Programmed Cell Death by hok/sok of Plasmid - Department of Plant ...

Programmed Cell Death by hok/sok of Plasmid R1: 

Processing at the hok mRNA 3 0 -end Triggers 

Structural Rearrangements that Allow Translation and 

Antisense RNA Binding 

Thomas Franch 1 , Alexander P. Gultyaev 2 and Kenn Gerdes 1 * 

1 Department of Molecular 

Biology, Odense University 

Campusvej 55, DK-5230 

Odense M, Denmark 

2 Leiden Institute of Chemistry 

Leiden University, P.O. Box 

9502, 2300 RA, Leiden 

The Netherlands 

*Corresponding author 

Introduction 

The hok/sok system of plasmid R1 codes for two 

RNAs, hok mRNA and Sok antisense RNA (Gerdes 

et al., 1988, 1990). The mRNA encodes two reading 

frames, denoted hok (host killing) and mok (modulation 

of killing), that overlap extensively (Figure 1). 

Translation of hok is coupled to translation of mok 

(Thisted & Gerdes, 1992). Sok-RNA represses 

translation of mok via binding to the translational 

initiation region (TIR) of mok (Thisted et al., 1994a). 

Therefore, Sok-RNA regulates hok translation 

Abbreviations used: SD, Shine-Dalgarno; ucb, 

upstream complementary box; dcb, downstream 

complementary box; tac, translational activator; TIR, 

translational initiation region; DMS, dimethyl sulphate; 

wt, wild-type; IPTG, isopropyl-b,Dthiogalactopyranoside. 

J. Mol. Biol. (1997) 273, 38±51 

The hok/sok locus of plasmid R1 mediates plasmid stabilization by killing 

of plasmid-free cells. The locus speci®es two RNAs, hok mRNA and Sok 

antisense RNA. The post-segregational killing mediated by hok/sok is governed 

by a complicated control mechanism that involves both post-transcriptional 

inhibition of translation by Sok-RNA and activation of hok 

translation by mRNA 3 0 processing. Sok-RNA inhibits translation of a 

reading frame (mok) that overlaps with hok, and translation of hok is 

coupled to translation of mok. In the inactive full-length hok mRNA, the 

translational activator element at the mRNA 5 0 -end (tac) is sequestered by 

the fold-back-inhibitory element located at the mRNA 3 0 -end (fbi). The 5 0 

to 3 0 pairing locks the RNA in an inert con®guration in which the SD mok 

and Sok-RNA target regions are sequestered. Here we show that the 3 0 

processing leads to major structural rearrangements in the mRNA 5 0 -end. 

The structure of the refolded RNA explains activation of translation and 

antisense RNA binding. The refolded RNA contains an antisense RNA 

target stem-loop that presents the target nucleotides in a single-stranded 

conformation. The stem of the target hairpin contains SD mok and AUG mok 

in a paired con®guration. Using toeprinting analysis, we show that this 

pairing keeps SD mok in an accessible con®guration. Furthermore, a mutational 

analysis shows that an internal loop in the target stem is prerequisite 

for ef®cient translation and antisense RNA binding. 

# 1997 Academic Press Limited 

Keywords: Sok; hok; antisense RNA; translational initiation; RNA 

rearrangements 

indirectly through mok. The 64 nt Sok-RNA consists 

of a stem±loop with an 11 nucleotide 5 0 

single-stranded extension (Figure 2A). The 5 0 extension 

is complementary to the part of the mok TIR 

that we denote sokT 0 (Figure 2) and is responsible 

for the initial recognition reaction between the 

RNAs (Thisted et al., 1994a). After initial recognition, 

more extensive duplex formation progresses 

from the Sok-RNA 5 0 -end by a zippering-like 

mechanism. Subsequently, such RNA duplexes are 

cleaved by RNase III (Gerdes et al., 1992). Thus, 

binding of Sok-RNA to hok mRNA occurs by a 

one-step binding mechanism similar to the RNA- 

IN/RNA-OUT pairing in Tn10 (Kittle et al., 1989; 

Case et al., 1989). 

The complex RNA metabolism prerequisite for 

activation of hok translation in plasmid-free cells is 

described in detail in the accompanying paper 

(Gultyaev et al., 1997) and is brie¯y outlined 

0022±2836/97/410038±14 $25.00/0/mb971294 # 1997 Academic Press Limited

Structure and Function of hok mRNA of Plasmid R1 39 

Figure 1. Overview of the structural elements of the hok/ 

sok system from plasmid R1. FL-1 and FL-2 denote the 

full-length hok mRNA-1 and mRNA-2, respectively. TR 

denotes the 3 0 truncated hok mRNA. Other abbreviations 

used: hok, host killing; mok, modulation of killing; Sok, 

suppression of killing;, sokT, Sok-RNA target; fbi, foldback-inhibition; 

and tac, translational activation. 

below. We showed previously that hok mRNA 

exists in two variants in vivo (Gerdes et al., 1990; 

Thisted et al., 1994a). A truncated, translationally 

active mRNA is generated by slow exonucleolytic 

removal of 40 nt from the 3 0 -end of the full-length 

mRNA. Thus, the full-length mRNA acts as a reservoir 

from which the active mRNA is continuously 

generated. In plasmid-carrying cells, the presence 

of Sok-RNA prevents translation of the truncated 

mRNA. However, in plasmid-free cells, in which 

Sok-RNA has decayed, the continuous slow processing 

of the full-length mRNA leads to accumulation 

of the translatable truncated mRNA. 

Eventually, this leads to synthesis of the toxic Hok 

protein, and killing of the plasmid-free cells 

ensues. 

Recently, we showed that the stable, inactive hok 

mRNA is compactedly folded with only a few 

single-stranded regions, apart from loops (Franch 

& Gerdes, 1996). A genetic analysis established 

that the full-length mRNA contains a peculiar 5 0 to 

3 0 long-range pairing in which the very ®rst 

nucleotides pair with the very last ones (Figure 2B). 

This interaction is prerequisite for the post-segregational 

killing mechanism, since its disruption (by 

mutation) leads to mRNA instability and depletion 

of the activatable pool of full-length mRNA 

(Franch & Gerdes, 1996). The 3 0 -element responsible 

for inhibition of translation and antisense 

RNA binding was called fbi (fold-back-inhibition). 

The 5 0 element that pairs with fbi was denoted tac 

(translational activation). The tac element is 

required for activation of translation (Franch & 

Gerdes, 1996). 

In full-length hok mRNA, the SD mok element is 

sequestered by an upstream sequence that we 

denote ucb (upstream complementary box). The 

ucb element is shown in red in Figure 2. Previously 

we suggested that pairing between SD mok and ucb 

prevents ribosome entry at SD mok and thereby inhibits 

translation of mok (and hok; Nielsen & Gerdes, 

1995). A nearly perfect repeat of the 9 nt ucb 

element is located 58 nt downstream of ucb 

(Figure 2). This element, which pairs with the 

translation initiation region of hok (TIR hok) in fulllength 

hok mRNA, was denoted dcb (downstream 

complementary box). This nomenclature is introduced 

to illustrate the complex structural 

rearrangements that occur in hok mRNA (see 

below). As a consequence of their repetitive nature, 

the ucb and dcb elements are both complementary 

to SD mok (compare Figure 2B and D). 

The structure of the antisense RNA target in hok 

mRNA is yet unknown. However, the genetic 

algorithm calculations and phylogenetic comparisons 

presented in the accompanying paper suggest 

that the exonucleolytic removal of fbi triggers 

major structural rearrangements at the mRNA 5 0 - 

end. More precisely, the computer-simulated RNA 

foldings predict that in truncated mRNA tac pairs 

with ucb, and SD mok with dcb. The latter interaction 

is involved in the formation of an antisense RNA 

target hairpin (Figure 2D). The target hairpin is 

topped by a 7 nt loop complementary to the very 

5 0 -end of Sok-RNA. 

