Activation of Transcription at Specific Promoters by Glycerol

Abstract Glycerol added to a transcription system containing λgal DNA and Escherichia coli RNA polymerase holoenzyme stimulates total RNA synthesis and replaces the requirement for cyclic adenosine 3' : 5'-monophosphate (cAMP) and cAMP receptor (CRP) for promotion of gal RNA synthesis. The stimulatory effect which is proportional to the concentration of glycerol up to 20% (v/v) occurs at the level of preinitiation complex formation. In stimulating gal transcription, glycerol appears to act at or close to the same site as cAMP and CRP because: (a) glycerol has little effect on gal transcription at saturating levels of cAMP and CRP; (b) although glycerol stimulates transcription from λgal DNA bearing revertible gal promoter mutations, transcription from DNA containing a gal promoter deletion is not stimulated; (c) glycerol stimulation of gal promoter mutants is inhibited by cAMP and CRP; and (d) with either glycerol or cAMP and CRP, transcription of the promoter-proximal galE region precedes transcription of the galKT region. The effect of glycerol on λ early transcription is to stimulate early r-strand RNA synthesis but not early l-strand RNA synthesis. However, if the DNA template contains a defective λsex promoter, the decreased levels of early l-strand RNA are restored to normal by glycerol. Ethylene glycol, dimethylsulfoxide, sucrose, and 1,3-propanediol, all of which lower the Tm of DNA, also stimulate both total RNA and gal RNA synthesis. Furthermore, the glycerol-promoted formation of the gal preinitiation complex is strongly dependent on the preincubation temperature. At lower temperatures, complex formation requires increased glycerol or decreased KCl concentrations. These findings suggest that glycerol may act by melting or changing the conformation of the DNA promoter regions.

In stimulating gal transcription, glycerol appears to act at or close to the same site as CAMP and CRP because: (a) glycerol has little effect on gal transcription at saturating levels of CAMP and CRP; (b) although glycerol stimulates transcription from Xgal DNA bearing revertible gal promoter mutations, transcription from DNA containing a gal promoter deletion is not stimulated; (c) glycerol stimulation of gal promoter mutants is inhibited by CAMP and CRP; and (d) with either glycerol or CAMP and CRP, transcription of the promoter-proximal gaZE region precedes transcription of the gaZKT region. The effect of glycerol on X early transcription is to stimulate early r-strand RNA synthesis but not early Z-strand RNA synthesis.
However, if the DNA template contains a defective Xsex promoter, the decreased levels of early Z-strand RNA are restored to normal by glycerol.
Ethylene glycol, dimethylsulfoxide, sucrose, and 1, 3propanediol, all of which lower the T, of DNA, also stimulate both total RNA and gal RNA synthesis.
Furthermore, the glycerol-promoted formation of the gal preinitiation complex is strongly dependent on the preincubation temperature. At lower temperatures, complex formation requires increased glycerol or decreased KC1 concentrations. These findings suggest that glycerol may act by melting or changing the conformation of the DNA promoter regions.
Cyclic AX" and CRP act together with RX-1 polymerase to form a rifampill-resistant preinitiation cornplrs at the gal promoter. gal repressor interacts with th(x gnl operator to prcvcnt thcx formation of this complex. l'hr~ syntlirsis of early X transcripts, on the other hand, is not bcliclvcxd to b(, und(~r positive control; RNA polymcrase per se forms n c~omplrs witli the X early promotrrs, sez+ and z+ (5,6). X rc~prc~ssor csc>rts nrgativc control by binding to the X operators 0, and OH and iutrrfcring with complex formation arid possibly c+mgntion as well (7,8). Another protein factor, rko, is probably illvolvc,cl in the termination of both gal and X transcription (6,9). I'rclviously WC have> rcportcd that the glycrrol present in the gal rrpressor storagr buffer increased gal and total RNA syrlthesis (4). In this paper we report that glycerol will: (a) substitutcl for ca;\.\II' and ('RI' to stimulate gal trallscriptiori; (b) incarc,asc X Carl\-r-strand transcription from the z+ promoter but not rarly l-strantl transcription from the sex+ promoter; and (c) artivatr transcription from defective promoters in gal or X T)S:\.
