Biochimie (1996) 78, 1035-1042 Soci6t6 fran~;aise de biochimie et biologie mol6culaire / Elsevier, Paris

Differential contributions of two elements of rho-independent terminator to transcription termination and mRNA stabilization
H Abe, H Aiba*
Department of Molecular Bioiog3, School of Science, Nagoya Universi~.; Chikusa, Nagoya 464-01, Japan

(Received 3 December 1996: accepted 12 December 1996) Summary - - The hallmark features of rho-independent transcription terminators are a G+C-rich dyad symmetry sequence followed by a run of T residues on a sense strand. Both of these structural elements are required for efficient transcription termination. Besides its primary function, rho-independent terminators are also known to enhance expression of an upstream gene by stabilizing RNA in a few cases. The Escherichia coli crp gene encoding cAMP receptor protein (CRP) contains a typical rho-independent terminator. To gain further insight into the roles of the G+C-rich dyad symmetry sequence and the poly(T) tract both in transcription termination and mRNA stabilization, we constructed a series of variant crp terminators and analyzed their abilities regarding these two functions. Disruption of the G+C-rich dyad symmetry sequence almost completely eliminated terminator activity while disruption of the poly(T) tract reduced terminator activity significantly but not completely. Thus, the contribution of the G+C-rich dyad symmetry sequence to transcription termination is larger than that of the poly(T) tract. Disruption of the G+C-rich dyad symmetry region reduced expression of the upstream cip gene by accelerating the rate of mRNA degradation. However, disruption of the poly(T) sequence had no effect on the stability of the crp mRNA, indicating that the poly(T) tract plays no role in mRNA stabilization. When the crp terminator was replaced by terminators derived from other genes, the fusion genes expressed the crp mRNA at the same level as did the native crp gene, suggesting that the mRNA stabilization effect is probably a general nature of rho-independent terminators. rho-independent terminator / transcription termination / mRNA stability / poly(T) tract / stem-b~op structure introduction Rho-independent transcription terminators, also called factor-independent or intrinsic terminators, specify the 3' ends of many transcription units in prokaryotes [ ! ]. They arc also used as regulatory signals before structural genes in a number of operons to modulate transcription of the downstream genes [2]. The terminators consist of a G+C-rich dyad symmetry sequence immediately followed by a poly(T) tract on the sense strand, yielding an RNA with a stem-loop structure followed by a run of U residues. While mutational studies on several rho-independent terminators have revealed that both of these structural elements are required for efficient termination [3-7], how these elements are involved in transcription termination is not fully understood. In particular, the role of the poly(T) tract is rather , , ~ : o v e r s i a l . In a classical two-step model, the 3' tail of the nascent transcript is assumed to stimulate release of the RNA due to the low stability of the dA-rU hybrid after the formation of the G+C-rich RNA hairpin [3, 8, 9]. However, the role of the instability of the dA-rU base pairs in transcription termination has been questioned since there is no simple correlation of termination efficiency with the length of poly(T) tract [10]. An alternative model argues that transcription termination is closely coupled with inchworm movements of RNA polymerase and the poly(T) tract may act as an inchworming signal for elongating RNA polymerase [11 ]. Thus, the role of the poly(T) tract in rho-independent terminators remains to be an important subject for further studies. Besides transcription termination, rho-independent terminators are known to enhance gene expression by stabilizing upstream mRNA in some cases such as the E coli ti'p operon [12] and the lambda phage int gene [I 3]. However, only a small number of rho-independent terminators have been characterized regarding the mRNA stabilization effect. In other words, it has not yet been established whether the mRNA stabilization effect is generally applicable to other rho-independent terminators. Furthermore, structural determinants of m R N A stabilization have been poorly studied. For example, there is no report that addressed the role of the poly(T) tract in mRNA stabilization. The terminator of the Escherichia coli crp gene encoding cAMP receptor protein (CRP) fits closely the paradiem of a rho-independent terminator [14]. We showed previously that the ctp terminator acts efficiently to terminate transcription both in vitro and in vivo [15]. We also demon-