Here, we investigate the biological properties 

and secondary structures of the full-length and 

truncated hok mRNAs. We show that the binding 

of Sok-RNA to full-length hok mRNA is negligible 

both in vivo and in vitro. Using in vitro toeprint 

analyses, we show that the mok TIR in full-length 

hok mRNA is unable to bind ribosomal 30 S subunits, 

in accordance with the observation that 

full-length hok mRNA is not translated in vivo. 

Most importantly, we obtain ®rm evidence for 

the postulated structural rearrangements triggered 

by 3 0 processing of the full-length mRNA. In the 

truncated mRNA, tac was found to pair with ucb, 

and SD mok with dcb, thus con®rming the presence 

of the tac and target hairpins. A mutational analysis 

shows that the structure of the target hairpin 

is crucial for rapid antisense RNA binding and 

for translation of mok, and that even modest 

changes in its structure have profound effects on 

these parameters. We further show that the pairing 

between SD mok and dcb is required to maintain 

truncated hok mRNA in a translatable 

con®guration. 

Results 

Sok antisense RNA binds to truncated but not 

to full-length hok mRNA in vivo 

We have previously examined the metabolism of 

the hok mRNAs in vivo (Gerdes et al., 1990, 1992; 

Thisted & Gerdes, 1992; Thisted et al., 1994a). Indirect 

evidence from Northern analyses suggested 

that Sok-RNA binds to truncated hok mRNA much 

more ef®ciently than to the full-length RNA. To 

test this inference more directly, we constructed a 

plasmid vector carrying a hok/sok system in which 

P hok was precisely replaced by the strong, LacI 

repressed P A1/04 promotor (Lanzer & Bujard, 1988). 

In this system, synthesis of hok mRNA is halted by 

the removal of isopropyl-b,D-thiogalactopyranoside 

(IPTG) from the growth-medium. Thus, synthesis

40 Structure and Function of hok mRNA of Plasmid R1 

of hok mRNA could be blocked without the 

addition of rifampicin. 

The in vivo processing pattern of the hok mRNAs 

was analysed by Northern blotting in the presence 

(sok ‡ ) or absence (sok ) of antisense RNA. In its 

presence, two hok mRNAs, denoted full-length 

mRNA-1 and mRNA-2, were observed (Figure 3A, 

®rst panel). Both mRNAs were quite stable with a 

half-life of approximately 20 minutes. Previous 

analyses using rifampicin to inhibit transcription 

yielded half-lives of the order of 20 to 40 minutes 

(Gerdes et al., 1990, 1992; Thisted et al., 1994a). hok 

Figure 2. Secondary structures of Sok-RNA and hok mRNAs. A, Secondary structure of Sok-RNA. B, Secondary structure 

of full-length hok mRNA. C, Con®guration of non-refolded truncated hok mRNA as predicted by computerassisted 

structure calculations. D, Putative truncated, refolded mRNA supported by the structure-probing analyses 

presented in Figure 4 and the RNase H cleavage pattern shown in Figure 5. Shine-Dalgarno and start codons of the 

mok and hok reading frames are shown in blue, and the ucb and dcb repetitive sequences are shown in red. C74 in 

ucb, marked in green, emphasizes a heterogeneity between the ucb and dcb elements. Thin numbered lines mark the 

nucleotides complementary to the three DNA oligonucleotides used in the RNase H cleavage assay. The antisense 

RNA target that is complementary to the 5 0 -tail of Sok-RNA is denoted sokT 0 (B). The sequence and structure of 

super-tac truncated RNA is shown below the wt tac/ucb pairing in D. The non-refolded (C) and refolded (D) con®gurations 

are proposed to exist in a dynamic equilibrium.


Figure 3. A, Effect of Sok antisense RNA on the stability 

and processing pattern of hok mRNAs as shown by 

Northern blotting. The ®rst panel shows hok mRNAs 

after arrest of hok transcription in the presence of Sok- 

RNA, the second panel shows hok mRNAs after arrest 

of transcription in the absence of Sok-RNA. Cells 

(CSH50 containing either pKG1540 (sok ‡ ) or pKG1541 

(sok )) were grown in LB medium containing 1 mM 

IPTG to an A 450 of ca 0.2, harvested, and resuspended in 

prewarmed medium without IPTG. Cell samples for 

total RNA preparation were taken at the intervals indicated. 

The mRNAs contained stop codon mutations in 

the hok gene to prevent detrimental hok expression (from 

Thisted & Gerdes, 1992). B, In vitro binding assay 

between 32 P-labelled Sok-RNA and unlabelled ( 3 H) truncated 

hok mRNA (®rst panel) and full-length hok 

mRNA-1 (second panel). Labelled antisense RNA was 

added to a tenfold excess of target RNA. For details, see 

Materials and Methods. 

mRNA-2 is generated from hok mRNA-1 by an as 

yet unknown mechanism (Nikolaj Dam Mikkelsen, 

unpublished results). The absence of Sok-RNA 

caused the appearance of a third band corresponding 

to truncated hok mRNA (Figure 3A, second 

panel). This RNA is generated by the exonucleolytic 

removal of 40 nt at the 3 0 -end of hok mRNA-2 

(Gerdes et al., 1990). Comparison of the two panels 

of Figure 3A yields valuable new information on 

the in vivo metabolism of hok mRNA: ®rst, the 

absence of Sok-RNA has no in¯uence on the 

amounts or the stabilities of the two full-length 

species. Since antisense RNA binding leads to 

duplex formation and rapid RNase III cleavage, 

this observation indicates that Sok-RNA does not 

bind to hok mRNA-1 or mRNA-2 at a signi®cant 

rate. Secondly, the absence of Sok-RNA resulted in 

the accumulation of the truncated mRNA. Since 

this RNA is absent from the ®rst panel, this observation 

indicates that Sok-RNA binds avidly to 

truncated hok mRNA and thereby confers its rapid 

decay via RNase III-mediated hydrolysis. Taken 

together, these results show that Sok-RNA binds 

full-length hok mRNA at a negligible rate in vivo, 

whereas the truncated version of hok mRNA is scavenged 

rapidly. 

Sok-RNA binds 100-fold more rapidly to 

truncated than to full-length hok mRNA 

in vitro 

The above results indicate that Sok-RNA binds 

much more rapidly to truncated than to full-length 

hok mRNA in vivo. Using a standard in vitro binding 

assay, we showed previously that full-length 

hok mRNA bound Sok-RNA with an approximately 

30% reduced rate (Thisted et al., 1994b). To 

resolve this apparent discrepancy, we repeated the 

in vitro binding assay using a more physiologically 

correct binding buffer (i.e. 200 mM potassium glutamate 

was included). The resulting gel-shift experiment 

is shown in Figure 3B. As seen, Sok-RNA 

bound readily to the truncated RNA (Figure 3A). 

However, binding to the full-length RNA was 

inhibited (Figure 3B). The apparent second-order 

binding-rate constants (K app) were estimated to 

be 3 10 5 M 1 s 1 for truncated and less than 

3 10 3 M 1 s 1 for full-length hok mRNA, corresponding 

to a more than 100-fold difference in 

binding-rates (for calculations, see Materials and 

Methods). This result suggests that the sokT 0 region 

exists in different con®gurations in the two RNAs 

(see Figure 2B and D). Now we can address this 

important question experimentally. 

Secondary structure analyses of fulllength, 

truncated and super-tac truncated 

hok mRNAs 

Mutational analyses, computer calculations and 

phylogenetic comparisons suggested that translational 

activation of hok mRNA involves a substantial 

structural rearrangement at the mRNA 5 0 -end, 

and that this rearrangement is triggered by the 3 0 

processing of full-length hok mRNA (Franch & 

Gerdes, 1996; Gultyaev et al., 1997, accompanying 

paper). More speci®cally, our analyses predicted 

that in truncated RNA, tac preferentially pairs with 

ucb thereby leading to formation of the extended 

tac-stem (see Figure 2D). Although a large internal 

loop destabilizes the tac/ucb interaction, it contains 

the possibility of non-canonical GA/AG base-pairing 

(Turner & Bevilacqua, 1993), which may add to 

the energy of the tac stem. Most importantly, 

the 5 0 refolding was accompanied by structural 

rearrangements further downstream in the molecule. 