Glycnerol arts 1)~ stimulating preinitiatiou complex formatioil at gal autl X promoters, probably by producing local alterations in thr structure of the I)N,\ templatc~. tion mixture was incubated at 37' for 10 miti to form a preinitiation complex. Transcription was started by the addition of MgC12 (final concentration, 10 mM) and rifampin (10 rg per ml) and the RNA made was measured after a lo-min incubation, unless otherwise noted. The amount of gal RNA was measured by DNA-RNA hybridization (2). To measure the amount of X early Z-strand or X early r-strand RNA, the [sH]RNA was first hybridized to filters containing Xirnrnzl DNA or Ximm4 DNA, respectively.
Portions of the supernatant fluids from these prehybridization mixtures were then hybridized to the Z-strand or r-strand of X DNA by liquidliquid hybridization.
Conditions of filter prehybridisation and liquid-liquid hybridization for X RNA were the same as those described in the assay of gal RNA (2).

Glycerol Stimulates
Transcription-The DNA of Xgal bacteriophage contains both X and E. coli sequences as shown in Fig. 1. When used as a template for transcription, RNA polymerase by itself will transcribe certain phage sequences but will not tran- scribe the bacterial gal operon. The addition of CAMP and CRP to the transcription system has no detectable effect on transcription initiated at X promoters, but stimulates the transcription of the gal DNA sequences at least lo-fold (2). Cyclic AMP and CRP act by increasing the formation of rifampinresistant preinitiation complexes between RNA polymerase and the gal promoter (2).
We observed that the addition of glycerol to a reaction mixture containing Xgal DNA and RNA polymerase stimulated total RNA synthesis (Fig. 2). The st#imulation was proportional to the glycerol concentration up to 20% (v/v) glycerol, 2 (left) and 3 (center). Effect of glycerol on rifampinresistant gal RNA and total RNA synthesis on wild type and gal promoter-deletion templates. DNA, nucleoside triphosphates, RNA polymerase and salts as described under "Materials and Methods" with or without CAMP and CRP were mixed together with various concentrations of glycerol.
After a lo-min preincubation at 37", MgCL and rifampin were added and the amounts of gal RNA and total RNA synthesized for 10 min at 37" were determined. XgaZ and xgu& DNAs were used as templates in the experiments of Figs was mixed together with either 15% (v/v) glycerol (0, 0) or CAMP and CRP (m, 0) at 0". After a lo-min preincubation at 37", MgClt and rifampin were added and the reaction was stopped at various times. The RNA products were hybridized either to the l-strand of XgaZDNA (galKTE RNA) (0, at which point RNA synthesis was increased 3.fold.
This assay measures the number of rifampirl-resistant RNA polymcrase-DNA complexes formed during a IO-min preincubation.
These complexes are thought to represent associations between RNA polymerase and 1)X-i promoter regions (14). Glycerol was therefore increasing the affinity between the promoters and RKA polymerase and/or stimulating the formation of new rifampitlresistant associations betwvten RNA polymerase and DSA. This finding supports the idea that glycerol affects promoter-RX. 4 polymerase interactions; if glycerol introduced new RNA initiation sites on DNA, an additive effect of cA;\lP and CRP and glycerol might have been espected. Requirement of cs factor for gal RNA synthesis kgal DNA, nucleoside triphosphates, and the other components described under "Materials and Methods" were mixed at O", together with the various additions indicated in the first and second columns. The concentration of RNA polymerase core enzyme and c factor were 25 pg per ml and 25 pg per ml, respectively.
After a IO-min preincubation at 37", RNA synthesis was initiated by the addition of MgClz and the amount of gaZItNA and total RNA synthesized after a 10.min incubation at 37" were determined Xga1326 which is deleted for the gal promoter (as well as for a portion of galE), should not serve as a template for gal RNA synthesis.
Indeed, this is the case; neither c-UII' and CRI' nor glycerol will stimulate gal RSX synthesis with this DNA (Fig. 3). Thus glycerol-promoted gal transcription must initiate within the 326 deletion.
Note that while gal RN;h synthesis from this DNA was not stimulated by glycerol, total RNA synthesis was increased about 3.fold.
This finding indicates that glycerol also activates transcription of X genes. We shall return to this point below. 2. To characterize further the glycerol-dependent site of gal RNA initiation, we compared the rate and sequence of transcription of portions of the gal operon in the presence of glycerol and in the presence of cALlI' and CRI'.