*Correspondence and reprints

1036 strated that the crp terminator functions to increase the crp expression by stabilizing the crp mRNA [15]. In this report, we have investigated the relative contributions of the G+Crich dyad symmetry sequence and the poly(T) tract to transcription termination and to mRNA stabilization using the crp gene as a model system. This is the first systematic study that addresses how two elements of rho-independent terminators contribute to both transcription termination and mRNA stabilization. We show that the contribution of the G+C-rich dyad symmetry sequence to transcription termination is more important than that of the poly(T) tract although both of these elements are required for efficient termination. In contrast, we found that the poly(T) tract plays no role in m R N A stabilization. In other words, the G+C-rich dyad symmetry sequence is solely responsible for mRNA stabilization. Thus, the structural determinant for mRNA stabilization does not coincide with that for transcription termination. We also demonstrate that several heterologous terminators can act an RNA stabilizer when placed downstream of the crp structural gene instead of the crp terminator. tants), the DNA sequence between the BssHll and Smal sites was replaced by each of the BssHll-Smal DNA fragments. To construct pHA7 derivatives in which the crp structural gene is fused to heterologous terminators, an additional restriction site (Mlul) was created near the translation termination codon in pHA7M0. The Mhd and Sinai regions of the resulting plasmid were replaced by synthetic DNA fragments corresponding to the trp attenuator (opAL T7 early terminator (T7Te), and rrnB terminator (rrnBTl) [10l with Mhd and Smal half-sites. The sequences of variant terminators were verified by direct DNA sequence analysis. The names and sequences of terminator variants are given in figure 2.

Construction of tetTninator-galK filsion plasmids and termination assay

Termination efticiencies of terminator variants were measured by using plasmid vector pKG!00, designed to quantitate terminator activity in vivo !161. The Haelll fragments containing each of the crp terminator variants were prepared from pHA7 derivativ~ s and ligated into the Smal site of pKGI00 (see fig 3). The resulting pKG i 00 derivatives in which the terminator was placed in the correct orientation were introduced into gall(- strain, R594 (gal-I gal2 lac rpsL s,pO) 1171. Cells harboring each of the pKGI00 derivatives were grown in LB medium containing 50 lag/mL of ampicillin at 37C to an OD~,()0of 0.8. Galactokinase activities were assayed according to the procedure of Adhya and Miller [ 18]. Each sample was assayed in triplicate, and the specific activities were calculated and averaged.

Materials and methods

Construction of terminator variants
The parent plasmid used in this study is pHA7 in which the crp gene is under the control of the bin promoter [ 141. A unique BssHil site is located in the dyad symmetry region of the crp terminator (see fig 1). Two new restriction sites ~'ere introduced in the flankin~ regions of the crp terminator by PCR mutagenesis: a SpI! site downstream of the translation termination codon of the c~p and a Sinai site downstream of the poly(T) tract. The resulting plasmid was designated pHA7M0. DNA fragments corresponding to each terminator variant with Sinai and Spll or BssHll half-sites at the ends were prepared by annealing complementary synthetic oligonucleotides. To construct variants with the dyad symmetry region (stem mutants), the DNA region between the Spli and Sinai sites was replaced by each of the synthetic SpII-Smal DNA fragments. To construct variants with the poly(T) tract (T tract mu-

Analysis of protein
TP2339 cells (F- xyl ih,A argHi AlacX74 Acva AcqJ) 1191 harboring each of the pHA7 derivatives were grown in LB medium containing 50 lag/mL of ampicillin at 37C. At ODt,,0 = 0.8, cells equivalent to 1.00D~,q,) were centrifuged and the pellets were suspended in 50 HL of H20. Tt~ solubilize cells, the cell suspensions were mixed with 50 laL of 2 x loading buffer (4% SDS, 10% 2mercaptoethanol, 125 mM Tris-HCI, pH 8.0, 20% glycerol, 0.2% bromophenol blue) and heated for 5 min at 90C. Total proteins equivalent to 0. ! ODt,o~ were loaded onto 0. 1% SDS-15% polyacrylamide gels and electrophoresed. After electrophoresis, the gels were stained with Coomassie brilliant blue. To quantify the amount of CRE the stained gels were scanned with an LKB UItrascan XL laser densitometer.