Thus, a stem±loop that we coin the antisense


Figure 4. Secondary structure probing by DMS (®rst panel), RNase T 2 (second panel) and RNase V 1 (third panel) of 

full-length, truncated and super-tac truncated mRNA. The sequence and secondary structure of the super-tac mutation 

are shown in Figure 2D. Modi®cation/cleavage bands mentioned in the text are marked by arrowheads. 

RNA target hairpin was formed (Figure 2D and 

the accompanying paper, Gultyaev et al., 1997). 

Construction of a ``perfect'' tac stem by the introduction 

of the super-tac mutation (see Figure 2D) 

increased translation of the truncated hok mRNA 

twofold (Franch & Gerdes, 1996). The super-tac 

mutation greatly increases the stability of the tac 

stem. Consequently, the presence of super-tac predictably 

would lead to refolding of the entire 

population of hok mRNA molecules present in a 

pool of in vitro prefolded RNA. This inference is in 

accordance with the increased translation rate of 

super-tac truncated RNA. 

In order to verify the predicted secondary structure 

transitions, structure probing analyses were 

conducted on full-length, truncated and super-tac 

truncated RNAs (Figure 4). Chemical modi®cation 

by dimethyl sulphate (modi®es N1 of unpaired 

A > N3 of unpaired C), cleavage by RNase T 2 

(cleaves 3 0 of unpaired nucleotides) and RNase V 1 

(cleaves paired and stacked nucleotides) were 

detected by primer extension using the hok1 primer. 

For convenience, only the regulatory region 

covering nucleotides 70 to 160 is shown in 

Figure 4. However, the entire mRNA 5 0 -ends 

were also extensively analysed during this work 

(data not shown). 

Full-length hok mRNA 

In the full-length mRNA, the proposed position 

of the loop in the ucb/SD mok stem±loop structure


was supported by T 2 cuts in the region U89-U91 

and DMS modi®cation at C92. The DMS modi®cation 

at A81 supports the internal loop at the top 

of the stem. Furthermore, V 1 cuts at 110 to 112 

could support the double-stranded bottom stem. 

Three A-C mismatches (C74-A107, A78-C103 and 

A82-C100) were predicted in the ucb/SD mok stem. 

No signi®cant modi®cation of these nucleotides 

was observed, probably due to the formation of 

A-C reverse wobble pairings (Saenger, 1984). The 

modi®cations at C115 and A117, and to a lesser 

extent at C113 and C114, suggest that this region is 

at least partially single-stranded. The secondary 

structure probing of the full-length hok mRNA is 

consistent with the structure presented in Figure 2B 

in which sokT 0 is in a double-stranded conformation. 

Wild-type and super-tac truncated hok mRNAs 

The wild-type and super-tac truncated mRNAs 

exhibited nearly identical modi®cation/cleavage 

patterns and comparison with that of the fulllength 

RNA revealed important differences. Cleavage 

by T 2 at A78 and A81-A82 in the truncated 

mRNA suggests the presence of a single-stranded 

region not present in the full-length transcript. 

These cuts appeared enhanced in super-tac mRNA. 

In the SD mok region signi®cant DMS modi®cations 

were observed in super-tac mRNA compared to the 

weak probing of the wt truncated mRNA, consistent 

with incomplete 5 0 -end refolding of the latter 

species. However, DMS and T 2 probing at C120- 

U124 con®rmed the single-stranded loop of the target 

stem. In addition, V 1 cuts at C125, C126 and 

A127 are consistent with stacking of the CC loop 

dinucleotide on a stem adenine nucleotide. The 

super-tac mutation increased the T 2 cleavages in the 

loop region (C120 to U124), whereas it reduced 

cleavage at U117, G118 and A127-A in the stem. 

Thus, the super-tac mutation seems to stabilize the 

target hairpin, consistent with incomplete refolding 

of the wt truncated RNA. This conclusion was corroborated 

by the V 1 cuts at C139 in the dcb 

element. These cuts increase gradually in the order: 

full-length < truncated < super-tac truncated mRNA 

consistent with the proposed transition of the dcb/ 

TIR hok interaction in full-length RNA to the SD mok/ 

dcb interaction in truncated RNA. 

The secondary structure model of truncated hok 

mRNA shown in Figure 2D predicts that sokT 0 and 

the mok TIR are located within the same hairpin. 

The validity of our secondary structure model of 

the truncated mRNA was further investigated 

using an RNase H cleavage assay that tests the 

propensity of an RNA to hybridize with different 

DNA oligonucleotides (Zarrinkar & Williamson, 

1994). The DNA oligonucleotides used here are 

indicated in Figure 2. 

The cleavage pattern of full-length hok mRNA is 

shown in the ®rst panel of Figure 5. Oligo 1 

resulted in 6% cleavage only, and oligos 2 and 3 

both yielded negligible cleavage. This pattern is 

Figure 5. RNase H cleavage patterns of 32 P-labelled fulllength 

(®rst panel), truncated (second panel) and supertac 

truncated mRNA (third panel). The regions of complementarity 

to the oligonucleotides used are shown in 

Figure 2. Cleavage products are marked by arrowheads. 

consistent with a compact secondary structure, 

especially in the region of sokT 0 . In contrast, all 

three oligonucleotides yielded signi®cant RNase H 

cleavage of truncated hok mRNA (second panel). In 

this case, 38%, 14% and 50% cleavage was 

observed with oligonucleotides 1, 2 and 3, respectively. 

The third panel shows the cleavage pattern 

of super-tac truncated mRNA. Here, oligos 1, 2 and 

3 yielded 77%, 19% and 65% cleavage, respectively. 

These two latter sets of mapping data are consistent 

with the formation of the antisense target 

stem±loop presented in Figure 2D. In addition, 

oligo 3, which is complementary to sokT 0 , yielded 

more than a 100-fold difference in cleavage activity 

between full-length and truncated mRNA. This 

pattern is consistent with the difference in antisense 

RNA binding described above. Oligo 1, 

which is complementary to SD mok, showed an 

increased cleavage in super-tac truncated mRNA as 

compared to the wild-type, consistent with the 

observed twofold increase in translation (Franch & 

Gerdes, 1996). In addition, the increased cleavage 

of the super-tac mRNA with all three oligonucleotides 

corroborates the suggestion that the 

wild-type truncated mRNA in vitro exists in two 

con®gurations, a refolded and a non-refolded. This 

further suggests that the antisense RNA target 

stem±loop favoured by the tac/ucb and SD mok/dcb 

pairings may be in a dynamic equilibrium with the 

non-refolded structure shown in Figure 2C.


Figure 6. Primary and proposed local secondary structures of wild-type and mutated antisense RNA target hairpins. 

The SD mok and AUG mok sequences are marked by a black line and the dcb box by a hairline. Mutations are shown in 

bold. The energy of the local structures is given below each structure (calculated by MFOLD). The in vitro binding 

assays and translation data from Figures 7 and 8 are presented here in a quantitative form. 

The internal loop of the target hairpin is 

required for antisense RNA binding 

The loop of the target hairpin in truncated hok 

mRNA is complementary to the 5 0 single-stranded 

end of Sok-RNA (Figure 2A and D). This suggests 

that the stem±loop structure is crucial for optimal 

antisense RNA binding. Therefore, we tested the 

function of the internal loop in antisense RNA 

Figure 7. In vitro binding between truncated hok mRNA 

containing the tst mutations and 32 P-labelled Sok-RNA 

shown by gel-shift. The Sok-RNAs used to bind tst2 and 

tst1,2 target RNAs carried cognate mutations to maintain 

complete antisense/target complementarity. 

binding. Structure-closing mutations yielding a 

non-interrupted upper stem were introduced at 

both sides of the target stem (tst1 and tst2). The 

double tst1,2 mutation was constructed in order to 

regain the internal loop. The secondary structures 

of the mutated target RNAs are shown in Figure 6. 