In these experiments, RNA was synthesized from a Xgal DNA template and hybridized to the l-strand of Xgal DNA or to the l-strand of XgabT which is deleted for the galE region.
The former hybridization measures total gal RNA (KTE) ; the latter, the gal RNA derived from the promoter-distal portion of the operon (KT). As is shown in Fig. 4 the chronology of transcription is identical in glycerol and in CAMP and CRP. gdlE RNA appears without lag and about 1 min before galKT RNA is seen, a value consistent with prior transcription of the galE gene.2 Since RNA polymerase in vi&o at 37" catalyzes the incorporation of about 1100 nucleotides per min, we estimate that the initiation site for glycerol-dependent gal transcription is located within a few hundred nucleotide pairs of the CAMP and CRPdependent initiation site.
synthesis of gal RNA in the presence of cALlI' and CRP, indicating that they are defective in gal transcription. However, both DNAs are excellent templates for gal transcription when glycerol is added to the preincubation misture. Note that CAMP and CRI' partially inhibit the action of glycerol indicat,ing that they bind to the mutant template.
This observation is cow sistent with the notion that CAMP and CRY and glycerol affect the same IINrZ region, and in the case of the gulp-3 and gulp-21 1 templates, their actions are antagonistic.
Glycerol stimulates transcription from these mutant promoter templates by activating the formation of preinitiation complexes. In the esperiments reported in Table III, we used the synthetic  polynucleotide, poly(rI), to distinguish between free RNA polymerase and RNA polymerase bound in a preinitiation complex. l'oly(r1) has one advantage over rifampin; poly(r1) does not inactivate RNA polymerase bound to a promoter, whereas rifampin eventually inactivates even the bound form of the enzyme (4). Both compounds rapidly inactivate free RNA polymerase.
3. The binding of RNA polymerase to promoter sites requires holoenzyme, i.e. core plus sigma. We therefore studied the effect of glycerol on gal transcription with RNA polymerase core enzyme and reconstituted RNA polymerase holoenzyme. The results are presented in Table Il.
Note that the stimulation of gal RNA synthesis by either CAMP and CR1 or glycerol depends upon the presence of sigma; this result lends further support to the idea that glycerol acts at promoter sites.

Glycerol
Activates Defective gal Promoters--Although glycerol is ineffective in stimulating the template activity of DNA bearing a gal promoter deletion mutation, it strongly stimulates gal transcription from DNA with revertible gal promoter mutations (galP-211, gulp-3) (see Ref. 15). As seen in Fig. 5, A and B, DNA extracted from XgalP-211 or XgalP-3 will not support the Note that glycerol activates gal transcription from XgalP-211 DNA when added prior to poly(rl) but not after (Table III, Lines 2 and 3). We conclude that glycerol promotes the formation of gal preinitiation complexes on this template and these complexes: (a) are stable for at least 7 min in the presence of poly(rI) ; and (b) can form in the absence of the four ribonucleoside triphosphates (Line 4). Preincubation of XgalP-211 DNA with RNA polymerase and cAXP and CRP does not yield a poly(rI)-resistant complex; addition of CAMP, CRP, and RNA polymerase prior to poly(r1) and glycerol as shown in Line 5 does not result in gal RNA synthesis.
Since cAlM1' and CRP produce only a partial inhibition of the glycerol-stimulated gal transcription from the template (see Fig. 5A), even when added prior to glycerol (data not shown), we conclude that a preinitiation complex will not form between RNA polymerase and the mutant DNA, even in the presence of CAMP and CRI'. The P-211 mutation, is, therefore, a mutation affecting the binding of RNA polymerase or CAMP and CRP to gal DNA whose phenotype is overcome in vitro by glycerol. These data are inconsistent with an alternative model in which the P-211 mutation only affects gal RNA propagation. 2 Richard Musso, manuscript in preparation. The initial components were mixed at 0", the temperature raised to 37", and various additions made at 7-min intervals.
Seven minutes after the last addition, MgCl, and rifampin were added. In the experiment of Line 4, four ribonucleoside triphosphates were omitted from the preincubation mixture. RNA synthesis was started by their addition together with MgC12 and rifampin.
The final concentrations of glycerol and poly(r1) are 10% (v/v) and 309 pg per ml, respectively.
RNP is RNA polymerase. We therefore tested the effect of glycerol on X early RNA synthesis.