Analysis of RNA
The pHA7 derivatives containing the terminator variants were introduced into TP2339. The crp mRNA levels were analyzed by a quantitative S I nuclease assay as described [20]. Cells harboring each of the pHA7 derivatives were grown in I0 mL of LB medium containing 50 ~g/mL of ampicillin at 37C. Ai ODt,00 = 0.8, cells were harvested and total cellular RNAs were prepared. Tile lower strand of the 488-bp Hpall fragment labeled at its 5' end with 32p was hybridized with each of the RNA samples and treated with S I nuclease. Products of S l digestion were analyzed by electrophoresis on 6% polyacrylamide-8 M urea gels. To determine the degradation rate of the t'q~ mRNA, cells harboring each of the pHA7 derivatives were grown in 100 mL of LB medium containing 50 pg/mL of ampicillin at 37C. At OD600 = 0.8, rifampicin was added to 200 pg/mL and incubation was continued. Total cellular RNAs were isolated at time intervals after the addition of rifam-

. I

'Ill'

pHA7

488.bp

333-bp

Fig 1. Diagram of the crp region in pHA7. The relevant restriction

sites are shown. The black bar indicates the ceding region of the
crp which is under the control of the bla promoter. The arrow indicates the start site and direction of transcription. The major transcript is shown schematically below the restriction map. The 333-bp Haelll fragment was used for b~ vivo termination assay. The 488-bp Hpall fragment was used as a DNA probe for S I assay.

103/ BssHH

CACTCG~TCCCGTCGGAGTGGCGCGTTACCTGGTA~CATTTTGTTTCCCCCGA
M0 MI

Sp~ Sm~ CACTCG~TCCCGTACGGAGTGGCGCGTTACCTGGTAGCGCGC_CATTTTGTTTCCCGG_.G_GA CACTCGq~------~TCCCGTACGGAGTGGCGCGTTACCTGGTAGCGCGCCATTTTGTTCCCCGGGA CACTCGq~--~TCCCGTACGGAGTGGCGCGTTACCTGGTAGCGCGCCATTTTGTCC~CCGGGA CACTCG~TCCCGTACGGAGTGGCGCGTTACCTGGTAG .GCG~_CAFTTTGCCCCCCGGGA

M2
M3

M4
M5

CACTCGZ~TCCCGTACGGAGTGGCGCGTTACCTGGTAGCGCGCCAGAGAGCCCCCCOa.~QA CACTCGq~-~TCCCGTACGGAGTGGCGCGTTACCTGGTAGCGCGCCAAAAAAAAACCCGGGA CACTCG~TCCCGTACG-AGTTTATATGGAAATGGTAGCGCGCCATTTTGTTTCCCGGGA CACTCG~TCCCGTACGGAGTTTATATTTAAATTTTAGATATAAATTTTGTTTCCCGGGA CACTCGq~--~TCCCGTACG-AGTTTATATGGAAATGGTAGCGCGCCAAAAAAAAACCCGGGA

Mlui

M6
M7

M8

trpA CACGCG~TCAGATACCCAGCCCGCCTAATGAGCGGGCTTTTTTTTCCCGGGA rrnBT1

CACGCG~TAAAACGAAAGGCTCAGTCGAAAGACTGGGCCTTTCGTTTTCCC-G-C~A

77Te

CACOC~GTAATCACACTGGCTCACCTTCGGGTGGGCCTTTCTGCGTTTAT~A

Fig 2. The nucleotide sequences around the wild-type crp terminator and its variants. The mutational changes in the crp terminator are indicated as bold letters. Bold italic lettzrs indicate the sequences derived from trI,A, rrnBT i, and T7Te. Translation termination codon (TAA) of the crp gene is boxed. The relevant restriction sites are underlined. The inverted repeat sequences are shown by horizontal arrows.

picin. The RNA samples were subjected to S! nuclease assay and the S I digestion products were analyzed by electrophoresis on 6% polyacrylamide-8 M urea gels. DNA bands protected from S I digestion were visualized by autoradiography and quantified by Bioimage Analyzer BAS2000 (Fujix).

Results
Structures o f the crp terminator and its variants
DNA sequences of variant terminators used in this study are shown in figure 2. M0 is a control terminator which retains

the wild-type terminator sequence and has two base substitutions in the flanking region of the terminator. As shown later, these base changes did affect neither transcription termination nor mRNA stability. Variants M I, M2 and M3 are T cluster mutants tha. have one, two and three T to C changes in the distal ~'egion of the poly(T) tract, respectively. M4 and M5 are T cluster mutants in which the T cluster has been completely substituted by a random sequence without T and by an A cluster, respectively. Variants M6 and M7 are stem mutants which have base substitutions in the dyad symmetry region. Variant M6 has several base changes which completely disrupt the hairpin-forming sequence. M7 is a variant in which the G+C-rich stein has

1038
Haelll

terminator

HaeUl

"%',.,"..%. . o.
Sinai

--

! ,..o" ...,"

....'""