Antisense binding assays were accomplished as 

described above using antisense RNAs with 5 0 - 

ends completely complementary to the mutated 

hok target RNAs. Typical binding assays are shown 

in Figure 7, and the relative binding-rates calculated 

from three independent experiments are 

given in Figure 6. The observed binding rates for 

the truncated mRNA carrying tst1 and tst2 were 

reduced approximately eightfold and 16-fold, 

respectively. The double mutation (tst1,2) completely 

restored the binding rate to that of the wildtype 

RNAs. These results show (i) that interrupted 

helicity in the target stem is required for optimal 

antisense RNA binding and (ii) that the postulated 

secondary structure of the target hairpin is valid. 

The internal loop of the target hairpin is 

required for translation of hok (and mok) 

The effect of the tst mutations on hok expression 

was tested by in vitro translation of truncated 

mRNAs. Note, that this assay detects translation of 

hok, not mok. As shown in Figure 8A, the tst1 and


Figure 8. A and B, SDS-PAGE of in vitro translation 

reactions of truncated hok mRNA. Relative translation 

ef®ciencies are normalized to LacZ(a) and b-lactamase 

internal standards, and their quanti®cations are given in 

Figure 6. 

tst2 mutations abolished hok translation. The tst1,2 

mutation partially restored translation (to 9% of 

that of the wild-type). The tst1,2 mutation increases 

the G value of the hairpin by 3.4 kcal/mol 

because of the G-U base-pairing partially closing 

the internal loop (Figure 6). Therefore, an 

additional mutation containing an inverted internal 

loop sequence was constructed and tested (tst-inv). 

This mutation reduced the hok translation rate to 

17% (Figure 8A). These results are consistent with 

the notion that the internal loop is responsible for a 

suf®ciently low hairpin stability as to allow loading 

of ribosomes at the mok TIR. 

Mutations in the downstream complementary 

box (dcb) reduce hok expression 

In the target hairpin, the SD mok element is proposed 

to base-pair with the downstream complementary 

box, dcb (Figures 2D and 6). This 

sequestration was puzzling, since it is known that 

secondary structures tend to reduce or even prevent 

translation (de Smit & van Duin, 1990a,b), 

and hok mRNA is translated at a signi®cant rate 

(Figure 8). To investigate this apparent discre- 

pancy, mutations that reduce the pairing to SD mok 

were introduced into the dcb element (Figure 6). 

Surprisingly, both dcb mutations resulted in a 

severe reduction of hok translation (Figure 8A). 

Thus, the SD mok/dcb interaction appears to be 

required to maintain the wild-type translation rate 

of hok mRNA. 

A straightforward explanation for this unexpected 

result could be that the SD mok/dcb interaction 

is required for 5 0 refolding and therefore for 

translational activation. Hence, weakening of the 

SD mok/dcb interaction would favour the nonrefolded 

ucb/SD mok con®guration, thus explaining 

the effect of the dcb mutations on translation. To 

test this inference, the super-tac mutation was 

combined with the dcb1 and dcb2 mutations and 

the respective RNAs subjected to translation 

(Figure 8B). The tst1,2 and tst-inv mutations were 

also combined with super-tac and translation of the 

respective RNAs included for comparison. The 

relative translation-rates of tst1,2 and tst-inv were 

not signi®cantly increased by the super-tac refolding 

mutation. This was expected, since the tst1,2 

and tst-inv mutations stabilize the target stem (see 

Figure 6). In contrast, the super-tac mutation 

resulted in a three- to sixfold increase in hok translation 

of RNAs carrying the dcb1 and dcb2 

mutations (Figure 6). Thus, the super-tac mutation 

could, at least partially, reverse the translational 

defect conferred by the dcb mutations. However, 

the super-tac dcb mRNAs were still translated at a 

rate signi®cantly lower than that of the wild-type 

truncated RNA (Figure 8B). 

Toeprinting of the mok TIR in full-length and 

truncated hok mRNAs 

To investigate the accessibility of SD mok more 

directly, we employed the toeprinting technique. 

The toeprint assay semi-quantitatively monitors 

the binding af®nity between a TIR and the ribosomal 

30 S subunit (Hartz et al., 1988). Our toeprint 

analysis of the mok TIR is shown in Figure 9. In 

truncated hok mRNA, a weak, but clearly discernible 

band corresponding to a reverse transcriptase 

stop at ‡15 relative to the mok AUG start-codon 

was detected (lane 4). The position of this toeprint 

is consistent with a ternary complex formed at the 

mok AUG codon. We were not able to detect a 

similar toeprint in full-length hok mRNA (lane 2). 

This correlates well with the lack of translation of 

full-length hok mRNA in vivo and in vitro (Gerdes 

et al., 1990; Thisted et al., 1994a). Furthermore, the 

weak mok toeprint in truncated hok mRNA 

suggests that it is translated relatively infrequently, 

consistent with the pairing of the mok 

TIR to the dcb element in the target RNA hairpin. 

Similar results were obtained by the use of 

another oligonucleotide as primer in the toeprinting 

assay, indicating that the quantitative aspects 

described here are independent of the actual 

experimental setup.


Figure 9. In vitro toeprinting analysis of full-length and truncated hok mRNAs. The hok1 primer was used in all reactions 

(see Materials and Methods). Reverse transcription stops (toeprints) on wild-type and mutated hok mRNAs at 

AUG mok ‡ 15 are marked by squares. Lanes ‡ contain mRNA incubated with tRNA fMet and 30 S subunits. Lanes 

are control reactions without tRNA fMet and 30 S subunits. The type of RNA used in each toeprint is indicated for each 

pair of lanes. All mRNAs used are truncated hok mRNA derivatives, except for lanes 1 and 2. The tst and dcb 

mutations are shown in Figure 6, and super-tac in Figure 2. 

The accessibility of the mok TIR determines 

the rate of hok translation 

To investigate whether the translatability of hok 

mRNA and toeprinting at the mok TIR appear to 

be correlated, we performed toeprinting analyses 

of mutated hok mRNAs. As seen from lanes 6 

and 8 of Figure 9, the presence of the tst1 and 

tst2 structure-closing mutations completely abolished 

toeprinting at the mok TIR, consistent with 

the lack of translation of these RNAs (Figure 6). 

The tst1,2 and tst-inv mRNAs yielded reduced 

toeprinting signals (lanes 10 and 12), again consistent 

with the degree of translation of the 

mRNAs. The mRNAs containing the dcb1 and 

dcb2 mutations yielded slightly reduced toeprint 

signals (lanes 14 and 16) and their translation 

rates were also clearly reduced (Figure 6). Furthermore, 

the super-tac mutation that previously 

was shown to enhance translation of hok by twofold, 

yielded a threefold increase in the mok toeprinting 

signal (lane 18). These results indicate 

that the strength of the toeprinting signal from 

the mok TIR and the hok translation rate are correlated. 

We next asked if the combination of the 

super-tac mutation and the tst1,2 and tst-inv 

mutations would enhance the toeprinting signal 

(lanes 20 and 22, respectively). As seen, the 

super-tac mutation in these cases did not lead to 

an increased toeprinting signal. This is consistent 

with the lack of enhanced hok translation by 

super-tac in these two cases (Figure 6). Thus, all 

our observations indicate that within limits, the 

rate of ribosome loading at the mok TIR determines 

the rate of hok translation. 

The SD mok/dcb interaction is required for 

loading of ribosomes at the mok TIR 

Even though super-tac to a signi®cant degree 

reverted the negative effect of the dcb mutations, 

the double super-tac dcb mutants were not translated 

as ef®ciently as the wild-type hok mRNA 

(Figure 6). One explanation could be that in the 

double super-tac dcb mutant mRNA mok translation 

is so intense that it reduces translation of hok. 

Examples of occluding overlapping translation 

have been described (Berkhout et al., 1985). In the 

latter case, toeprinting at the mok TIR yields much 

more direct information than the hok translationrate. 