Our analysis demonstrates that glycerol stimulates X early r-strand RNA but not X early I-strand RNA (Fig. 6). In this experiment, a template which contains the wild type promoters for these regions, Z+ and sex+, respectively, was used. However, when the template carries a defective promoter for X early l-strand RNA, sex1 (6) or sex3, the diminished synthesis of this transcript is restored to normal by glycerol (Fig. 7). It is of interest that the seal defect, which is less severe than the sex3 defect, is restored by a lower concentration of glycerol.
As in the case with gulp-3 and gulp-211, glycerol must be present during the preincubation period (data not shown).
These observations support the idea that glycerol can act directly on defective promoters to stimulate transcrip- DNAs from XgaZP-211 and XgalP-3 o_ : were used as template in the experiments of A and B, respectively. n 0 tion. It further suggests that the sex1 and sex3 mutations affect RNA polymerase binding, in agreement with the data of Chadwick ef al. (16). Mechanism of Action of Glycerol on Transcription--The proportional relationship between glycerol concentration and the stimulation of defective promoters, as well as the difference in glycerol concentration required to restore full activity to sex1 and sex3 mutants, suggested that glycerol works through interactions with DNA rather than with RNA polymerase.
Since glycerol has been reported to decrease the T, of DNA (lT), we measured the effect of preincubation temperature on preinitiation complex formation.
As shown in Fig. 8, the formation of preinitiation complexes between RNA polymerase and the gulP-211 promoter is strongly temperature-dependent; the lower the preincubation temperature, the higher the glycerol concentration required. Prolonging the preincubatjion time did not increase preinitiation complex formation; evidently the reaction is complete by 10 min.
We also tested the effect of DNA denaturing reagents related to glycerol (17,18) on both total RNA synthesis and on their ability to "suppress" the effect of the gulP-211 mutation.
We define suppression here as the restoration of promoter activity to t.he DNA template.
Dimethqlsulfoside and sucrose are even more potent than glycerol in stimulating total and gal RNA synthesis (Table IV) whereas isopropyl alcohol and n-propyl alcohol inhibited the reaction (data not shown).
Although glycerol and the other reagents mentioned above do reduce the dielectric constant of the reaction medium, this cannot explain their mechanism of action.
Xs shown in Fig. 9, omitting KC1 entirely from the mixture does not restore activity to the gulp-211 promoter.
Nevertheless KU, which stabilizes the DNA helix (19) and glycerol which destabilizes it, also have antagonistic effects on transcription from the gulp-21 1 template. Stimulation by 15% (v/v) glycerol is maximal when no KC1 is added. The inhibition of glycerol promoted gal RNA synthesis by KC1 can be overcome by raising the preincubation temperature (Fig. 10).
The relationship between preincubation temperature and X early I-strand RNA synthesis is demonstrated in Fig. 11A. Transcription from the wild type promoter is independent of temperabure over a 20" range, whereas the suppression of the sex3 mut.ation by glycerol shows a marked temperature de-  10 (left). Effect of KC1 on formation of preinitiation complexes in the presence of glycerol.
Conditions were similar to those in Fig. 8, except that the concentration of glycerol was fixed at 15% (v/v), and either KC1 was omitted (0) or 0.1 M KC1 was used (0).

FIG.
11. Effect of preincubation temperature on formation of preinitiation complex. Conditions were similar to those of Fig. 8, except for the following: A, AgulP-211 (0) or Xsez3 DNA (0) was used as a template and the concentration of glycerol was 15% 3. Suppression by glycerol of the Xsez3 mutation is achieved at a lower temperature than the suppression of gulp-(compare Fig. 11, A and B). These results may indicate that the sex3 mutation decreases the binding of RNA polymerase to the sex promoter while the galP-3 and galP-211 mutations affect the interaction between the gal promoter region and CAMP and CRI'.
Glycerol would then suppress the galP mutations indirectly by converting the gal promoter region to a CAMP and CRP independent form. An alternate possibility is that glycerol initiates gal transcription at a second site close to the CAMP and CRl' dependent promoter.
This model, however, requires a second, ad hoc, assumption that activation of one promoter in gal inhibits the activity of the other. DISCUSSION We have reported in this paper that the presence of glycerol in a purified in vitro system increases transcription by stimulating the formation of preinitiation complexes at bacterial and X promoter regions.