Pgal galK

pKG100 A

iv

activity compared to the control plasmid pKG 100. This reduction would be caused by possible rho-dependent and/or non-specific terminations within the inserted fragment. In any case, we used the pKGI00M8 as a terminator-deficient control to evaluate the effect of mutations on the termination. Namely, the expression of galactokinase from this plasmid was set as 100% read-through or 0% termination. The termination efficiency for the wild-type crp terminator was calculated to be 87%. Variant M0 retained the termination efficiency similar to the wild-type terminator, indicating that the introduction of two restriction sites in the flanking regions has little effect on transcription termination. The complete disruption of the dyad symmetry sequence by base substitutions (M6) essentially eliminated terminator activity. When the G+C-rich stem was converted to an A+T-rich stem (M7), terminator activity was also heavily impaired The data clearly indicate that the G+C-rich dyad symmetry sequence makes a crucial contribution to transcription termination, as expected. We next analyzed the effect of the mutations in the poly(T) tract region on termination. The substitution of the last T residue in the poly(T) tract to C residue (variant M l) reduced termination efficiency to 56%. Further substitutions of T residue to C residue (M2 or M3) caused little reductions in termination efficiency. Furthermore, variants M4 and M5 in which the T tract is totally disrupted showed termination efficiency that is only slightly lower than variant M 1. In other words,

Fig 3. Clonings of the crp terminator variants into the pKGI00 vector. The Haelll terminator fragments taken from pHA7 or its derivatives were cloned into the Smal site between the gal promoter and galK gene.

Table 1. hz vivo transcription termination efficiencies of the crp

terminator variants. The derivatives of pKGI00 carrying the terminator fragments were introduced in R594 cells and the galactokinase activities of the transformants were determined. The galactokinase units are expressed as nanomoles of galactose phosphorylated per min/A600..~. The percent termination is calculated from the ratio of the galactokinase units of each plasmid to that of pKG 100M8.

been converted to an A+T-rich stem. Finally, M8 is a double mutant in which both the G+C-rich stem and T cluster sequences have been totally disrupted by base substitutions.

Termination efficiencies of variant terminators in vivo

To assess the roles of the G+C-rich dyad symmetry sequence and the poly(T) tract in transcription termination, we measured the termination efficiencies of the mutated terminators by using the galK plasmid system. Haelll DNA fragments containing the terminator variants were inserted between the gal promoter and the galK reporter gene of pKGI00. The resulting plasmids (pKGI00 derivatives) were introduced into the gall(,- cells, and galactokinase activities of the cells were determined to monitor terminator activity of inserted fragments. The results are presented in table I. The galactokinase activity was significantly reduced when each of the terminator variants was inserted. It should be noted that even M8 DNA fragment in which two elements of the rho-independent terminator were totally eliminated exhibited about 50% reduction in galactokinase

Table.1. Termination efliclencies of the crp terminator variants

Plasmids Galacfokinase units

% Transcription termination

pBR322 pKG100 pKG-WT ~ ) pKG-M0 -JLu)n pKG-M1 pKG-M2 pKG-M3 pKG-M4 pKG-M5 pKG-M6 pKG-M7 pKG-M8 C) ._]k

<1 235.4 15.4 16.0 52.7 56.9 52.3 72.7 74.7 109.1 109.4 119.9

87.2 66.7 56.0 50.9 56.4 39.4 37.7 9.0 8.8 0

~ ~)n ~,

! 039 significant termination activity was retained even without the poly(T) tract. Taken together, we conclude that the poly(T) tract certainly plays an important role in transcription termination but its contribution is less prominent compared to that of the G+C-rich dyad symmetry sequence at the crp terminator.