As seen from Figure 9, the super-tac mutation 

enhanced the toeprinting signals of both dcb 

mutant mRNAs more than tenfold (compare lanes 

14 and 16 with lanes 24 and 26), and these RNAs 

clearly yielded the strongest signals of all the 

mutants. These results are most readily explained 

by the assumption that super-tac forced refolding of 

the mRNA into the SD mok/dcb con®guration, combined 

with the reduced base-pairing between 

SD mok and dcb together cause the severe increase in 

the toeprinting signals of the doubly mutated 

mRNAs. Taken together, our results show that the 

antisense RNA hairpin is ®nely tuned to its two 

functions, translation and antisense RNA binding, 

since virtually any change in its structure has detrimental 

effects. 

Discussion 

Here, we obtain direct evidence for the refolding 

model proposed in the accompanying paper


(Gultyaev et al., 1997). We predict therein that the 

3 0 processing of the inactive full-length hok mRNA 

triggers major structural rearrangements at its 5 0 - 

end that activate translation and antisense RNA 

binding. Accordingly, we show here that the two 

versions of hok mRNA have highly different properties 

with respect to translation and antisense 

RNA binding, and subsequently correlate the biological 

properties of the molecules with their foldings. 

Finally, we dissect the antisense RNA target 

hairpin and show that this structure is required for 

optimal antisense RNA binding and for proper 

translation. 

Full-length hok mRNA bound Sok-RNA at a negligible 

rate in vivo (Figure 3A) and in vitro 

(Figure 3B). These results are consistent with the 

secondary structure model of the RNA as shown 

in Figure 2B: the Sok-RNA target region (sokT 0 ) is 

base-paired with the 3 0 -end of the RNA and is 

therefore unable to react with the 5 0 -end of the 

antisense RNA. Furthermore, the fbi element 

located at the very 3 0 -end of the mRNA pairs with 

the 5 0 tac element. This pairing prevents tac stem 

formation and forces the ucb element to pair with 

SD mok. By toeprinting analyses (Figure 9), we show 

that this pairing prevented access of the ribosomes 

at SD mok, which is consistent with the lack of translation 

of full-length hok mRNA (Thisted et al., 

1994a). 

In contrast, truncated hok mRNA bound Sok- 

RNA avidly (Figure 3) and was translated 

ef®ciently (Figure 6). Truncated hok mRNA is generated 

by slow continuous 3 0 processing of the 

full-length mRNA (Gerdes et al., 1990; Thisted et al., 

1994a) in both plasmid-carrying and plasmid-free 

cells. In the former case, the presence of Sok-RNA 

prevents translation of truncated hok mRNA and 

mediates its rapid decay via RNase III-mediated 

hydrolysis (Figure 3A, ®rst panel). Plasmid-free 

cells, however, do not contain antisense RNA, and 

truncated hok mRNA therefore accumulates 

(Figure 3A, second panel). We have shown previously 

that the 3 0 processing leads to Hok protein 

synthesis and killing of the plasmid-free cells 

(Gerdes et al., 1990; Thisted et al., 1994a). Thus, the 

post-segregational killing mechanism is dependent 

on the formation of truncated hok mRNA. 

In the accompanying paper we suggest that the 

exonucleolytic removal of the hok mRNA 3 0 -end 

triggers refolding of the 5 0 -end with formation of 

the tac stem and the antisense RNA target hairpin. 

Figure 2 shows the structural rearrangements proposed 

to activate both translation and antisense 

RNA binding. In truncated RNA, tac now prefers 

to pair with the ucb element (Figure 2D). This interaction 

favours the formation of the antisense RNA 

target hairpin. In this hairpin, SD mok pairs with the 

dcb element (Figure 2D). The AUG mok start codon is 

also paired, but an internal loop is located between 

the start codon and SD mok. The 5 0 -end of Sok-RNA 

is complementary to six of the seven nucleotides in 

the loop of the target hairpin. We have shown 

recently that these nucleotides are involved in the 

initial recognition reaction between the antisense 

RNA and the mRNA, consistent with their singlestranded 

conformation (unpublished results). The 

analysis of the antisense RNA recognition reaction 

will be presented elsewhere. 

To obtain direct evidence for the proposed structural 

rearrangements, we accomplished secondary 

structure analyses (Figure 4) and RNase H cleavage 

mapping (Figure 5) of the two RNAs. Data 

obtained from both types of experiments are largely 

consistent with the structure models shown in 

Figure 2B and D. Clearly, the introduction of the 

super-tac mutation (Figure 2D) into truncated hok 

mRNA changed the cleavage pattern further 

downstream in the molecule (Figure 4). The altered 

cleavage pattern is best explained by the assumption 

that wild-type truncated RNA exists in two 

con®gurations, a refolded and a non-refolded 

form, and that the super-tac mutation forces refolding 

of the entire population of molecules. This 

interpretation is consistent with our translation 

and toeprinting data (see below). 

Phylogeny and genetic algorithm calculations 

support the suggestion that all the hok-homologous 

gene systems form antisense RNA target hairpins 

of similar sizes and energies (see the accompanying 

paper, Gultyaev et al., 1997). Therefore, we investigated 

the function of the target hairpin with 

respect to translation and antisense RNA binding. 

To this end, we introduced a number of speci®c 

mutations in the stem and assayed their effects 

(summarized in Figure 6). First, the stem-closing 

mutations (tst1 and tst2) completely abolished 

translation and severely impeded antisense RNA 

binding. The double tst1,2 re-opening mutation 

fully restored antisense RNA binding, and translation 

was partially regained. These results show 

that the internal loop is required for the proper 

function of the target hairpin. The tst-inv mutation, 

in which the nucleotides of the internal loop were 

inverted, also exhibited a reduced translation rate 

(Figure 6). The latter two mutations show that the 

wild-type stem is accurately tuned to its particular 

translation rate, since even modest changes in its 

stability profoundly in¯uenced the translation rate. 

In the RNAs described until now, the ( G values 

of the target hairpin and the translation-rates are 

inversely correlated with the stability of the hairpin 

(Figure 6). 

The pairing of SD mok with dcb was puzzling, 

since sequestration of SD elements is known to 

reduce or prevent translation (de Smit & van Duin, 

1990a,b, 1994). However, phylogeny and structureprobing 

data strongly support the postulated interaction. 

Point mutations in dcb that reduced the 

SD mok/dcb interaction therefore should lead to an 

increased translation rate. However, the dcb1 and 

dcb2 mutations both exhibited severely decreased 

rates of hok translation (eightfold and 16-fold, 

respectively; see Figure 6). To some extent this 

result was surprising and indicated that the point 

mutations perhaps shift the mok TIR into an 

alternative, non-translatable con®guration. One


straight-forward possibility would be that the dcb 

mutations favour the ucb/SD mok interaction present 

in the full-length RNA. As shown by toeprinting, 

this RNA con®guration prevents access of ribosomes 

at SD mok (Figure 9). To test this assumption, 

we combined the super-tac mutation with the dcb 

mutations, since super-tac predictably would shift 

the equilibrium towards the SD mok/dcb interaction 

and thereby stimulate ribosome-binding at the mok 

TIR and thus, translation of hok. The super-tac 

mutation was also combined with the tst1,2, and 

tst-inv mutations. The super-tac mutation stimulated 

hok translation of the wild-type mRNA twofold, 

consistent with the assumption that super-tac 

forces refolding of the entire population of molecules 

(Figure 6). In contrast, super-tac did not 

stimulate translation of RNAs containing tst1,2 and 

tst-inv. The latter result shows that translation of 

these RNAs is refractory to stimulation by supertac. 

The target hairpins containing tst1,2 and tst-inv 

are more energy-rich than the wt hairpin (Figure 6). 

Thus, the lack of stimulation by super-tac suggests 

that these RNAs are already fully shifted towards 

the refolded, active con®guration. 

In the cases of dcb1 and dcb2, super-tac stimulated 

translation threefold and sixfold, respectively (i.e. 

the stimulation was signi®cant). This result is consistent 

with the proposed structural change conferred 

by the dcb mutations. However, even 

though super-tac stimulated hok translation of the 

dcb mutant RNAs, the translation rates of the doubly 

mutated RNAs never exceeded that of the 

wild-type RNA. This was surprising, since supertac 

forces formation of the translatable con- 

®guration and the dcb mutations reduce the 

sequestration of SD mok. However, these observations 

are indirect, since we change the environment 

of the mok TIR and measure hok translation. 