We have studied the transcription of a variety of operons and found: (a) the transcription of X early rstrand RNA is strongly stimulated by glycerol; (b) the X early l-strand transcription is not stimulated. However, when this transcription is reduced by the defective promoter mutations, sesl and sex3, activity is restored to normal by glycerol; and (c) for E. coli gal transcription, glycerol can replace CAMP and CRP and will also restore gal RNA synthesis to normal, when defective promoter mutations are present.
Glycerol appears to affect promoter sites rather than to cause nonspecific initiation of transcription because : (a) transcription factor sigma, which is required for the stable binding of RNA polymerase to promoter sites, is also required for the glycerol effect; (b) at saturating concentrations of CAMP and CRP where 40 (v/v). X early I-strand RNA was measured as described under "Materials and Methods." B, a preincubation mixture containing Xgal DNA was mixed either with 15% (v/v) glycerol (0) or with CAMP and CRP (m). Similar mixtures containing kg&P-211 (0) or XgalP-3 DNA (0) were mixed with 15% (v/v) glycerol. After a lo-min preincubation at the indidated temperatures, MgClz and rifampin were added and the reaction was further incubated for 10 min at 37". the gal promoter is fully active, little increase in gal transcription is seen when glycerol is also present; (c) glycerol also fails to stimulate gal transcription from a template deleted for the gal promoter region; (d) transcription of the galE region precedes by about 1 min the transcription of the promoter-distal galKT region with either CAMP and CRP or glycerol; (e) two defective X promoter mutants, sex1 and ses3, show different glycerol concentration dependence curves.
We believe that glycerol acts by changing the conformation of the DNA template rather than altering the properties of RNA polymerase: (a) the stimulation of RNA synthesis is proportional to the concentration of glycerol and, in the case of Xsezl and Xsex3, the more defective the mutant the higher the required glycerol concentration; (b) whereas preinitiation complex formation is not markedly temperature-dependent (over a 8-37' range, data not shown), glycerol-stimulated complex formation is exquisitely temperature-dependent; (c) the midpoint of the temperature transition is lowered either by increasing the glycerol or reducing the salt concentration.
Glycerol and salt have antagonistic effects on the stability of DNA, the former reducing and the latter elevating the !Z',. Other denaturing agents, i.e. sucrose, ethylene glycol, dimethylsulfoxide, and 1, Q-propanediol (17, 18) also stimulate both gal and total RNA synthesis.
The effect of ethylene glycol on the T,,, of DNA and its apparent conversion of DNA from the 13 to the C form (20) reflect a general change in the structure of DNA.
In light of this, how then would glycerol act specifically to activate promoter regions? One possibility is that glycerol produces a general alteration in the structure of DNA which results in the activation of certain promoter regions. Another possibility is that the promoter regions are uniquely affected by glycerol.
There is evidence that the promoter loci might be palindromic and therefore have an atypical secondary structure.
In the lac promoter-operator 4056 regions, palindromic symmetry has been proposed by Sadler and Smith (21) on the basis of genetic evidence.
In the case of X, the seal and sex3 mutations eliminate a site of cleavage of the Hemophilus influenza enzyme restriction3; restriction nucleases are known to act at sites of rotational symmetry (22)~ The apparent activation of the X Z+ promoter by glycerol is of some theoretical interest.
If glycerol acts on wild type promoters only to replace positive control factors, then X early rstrand transcription may normally be controlled by elements analogous to CAMP and CRP.
However, transcription of the X c1 gene, thought to be under positive cont.rol, is not stimulated by glycerol (data not shown).
It is known from the work of Brady and Leautey (23) that high concentrations of ethylene glycol, which inhibit over-all transcription from T4 DNA templates, stimulate the synthesis of new GTP-initiated RNA molecules and that this stimulation requires u factor. Whether this may occur with our templates and under our experimental conditions, where glycerol is stimulatory, remains to be determined.
Preliminary evidence indicates that 15y0 (v/v) glycerol stimulates preinitiation complex formation of T2 and calf thymus DNA templates (data not shown).
The marked effect of glycerol or1 transcription and its ubiquitous use as a stabilizer of proteins or as a component of gradient centrifugation mixes should be noted. The present study was, in fact, initiated by the observation that, the glycerol present in gal repressor preparations antagonized the action of the repressor. Finally our observation that glycerol and related substances can influence specifically the transcription of DNA might prove useful in elucidating the mechanism of initiation of RNA synthesis.