r ~ ]

i F----7

WT MO M'l U2 U5 M6 M7

M8

Effects o f terminator mutations on the expression upstream gene

qf
CRP
===

N
. . ~ . -

To examine the effects of terminator mutations on expression of the crp, pHA7 derivativies were introduced in crp- cells. Total proteins prepared from cells harboring each plasmid were analyzed by SDS-PAGE. In this assay, the CRP protein was easily detected on a gel stained with Coomassie brilliant blue. As shown in figure 4, the levels of CRP in cells harboring stern mutants were about fourfold less compared to the wild-type terminator while the mutations in the poly(T) tract did not affect the level of CRP. We also examined the effects of mutations of the terminator region on the levels of crp mRNA by a quantitative S! nuclease assay by using the lower strand of the 488-bp Hpall fragment 32p-labeled at its 5' end as a DNA probe. In this assay an S l-resistant DNA of about 380 bases which correspond to the crp m R N A was produced. As shown in figure 5, the disruption of the stem region significantly reduced

e-

"~

lo

Fig 4. Effects of terminator mutations on the expression of CRE Total proteins from TP2339 cells harboring pHA7 derivatives were analyzed by SDS-PAGE using a 15% polyacrylamide gel as described in Materials and methods. The arrowhead indicates the CRP bands. Lane 1 is 0.5 lag of purified CRP. Lane 2 represents total proteins of cells harboring a control plasmid pBR322.

the crp mRNA level. The reduction of S i-resistant DNA in stem mutants was estimated to be about four-fold. On the other hand, all of the T tract mutations did not affect the

i w

j F--'--'--'7

WT
probe h=~

MO M1

M2

M5

M6

M7

M8

crp m R N A ~ .

..... . . . . . ~i!~

i~,~i!~!ii~ ~

Fig 5. Effects of terminator mutations on the expression of the crp mRNA. The pHA7 derivatives carrying the terminator variants were introduced in TP2339 cells. Total RNAs (30 lag) prepared from each transformant were subjected to S! assay as described in Materials and methods. Lane i is a DNA probe witheut S i treatment.

1040 levels of crp mRNA. The levels of csp mRNA were well correlated with CRP levels. The data clearly indicate that the G+C-rich stem-loop but not T tract sequences is responsible for the enhanced expression of upstream gene.
Effects of terminator mutations on mRNA stability

To test the effects of terminator mutations on the stability of the upstream mRNA, the decay rate of the crp mRNAs was determined. Cells harboring pHA7 derivatives were grown to exponential phase and rifamipicin was added to prevent further initiation of transcription. Cellular RNAs were isolated at various times after the addition of rifampicin and subjected to an El nuclease assay. As shown in figure 6, the mutations in the stem region significantly accelerated the degradation rate of crp mRN,,~. The radioactivity of the S l-resistant DNA band in each lane was quantified to determine the rate of mRNA degradation. The half-lives of crp mRNAs of M6 and M8 were estimated tc be about I min while the half-life of the wild-type c~7~ mRNA was about 3 min. The half-life of M5 crp mRNA was essentially the same with that of the wild-type. Similarly, all of the other T tract mutations had little effect on the stability of crp mRNA (data not shown). We conclude that the poly(T) tract which is required for efficient transcription termination plays no role in mRNA stabilization of the upstream gene.
Heterologous terminators can stabilize the ctp mRNA

characterized rho-independent terminators (oT)A, T7Te, and rrnBTl were prepared by DNA synthesis and inserted between the Mlul and Sinai sites of a variant of pHA7M0 which carries a M h d site near the termination codon. We examined how the heterologous terminators affect the crp mRNA levels by S I assay. As shown in figure 7, all three fusion genes expressed efficiently the upstream crp mRNA as did the wild-type crp gene, indicating that three terminators can act to stabilize the upstream mRNA. These results strongly suggest that rho-independent terminators generally can confer stability to the upstream mRNA. It should be noted that the~e is no simple correlation between the thermodynamic stability (AG) of the RNA secondary structure and the magnitude of enhancement in expression of the upstream gene.

Discussion
The primary role of rho-independent terminators is to specify the 3' end of transcription units or to attenuate transcriptic'a thereby controlling the level of transcription of downstream genes. Another less characterized function of rho-independent terminators is to stabilize the upstream RNA. In this paper, we have addressed a question of how two elements of rho-in(~ependent terminators contribute to transcription termination and to RNA stabilization of the upstream gene. For this, the effects of various mutations in the crp terminator on the transcription termination and on the stability of upstream clp mRNA have been investigated. Our data can be summarized as follows: I ) disruption of the dyad symmetry sequence of the crp terminator dramatically reduced termination eMciency; 2) conversion of the G+C,'ich to A+T-rich dyad symmetry sequences also eliminated

To examine the generality of the mRNA stabilization effect of rho-independent terminators, we constructed fusion genes in which the ctp terminator was replaced by other terminators. DNA fragments corresponding to three well-

WT Time (rain)

M5 Time (rain)

0 crpmRNA ,.~

4 --

8 crpmRNA ~ ,~

4 ~__

M6 Time (min) 0 2 4 8 arpmRNA ~ , ~ ~ 1 "" 1 0 0 ~ ...... tl/2=lmin

U8

Time (rain) 2 4 8
~ : ~ : ~2~

r ......

cmmnNA-- ~ L ~

...............