This is, of course, a technical caveat. 

To gain more direct insight into the status of the 

mok TIR, we turned to the toeprinting technique 

(Figure 9). Most importantly, we found that in all 

but two cases there was a correlation between the 

in vitro translation rates of hok and the strength of 

the corresponding SD mok toeprinting signal. In the 

cases exempt, dcb1 and dcb2, respectively, super-tac 

severely enhanced the toeprinting signals relative 

to the translation rates. These results show that the 

dcb mutations lead, as expected, to a highly accessible 

SD mok element when combined with super-tac. 

The relatively slow translation rates from the 

double super-tac dcb mutant RNAs may be caused 

by the highly increased mok translation rate, which 

inhibits hok translation (Berkhout et al., 1985). This 

inference is consistent with the presence of unexpected 

Mok protein bands appearing in the translation 

reactions of these RNAs (data not shown). 

The results discussed above suggest that the 

SD mok/dcb interaction prevents the RNA from folding 

into the non-translatable con®guration. Only 

when the non-translatable con®guration is disfavoured 

by other structural mutations (i.e. supertac), 

the dcb element appears, as expected, to have 

a negative effect on translation. These assumptions 

are supported by foldings of the mutated RNAs 

using the genetic algorithm (not shown). Thus, the 

toeprinting data explain the somewhat surprising 

observation that an anti-SD element (dcb) is 

required to maintain an SD element in a translatable 

con®guration. Our results are still compatible 

with and actually support the paradigm that secondary 

structure interactions at SD elements certainly 

reduce translation (de Smit & van Duin, 

1990a, 1994). 

In bacteria, exonucleolytic 3 0 processing is a 

major pathway of mRNA inactivation and decay 

(Higgins et al., 1988; Belasco & Higgins, 1988). We 

believe that the hok and hok-homologous mRNAs 

represent a unique class of mRNAs that are activated 

by 3 0 processing and we know of no other 

example in which an RNA is activated by 5 0 refolding 

triggered by the removal of its 3 0 -end. However, 

several other mRNAs contain alternative 

competing structures involved in translational control. 

The rpsO mRNA encoding the ribosomal protein 

S15 and the exoribonuclease PNPase is 

autoregulated by S15 that binds to the rpsO TIR. 

The rpsO mRNA leader region exists as a translational-competent 

pseudoknot structure in equilibrium 

with a double hairpin structure occluding 

SD rpsO (Phillipe et al., 1990). Interestingly, the S15 

protein binds and sequesters the pseudoknot structure 

and thereby inhibits initiation of translation 

(Phillipe et al., 1993). Also, the leader region of cIII 

mRNA from phage l adopts two alternative secondary 

structures with almost equal energies but 

different translational capacities (Altuvia et al., 

1989). Here, the RNase III protein is believed to 

trap, without cleavage, the active mRNA con®guration 

(Altuvia et al., 1991). 

Many antisense RNAs contain upper-stem mismatches 

in order to accommodate stable antisense 

RNA/target interaction (Hjalt & Wagner, 1995; 

Kittle et al., 1989; Wilson et al., 1997). The structure 

of the regulatory hairpin in hok mRNA contains a 

four nucleotide internal loop separated from the 

seven nucleotide loop by four base-pairs. The 

stem-closing mutations tst1 and tst2 severely 

impaired Sok-RNA binding. A similar effect was 

observed for stem-closing mutations in both CopA 

of the R1 plasmid replication system (Hjalt & 

Wagner, 1995) and the RNA-OUT of Tn10 (Kittle 

et al., 1989). In the hok/sok system, restoring the 

internal loop by the tst1,2 mutation regained binding 

kinetics, thus indicating the requirement of 

helix imperfection for ef®cient Sok-RNA/target 

pairing in vitro. Most likely, the initial binding of 

the 5 0 -leader of Sok-RNA to the complementary six 

nucleotides in the target loop creates the ®rst 

unstable binding intermediate. Our results suggest 

that subsequent progression to more complete 

duplex formation requires interrupted helicity to 

favour intrastrand opening. 

In conclusion, we show here that the processing 

at the hok mRNA 3 0 -end confers dynamic 

rearrangements at its 5 0 -end that mediate activation


of translation and antisense binding. Phylogeny 

and genetic algorithm calculations indicate that the 

other hok-homologous mRNAs exhibit similar complex 

rearrangements (Gultyaev et al., 1997). We 

believe that this is a unique example of a group of 

phylogenetically related mRNAs that are activated 

by structural rearrangements triggered by 3 0 processing. 

Materials and Methods 

Enzymes and chemicals 

Antibiotics and chemicals was added at the following 

concentrations: ampicillin, 100 mg/ml; IPTG, 1 mM. All 

enzymes were purchased from Boehringer Mannheim 

unless stated otherwise. 

Bacterial strains 

The Escherichia coli K-12 strain CSH50 [ (lac pro) rpsL] 

(Miller, 1972) was used in the Northern transfer analysis. 

Plasmids 

The plasmids used and constructed are as follows: 

plasmids pKG1540 and pKG1541 are pGEM3 (Promega) 

derivatives encoding the bla and lacI q genes. The plasmids 

carry a mutant 450 bp hok/sok system cloned into 

the EcoRI-BamHI sites of the polylinker. Plasmid 

pKG1540 is constructed by PCR on a template carrying a 

stop codon in hok (Thisted et al., 1992) using the oligos (i) 

and (ii). Oligo (i) is 5 0 -CGGATCCAAAAAGAGTG- 

TTGACTTGTGAGCGGATAACAATGATACTTAGAT-T- 

CGGCGCTTGAGGCTTTCTGCCTCATG, italics mark 

the BamHI restriction site and the underlined region is 

complementary to ‡131 to ‡ 156 in the hok/sok locus. 

This oligo contains the sequence for the strongly LacI 

repressed P A1/04 promoter. Oligo (ii) is 5 0 -CCCCGAATTC 

ACAACATCAGCAAGGAGAAA, italics mark the EcoRI 

restriction site and the underlined region is complementary 

to ‡580 to ‡561 in the 3 0 -end of the hok/sok system. 

Plasmid pKG1541 was constructed accordingly by PCR 

with the same oligos on a template containing a double 

mutation in the 10 sequence of the Sok promoter in 

addition to the hok mutation described above (Thisted & 

Gerdes, 1992). 

Total RNA preparation for Northern transfer analysis 

Preparation of total RNA from E. coli and Northern 

analysis were performed as described (Gerdes et al., 

1990). A culture carrying a high copy number plasmid 

(pUC derivative; Yanisch-Perron et al., 1985) containing 

the P A1/04::hok fusion and either the sok ‡ or sok genotype 

was grown exponentially at 37 C. At A 450 0.2, 

1 mM IPTG was added to induce hok transcription. After 

one hour the cells were pelleted and resuspended in prewarmed 

medium without IPTG. Samples were collected 

at the time-points indicated. In order to determine the 

precise half-life for the hok mRNAs, the dilution of 

mRNA after IPTG withdrawal (by cell growth) was compensated 

for by increasing the amount of total RNA 

loaded onto the gel as determined from the A 450 

measurements. The 32 P-labelled RNA probe used was 

generated by phage T7 RNA polymerase and the pGEMbased 

plasmid pGEM342 (hok probe; Thisted & Gerdes, 

1992). Calculation of mRNA half-lives was accomplished 

by evaluation of band intensity using an LKB Ultroscan 

laser densitometer and the GelScan XL software package 

(version 2.1) provided by Pharmacia. 

