-
`counter
`any problems with
`the new server,
`please
`report
`them
`to biosci-help@net.bio.net
`
`of DNA
`the small amount
`Although
`used for most molecular biology experi-
`ments
`(less than 2 kg) would cause
`the
`DNA sample
`to be invisible
`to the naked
`eye, some netters
`say that seeing a pel-
`let of DNA in the bottom of the micro-
`centrifuge
`tube can be a real psycho-
`logical
`boost
`along
`the way when
`several
`steps of DNA manipulations
`are
`to be performed
`in combination,
`es-
`pecially when
`the cleaning
`up process
`involves
`a precipitation
`at the end of
`each
`step. Most
`researchers
`would
`probably
`agree that a pellet
`is a positive
`sign that things are going well, and can
`even provoke a sigh of relief that, after
`rinsing with 70% ethanol,
`the DNA pel-
`let has not been
`accidentally washed
`out of the tube and lost down the sink.
`To see a pellet when precipitating very
`small amounts of DNA, a co-precipitant
`or
`carrier can be a real advantage. The type
`of carrier
`to be added will depend on
`what the DNA is to be used for after pre-
`cipitation,
`and the following
`tips should
`help you in selecting an appropriate
`one.
`Spermine or tRNA.
`Some people
`add
`0.1 vols of 100 mM spermine
`to precipi-
`tate DNA or use a final concentration
`of
`50 kgml-’
`bacterial
`or yeast
`transfer
`RNA as carrier2,“. However,
`spermine
`does not precipitate DNA below 60 bp
`long and can be tricky
`to remove
`later.
`Transfer RNA can also be a real prob-
`lem. In past discussions,
`one netter
`re-
`ported
`having
`a problem when doing
`RNase protection
`assays
`- extra pro-
`tec