~ lOO~.~:3mln]

~ ,oo~=3mln

" IO0~il/2:lmin "~:

.= -02468' time(mini 1

- 'i1\
timelmin)
"02iiik

101 .N~
]o z 4 6 k

time(rain)

time(rain)

Fig 6. Stability of the crp mRNAs expressed in cells harboring pHA7 derivatives. TP2339 cells harboring plasmids were grown to OD600 = 0.8 and total RNA was extracted at the times indicated after the addition of rifampicin (200 lag/mL). Total RNA (50 lag) was subjected to S I assay. The DNA bands were quantified using Bioimage Analyzer BAS2000 (Fuji). Semilogarithmic plots of the radioactivity of S l-resistant DNA versus time are shown below the autoradiograms. The halt-life (t ]/:,) of the each mRNA was determined based on these plots.

!04 !

1 probe crp mRNA

i
WT M8 =2.4 trpA -14.8 rrnBT1 TTTe -29.5 -18.8

AG(kcai/mol)

-21.1

Fig 7. Effects of heterologous terminators on the expression of the cq~ mRNA. The pHA7 derivatives containing the opA, rrnBTl, and T7Te were introduced into TP2339 cells. Total RNAs (30 tag) prepared from each trans|k)rmant were subjected to S ! assay as described in Materials and methods. The free energies of the predicted terminator RNA hairpins, calculated by the method of Zuker and Stiegler [21 ], were shown below the lanes.

termination efficiency; 3) partial or complete disruption of the poly(T) sequence significantly but not completely reduced termination efficiency; 4) disruption of the G+C-rich dyad symmetry region markedly reduced expression of the upstream crp gene through its influence on mRNA stability; 5) disruption of the poly(T) sequence did not affect the stability of the upstream RNA; and 6) rho-independent terminators derived from other operons enhanced expression of the crp gene when they were placed downstream of the o7, structural gone. Based on these results, we conclude th~ two structural elements of the rho-independent terminato~ contribute differentially to transcription termination and to mRNA stabilization. Namely, the G+C-rich dyad symmetry sequence but not the poly(T) tract is essential for mRNA stabilization while both elements are required for efficient transcription termination. Examples are known in which dyad symmetry sequences that can form a stem-loop structure act to increase expression of an upstream gene by stabilizing mRNA. Evidence for this phenomenon have been most thoroughly documented with the repetitive e.~tragenic palindromic (REP) seq~:ences [22]. It is generally believed that stem-loop structures stabilize upstream mRNA by acting as a barrier against 3'-5' exonucleolytic attack [22]. Rho-independent terminators are probably the most widely distributed dyad symmetry sequences in the E coli genome. However, only a few rho-independent terminators have been actually shown to stabilize upstream RNA. The first demonstration that a rho-independent terminator may act to enhance expression of an upstream gone came from studies on the phage lambda int gene [13]. The int gene is expressed efficiently from a transcript that terminates at a rho-inde