Site-directed mutagenesis 

The mutation shown in Figure 6 was introduced in 

the hok/sok system by double PCR as described by 

Nielsen & Gerdes (1995). PCR was performed using the 

pBR322 vector carrying the 580 bp wild-type hok/sok system, 

inserted in the EcoRI - BamHI sites, as template and 

the pBR322 EcoRI CW and pBR322 BamHI CCW external 

primers. Mutant primers are: tst1, 5 0 -GGACTAGACA- 

TAGGGGATGC (TF17) and 5 0 -GAGGCATCCCTATG- 

TCTAGTCCACATAGGGATAGCCTCTTACC (TF17b); 

tst1,2, 5 0 -ATCCTGATGTGGACTAGACATCAGGATG- 

CCTC (TF17-5) and 5 0 -GAGGCATCCTGATGTCTAGTC- 

CACATAGGGATAGCCTCTTACC, tst-inv, 5 0 - GGCA- 

TCCACATGTCTAGTCCACATTCGGATAGCCTCTTA- 

CC (TF20a) and 5 0 -GCTATCCGAATGTGGACTAGA- 

CATGTGGATGCCTCGTGGTGG (TF20b); dcb1, 5 0 - 

CACATCAGGATAGGCTCTTACCGCGC-3 0 (TF25a) and 

5 0 -GCGCGGTAAGAGCCTATCCTGATGTG-3 0 (TF24b); 

dcb2, 5 0 - CACATCAGGATAGCGTCTTACCGCGC 

(TF25a) and 5 0 -GCGCGGTAAGACGCTATCCTGATGTG 

(TF25b). The tst2 mutation was introduced by double 

template PCR as described (Franch & Gerdes, 1996) 

using the primer 5 0 -ATCCCTATGTGGACTAGACATA 

(TF17-4). Mutant hok/sok fragments were all cloned in the 

EcoRI-BamHI restriction-sites in pUC19 or mini-R1 

derivatives. The plasmids carrying a mutant hok/sok system 

were used for PCR to generate templates for in vitro 

synthesis of 32 P and 3 H-labelled truncated mRNA. These 

transcripts were used for in vitro experiments. 

Synthesis of hok mRNAs in vitro 

Large-scale wild-type and mutant in vitro hok transcripts 

with different lengths were synthesized using T7 

RNA polymerase and DNA templates generated by PCR 

according to Franch & Gerdes (1996). 

Secondary structure probing using DMS, T 2 and V 1 

DMS, T 2 and V 1 structure probings were conducted 

according to Thisted et al. (1995), with the following 

exceptions. The buffer used for DMS modi®cation was 

as follows; 20 mM Hepes (pH 7.8), 100 mM NH 4Cl, 

10 mM magnesium acetate, 1 mM DTT and 10% (v/v) 

glycerol. RNase T 2 and V 1 cleavage reactions were carried 

out in TMK-glutamate buffer (20 mM Tris-acetate 

(pH 7.5), 10 mM magnesium acetate, 200 mM potassium 

glutamate). 

Secondary structure probing by RNase H cleavage 

Uniformly 32 P-labelled full-length or truncated hok 

mRNA (20 fmol) was incubated in TMK-glutamate buffer 

(see above) supplemented with 1 M DTT, for three 

minutes at 37 C to equilibrate RNA folding. DNA oligo 

(10 pmol) and one unit of RNase H (Gibco) were added 

to give a total reaction volume of 10 ml. The reaction was 

incubated at 30 C for an additional 30 minutes and 

stopped by addition of Formamid Dye (FD) and withdrawal 

to 0 C. In the control reaction the oligo was 

omitted. Samples were heated at 80 C for three minutes 

prior to separation on a 5.5% polyacrylamide gel con-


taining 8 M urea. The oligonucleotides used were: oligo 

1 (SD), 5 0 -CCTCGTGGTG; oligo 2 (stem), 5 0 -CATAGG- 

GATG; oligo 3 (loop), 5 0 -GTGGACTAGAC. 

in vitro binding reactions between hok mRNA 

and Sok-RNA 

All binding reactions were performed according 

to Thisted et al. (1994b), except that the hok mRNA 

and Sok-RNA concentrations were 3 10 8 M and 

3 10 9 M, respectively. Also, the buffer used was TMK 

(see above). Band intensity quanti®cation was done by 

Phosphor Imager and the Imagequant software package 

(Molecular Dynamics). 

In vitro translation 

The E. coli coupled transcription/translation system 

(Zubay, 1973) was purchased from Promega. The translation 

reactions contained the following in a total volume 

of 20 ml: 3 pmol of 3 H-labelled truncated hok mRNA in 

5 ml of TE, 1 mg of pUC19 plasmid (internal standard), 

0.5 mM each of the 20 amino acids minus methionine, 

0.2 mM L-[ 35 S]methionine (1000 Ci/mmol; NEN), 2 mM 

ATP, 0.5 mM each of GTP, UTP and CTP, 210 mM potassium 

glutamate, 20 mM phosphoenolpyruvate, 35 mM 

Tris-acetate (pH 8.0), 27 mM ammonium acetate, 1 mM 

cAMP, 20 g/ml folinic acid, 100 mg/ml E. coli tRNAs, 

9 mM magnesium acetate, 0.8 mM IPTG, 2 mM DTT, 

35 mg/ml PEG 8000 and 6 ml of E. coli S30 extract. The 

reactions were allowed to proceed for 30 minutes. The 

samples were precipitated with four volumes of acetone 

at 0 C, dried and redissolved in 2 SDS sample buffer 

(per 10 ml: 2 ml of glycerol, 2 ml of 10% (w/v) SDS, 

0.25 mg of bromophenol blue, 2.5 ml of 0.5 mM Tris-HCl 

(pH 6.8), 0.4% SDS, 0.5 ml of b-mercaptoethanol) and 

denatured at 100 C for ®ve minutes prior to loading 

onto SDS/polyacrylamide gels according to Thisted et al. 

(1994a). 

Toeprinting analyses 

The 30 S ribosomal subunits were puri®ed as follows: 

100 ml of mid-log culture of the E. coli strain MRE600 

(RNase I-de®cient) was snap-cooled and cells were harvested. 

The cell pellet was resuspended in 1 ml of ribosome 

buffer (50 mM Tris-HCl (pH 7.5), 1 mM MgCl 2, 

100 mM NH 4Cl). Cell membrane fracturing was accomplished 

by three cycles of snap-freezing and thawing (on 

ice). The suspension was pelleted and the supernatant 

containing uncoupled ribosome subunits was ultracentrifuged 

(18,000 rpm, SW28 buckets) in a 5% to 20% (w/v) 

sucrose gradient in ribosome buffer at 4 C for 16 hours. 

The 30 S ribosome fractions were collected and pelleted 

by ultracentrifugation. The 30 S ribosome subunits were 

dissolved in ribosome buffer containing 10 mM MgCl 2. 

The toeprinting analysis was conducted as follows: 

0.4 pmol of 5 0 -end labelled hok1 primer was annealed to 

0.4 pmol of hok mRNA in a buffer containing 60 mM 

NH 4Cl, 10 mM Tris-acetate (pH 7.5), 8.5 mM b-mercaptoethanol, 

10 mM MgCl 2 supplemented with ten units of 

RNAguard (Pharmacia) and 100 mM of dNTP and preincubated 

at 37 C for ®ve minutes. The 30 S ribosomes 

(2 ml of 0.2 mM) were added followed by ten minutes 

incubation. Addition of 1 ml (10 mM) of uncharged 

tRNA fMet (Sigma) was succeeded by 15 minutes incubation 

before supplementing with 1 ml (two units) of 

AMV-RT. Reactions were stopped after 20 minutes by 

precipitation in ethanol and subsequently resuspended 

in 5 ml of FD. An RNA sequence reaction was made 

according to Thisted et al. (1995). Toeprint ef®ciency was 

determined by Phosphor Imager and the Image-quant 

software package (Molecular Dynamics). 

Acknowledgements 

This work was supported by the Center for Interaction, 

Structure, Function, and Engineering of Macromolecules 

(CISFEM) and the Plasmid Foundation. 

References 

Altuvia, S., Kornitzer, D., Teff, D. & Oppenheim, A. B. 

(1989). Alternative mRNA structures of the cIII 

gene of bacteriophage determine the rate of its 

translation initiation. J. Mol. Biol. 210, 265±280. 

Altuvia, S., Kornitzer, D., Kobi, S. & Oppenheim, A. B. 