pendent terminator whereas a second read-through transcript expresses very little int gene due to the rapid degradation of int mRNA triggered by the removal of the terminator hairpin by RNAase Ill. Subsequently, mRNA stabilization by rho-independent terminators was demonstrated in the E coli trp operon [I 2], the Bacillus thuringiensis co' gene [23], a bacteriophage ~X174 [24], and the E coli cq~ gene J 15 ]. We demonstrated here that three wellknown rho-independent terminators exhibit the ability to stabilize cq~ mRNA when they are placed downstream ol' the crp structural gone. The RNA stabilization effect of rhoindependent terminators does not appear to be directly correlated with the thermodynamic stability (AG) of the RNA hairpin. However, it is apparent that the stem-loop structure, which consisted of only A-T base pair'.~, has no ability to stabilize the upstream RNA. It is also known that there is no simple correlation between the stability of RNA secondary structure and the efficiency of transcription termination [ ! 0]. Our results strongly support the view that the mRNA stabiliz.-tion effect is an important general function of rho-independent terminators. An~:~ther implication ~f the present sludy is that :'~c contrJbu|ion of the G~-C-rich dyad symmetry sequence is more important than that of the poly(T) tract at the crp terminator although both elements are required for efficient termination. This finding is consistent ,,,ith the fact that the requirement of the number of T resid~aes varies depending on the terminator. For example, it is known tha.~ partial deletions of the poly(T) tract drastically reduce termination efficiency in some terminators while there are strong terminators with only a few T residues [10]. On the other hand, there is an example in which a stretch of A residues in the template

1042 strand without an RNA hairpin structure acts as an efficient terminator in vitro [251. A classical model concerning the mechanism of transcription termination postulates that the G+C-rich RNA hairpin first causes the RNA polymerase to pause in elongation and subsequently the U-rich 3'-tail stimulates release of the RNA transcript due to the low stability of the dA-rU hybrid [3, 8, 9]. However, recent studies have revealed that interactions between RNA and RNA polymerase play a major role in transcription termination [10, 26]. In particular, findings about structural rearrangements of elongating RNA polymerase resembling the movements of an inchworm [27-301 have led to an alternative model in which transcription termination is coupled with straining and relaxation of elongating RNA polymerase [i 11. This model assumes that formation of an RNA hairpin reduces the RNA-RNA polymerase interaction and the poly(T) tract acts as an inchworming signal. The different requirements of the poly(T) tract at different terminators can be easily explained by the inchworming model since it is known that sequences without a run of T residues can be used as inchworming signals [291. Although the structural features of these sequences are not defined yet, they could substitute for the poly(T) tract in some terminators. It is interesting to examine whether rho-independent terminators containing an inchworming signal other than the poly(T) tract naturally exist.
8 Platt T (1986) Transcription termination and the regulation of gene expression. Allnu Rev Biochem 55, 339-372 9 Yager TD, yon Hippel, PH {1987) Transcription elongation and termination in Escherichia colt. In: Escherichia colt and Sahnonelhl typhimuriun~: Celhdvr and Molecuho" Biology. (Neidhardt FC, Ingraham SL, Low KB, Magasanik B, Schaechter M, Umbarger HE, eds) American Society for Microbiology, Washington DC, Vol 2, 1241 p 10 Reynolds R, Bermudez-Cruz RM, Chamberlin MJ (1992) Parameters affecting transcription ternfination by Escherichia colt RNA polymerase: I. Analysis of 13 rho-independent terminators. J Mol Bio1224, 31-51 11 Nudler E, Kashlev M, Nikiforov V, Goldfarb A (1995) Coupling between transcription termination and RNA polymerase inchworming. Cell 81, 35 !-357 12 Mott JE, Galloway JL, Platt T (1985) Maturation of Escherichia colt tryptophan operon mRNAs: Evidence for 3' exonucleolytic processing after rho-independent termination. EMBO J 4, 1887-1891 13 Guarneros G, Montanez C, Hernandez T, Court D { 1082) Posttranscriptional control of bacteriophage ~, int gene expres fion from a site distal to the gene. Proc Natl 4cad Sci USA 79, 238-242 14 Aiba H, Fujimoto S, and Ozaki N (1982) Molecular cloning and nucleotide sequencing of the gene for E colt cAMP receptor protein. Nucleic Acids Res I O, 1345- ! 36 I 15 Aiba H, Hanamura A, Yamano H (1991) Transcription terminator is a positive regulatory element in the expression of the Escherichia colt crp gene. J Biol Chem 266, 172 I- ! 727 16 McKenney K, Schimatake H, Court D, Schmeissner U, Rosenberg M t 1981 ) A system to study promoter and terminator signals recognized by Escherichh~ colt RNA polymerase, hi: Gene Ampl~/?cation and Analysis. Analysis of Nucleic Acids Enzymatic Methods (Chirikjian JC, Papas, TS, eds) Elevier/North Holland Publishing Co, Amsterdam, Vol 2, 383 p 17 Campbell A ( 1961 ) Sensitive mutants of bacteriophage k. Virology 14, 22-32 18 Adhya S, Miller W (19791 Modulation of two promoters of the galactose operon of Escherichia colt. Nanoe 279, 492---494 19 Roy A, Haziza C, Danchin A (1983) Regulation of adenylate cyclase synthesis in Escherichia colt: nucleotide sequence of the control region. EMBO J 2, 791-797 2t) Mort K, Aiba H (1985) Evidence for negative control of t3,tl transcription by cAMP and cAMP receptor protein in intact Escherichia colt cells. ,I Biol Chem 260, 14838-14843 21 Zuker M, Stiegler P (1981) Optimal computer folding of large RNA sequences using thermodynamics and auxiliary information. Nucleic Acids Res 9, 133-148 22 Higgins CE Causton HC, Dance GSC, Mudd EA (1993) The role of the 3' end in mRNA stability and decay, h~: Conool of Messenger RNA Stabili~ (Bela~o J, Brawennan G, eds) Academic Press Inc, New York, 13 p 23 Wong HC. Hayashi M (1986) Identification of a positive retroregulator that stabilizes mRNAs in bacteria. Pioc NatlAcad Sci USA 83, 3233-3237 24 Hayashi MN, Hayasr, i M (1985) Cloned DNA sequences that determine mRNA stability of bacteriophage ~XI74 in vivo are functional. Nucleic Acids Res 13, 5937-5948 25 Tomizawa J, Masukata H (1987) Factor-independent termination of transcription in a stretch of deoxyadenosine residues in the template DNA. Cell 51,623-630 26 Reynolds R, Chamberlin MJ {1992) Parameters affecting transcription termination by Escherichia colt RNA polymerase: It. Construction and analysis of hybrid terminators. J Mol Biol 224, 53-63 27 Krummel B, Chambedin MJ (1992) Structural analysis of ternary complexes of Escherichia colt RNA polymerase: individual complexes halted along different transcription units have distinct and unexpected biochemical properties. J Mol Biol 225, 221-237 28 Krummel B, Chamberlin MJ (1992) Structural analysis of ternary complexes of Escherichia colt RNA polymerase: deoxyribonuclease I footprinting of defined complexes. J Mol Biol 225, 239-250 29 Nudler E, Goldfarb A, Kashlev M (1994) Discontinuous mechanism '~f transcription elongation. Science 265, 793-796 30 Wang D, Meier TI, Chan CL, Feng G, Lee DN, Landick R (1995) Discontinuous movements of DNA and RNA in RNA polymerase accompany formation of a paused transcription complex. Cell 81,341-350

Acknowledgment
This work was supported in part by Special Project Research Funds from the Ministry of Education, Science, Sports and Culture o f Japan.

References
I d'Aubenton Carafa Y, Brody E, Thermes C (!090) Pr,_'dictionof rhoindependent Escherichia colt transcription terminators: A statistical analysis of their RNA stem-loop structures. J Mol Biol 216, 835-858 2 Landick R, Turnbough CL, Yanofsky, C (1996) Transcription attenuation. hz: Escherichia colt and Sahnonelh~ typhimuriunc Celhdar and Molecular Bioloh:v (Neidhardt FC, Curtiss !11 R, Ingraham JL, Lin ECC, Low KB, Magasanik B, Reznikoff WS, Riley M, Schaechter M, Umbarger HE, eds) American Society for Microbiology, Washington DC, 2nd edn, vol I, 1263 p 3 Platt T (1981) Termination of transcription anta its regulation in the tryptophan operon of E colt. Cell 24, 10-23 4 Lynn SP, Kasper LM, Gardner JF (1988) Contributions of RNA secondary structure and length of the thymidine tract to transcription termination at the thr operon attenuator. J Biol Chem 263, 472-479 5 Cheng S-WC, Lynch EC, Leason KR, Court DL, Shapiro BA, Friedman D! (1991) Functional importance of sequence in the stem-loop of a transcription terminator. Science 254, 1205- i 207 6 Wilso~ KS, yon Hippel PH (1995) Transcription termination at intrinsic termi.'~ators: The role of the RNA hairpin. Ptoc Natl Acad Sci USA 92, 8793-8797 7 Yang M-T, Scott !I HB, Gardner JF (1995) Transcription termination at the thr attenuator: Evidence that the adenine residues upstream of the stem and loop structure are not required for termination. J Biol Chem 270, 23330-23336