(1991). Functional and structural elements of the 

mRNA of the cIII gene of the bacteriphage lambda. 

J. Mol. Biol. 218, 723±733. 

Belasco, J. G. & Higgins, C. F. (1988). Mechanisms of 

mRNA decay in bacteria: a perspective. Gene, 72, 

15±23. 

Berkhout, B., Kastelein, R. A. & van Duin, J. (1985). 

Translational interference at overlapping reading 

frames in prokaryotic messenger RNA. Gene, 37, 

171±179. 

Case, C. C., Roels, S. M., Jensen, P. D., Lee, J., Kleckner, 

N. & Simons, R. W. (1989). The unusual stability of 

the IS10 anti-sense RNA is critical for its function 

and is determined by the structure of its stemdomain. 

EMBO J. 8, 4297±4305. 

de Smit, M. & van Duin, J. (1990a). Secondary structure 

of the ribosome binding site determines the translational 

ef®ciency: a quantitative analysis. Proc. Natl 

Acad. Sci. USA, 87, 7668±7672. 

de Smit, M. & van Duin, J. (1990b). Control of procaryotic 

translation by mRNA secondary structure. Prog. 

Nucl. Acid Res. Mol. Biol. 38, 1±35. 

de Smit, M. & van Duin, J. (1994). Translational 

initiation on structured messengers another role for 

the Shine-Dalgarno interaction. J. Mol. Biol. 135, 

173±184. 

Franch, T. & Gerdes, K. (1996). Programmed cell death 

in bacteria: translational repression by mRNA endpairing). 

Mol. Microbiol. 21, 1049±1060. 

Gerdes, K., Helin, K., Christensen, O. W. & Lùbner- 

Olesen, A. (1988). Translational control and differential 

RNA decay are key elements regulating postsegregational 

expression of the killer protein 

encoded by the parB locus of plasmid R1. J. Mol. 

Biol. 203, 119±129. 

Gerdes, K., Thisted, T. & Martinussen, J. (1990). Mechanism 

of post-segregational killing by the hok/sok system 

of plasmid R1: the sok antisense RNA regulates 

the formation of a hok mRNA species correlated 

with killing of plasmid free cells. Mol. Microbiol. 4, 

1807±1818. 

Gerdes, K., Nielsen, A. K., Thorsted, P. & Wagner, 

E. G. H. (1992). Mechanism of killer gene reactivation: 

antisense RNA mediated RNaseIII cleavage 

ensures rapid turn-over of the hok, srnB and pndA 

effector mRNAs. J. Mol. Biol. 226, 637±649.


Gultyaev, A. P., Franch, T. & Gerdes, K. (1997). Programmed 

cell death by hok/sok of plasmid R1: 

coupled nucleotide covariations reveal phylogenetically 

conserved folding pathway in the hok family 

of mRNAs. J. Mol. Biol. 273, 26±37. 

Hartz, D., McPheeters, D. S., Traut, R. & Gold, L. (1988). 

Extension inhibition analysis of translation initiation 

complexes. Methods Enzymol. 164, 419±425. 

Higgins., C. F., McLaren, R. S. & Newbury, S. F. (1988). 

Repetitive extragenic palindromic sequences, 

mRNA stability and gene expression: evolution by 

gene conversion? ± a review. Gene, 72, 3±14. 

Hjalt, T. AÊ . H. & Wagner, E. G. H. (1995). Bulged out 

nucleotides in an antisense RNA are required for 

target RNA binding in vitro and inhibition in vivo. 

Nucl. Acids Res. 23, 580±587. 

Kittle, J. K., Simons, R. W., Lee, J. & Kleckner, N. (1989). 

Insertion sequence IS10 anti-sense pairing initiates 

by an interaction between the 5 0 -end of the target 

RNA and a loop in the anti-sense RNA. J. Mol. Biol. 

210, 561±572. 

Lanzer, M. & Bujard, H. (1988). Promoters largely determine 

the the ef®ciency of repressor action. Proc. 

Natl Acad. Sci. USA, 85, 8973±8977. 

Miller, J. (1972). Experiments in Molecular Genetics, Cold 

Spring Harbor Laboratory Press, Cold Spring Harbor, 

NY. 

Nielsen, A. K. & Gerdes, K. (1995). Mechanism of postsegregational 

killing by hok-homologue pnd of plasmid 

R483: two translational control elements in the 

pnd mRNA. J. Mol. Biol. 249, 270±282. 

Phillipe, C., Portier, C., Mougel, M., Grunberg-Manago, 

M., Ebel, J. P., Ehresmann, B. & Ehresmann, C. 

(1990). Target site of Escherichia coli ribosomal protein 

S15 on its messenger RNA. J. Mol. Biol. 211, 

415±426. 

Phillipe, C., Eyermann, F., Benard, L., Portier, C., 

Ehresmann, B. & Ehresmann, C. (1993). Ribosomal 

protein S15 from Esherichia coli modulates its own 

translation by trapping the ribosome on the mRNA 

initiation loading site. Proc. Natl Acad. Sci. USA, 90, 

4394±4398. 

Saenger, W. (1984). Forces stabilizing associations 

between bases: Hydrogen bonding and base stacking. 

In Principles of Nucleic Acid Structure, pp. 116± 

158, Springer-Verlag, New York. 

Siemering, K. R., Prazkier, J. & Pittard, A. J. (1994). 

Mechanism of binding of the antisense and target 

RNAs involved in the refulation of IncB plasmid 

replication. J. Bacteriol. 176, 2677±2688. 

Thisted, T. & Gerdes, K. (1992). Mechanism of post-segregational 

killingby the hok/sok system of plasmid 

R1. Sok antisense RNAregulates hok expression 

indirectly through the overlapping mok gene. J. Mol. 

Biol. 223, 41±54. 

Thisted, T., Nielsen, A. K. & Gerdes, K. (1994a). Mechanism 

of post-segregational killing: translation of 

Hok, SrnB and Pnd mRNAs of plasmids R1, F and 

R483 is activated by 3 0 -end processing. EMBO J. 13, 

1950±1959. 

Thisted, T., Sùrensen, N. S., Wagner, G. H. & Gerdes, K. 

(1994b). Mechanism of post-segregational killing: 

Sok antisense RNA interacts with Hok mRNA via 

its 5 0 -end single-stranded leader and competes with 

the 3 0 -end of Hok mRNA for binding to the mok 

translational initiation region. EMBO J. 13, 1960± 

1968. 

Thisted, T., Sùrensen, N. S. & Gerdes, K. (1995). Mechanism 

of post-segregational killing: secondary structure 

analysis of the entire hok mRNA from plasmid 

R1 suggests a fold-back structure that prevents 

translation and antisense RNA binding. J. Mol. Biol. 

247, 859±873. 

Turner, D. H. & Bevilacqua, P. C. (1993). Thermodynamic 

considerations for evolution by RNA. In The 

RNA World (Gesteland, R. F. & Atkins, J. F., eds), 

pp. 447±464. Cold Spring Harbor Laboratory Press, 

Cold Spring Harbor, NY. 

Wilson, I. W., Siemering, K. R., Prazkier, J. & Pittard, 

A. J. (1997). Importance of structural differences 

between complementary RNA molecules to control 

of replication of an IncB plasmid. J. Bacteriol. 179, 

742±753. 

Yanisch-Perron, C., Vieira, J. & Messing, J. (1985). 

Improved M13 phage cloning vectors and host 

strains: nucleotide sequences of the M13mp18 and 

pUC19 vectors. Gene, 33, 103±109. 

Zarrinkar, P. P. & Williamson, J. R. (1994). Kinetic intermediates 

in RNA folding. Science, 265, 918±924. 

Zubay, G. (1973). In vitro synthesis of protein in 

microbial systems. Annu. Rev. Genet. 7, 267±287. 

Edited by D. E. Draper 

(Received 14 April 1997; received in revised form 14 July 1997; accepted 17 July 1997)

Programmed Cell Death by hok/sok of Plasmid - Department of Plant ...

You also want an ePaper? Increase the reach of your titles

Delete template?

Save as template?