JOURNAL OF BACTERIOLOGY, Jan. 1980, p. 422-426 Vol. 141, No. 1
Direct Selection and Analysis of Mutational Events Which
Diminish the Level of Expression of the trp Operont
JOHN C. ANDREWSt AND RONALD L. SOMERVILLE*
Department of Biochemistry, Purdue University, West Lafayette, Indiana 47907
A new procedure was devised for selecting, from lac+ galE strains of Esche-
richia coli, mutants resistant to galactoside-induced lysis. When applied to tip-
lac fusions, our method yields down mutations in the trp promoter.
The metabolism of lactose and related ,B-D- into the tip-lac fusion W205 (23) (Fig. 1). We
galactosides in Escherichia coli is mediated by then perfected a selection procedure which gen-
the gene products of the lac operon. The hy- erated several galactose-sensitive, galactoside-
drolysis of such disaccharides, catalyzed by ,B- resistant mutants, one of which was character-
galactosidase (lacZ gene product), produces in- ized as a promoter mutation. The bacterial
tracellular D-galactose, which is further metab- strains utilized (Tables 1 and 2) were all sub-
olized by the enzymes of the gal operon. In the strains of E. coli K-12. The F'tip+ episome (5)
presence of a neutral carbon source, exposure of carries the chromosomal markers att80, tonB+,
a galE strain (deficient in uridinediphosphoga- trpPOEDCBA+, and cysB+.
lactose 4-epimerase) to galactose or a hydrolyz- Because galactoside-induced lysis of sensitive
able ,B-D-galactoside causes cell lysis (21), owing cells ultimately leads to the release of free galac-
to the accumulation of UDP-galactose (6). Bac- tose, precautions must be taken in any selective
teriolysis induced in this manner constitutes se- scheme to prevent the subsequent lysis of the
lective pressure suitable for examining a range desired galactoside-resistant, galactose-sensitive
of genetic events that reduce lac operon expres- cells. Seeding selective medium with an auxo-
sion (14). The acquisition of resistance to f8-D- trophic gal+ gaiR strain prevents such undesir-
galactosides by galE mutants may be attribut- able lysis. Strain B78AHis, a histidine auxotroph
able to primary events which generate galE+ of a gaiR strain first isolated by Buttin (4),
revertants or to secondary mutations within the fulfills this "scavenger" function.
lac and gal operons. By screening a collection of Selections (Table 2) were performed in both
,B-D-galactoside-resistant galE mutants for re- tipR (A8-W205) and trpR+ (A7-W205) back-
tention of galactose sensitivity, Malamy identi- grounds. Independent L-broth cultures of each
fied several lac operon-specific mutational parental strain were grown overnight at 37°C.
events (14). The samples were centrifuged, washed twice,
In strains carrying compound operons, such and resuspended in twice the original volume of
as the trp-lac fusions X7713 (18, 20, 22) and saline. Samples (0.1 ml) of these suspensions
W205 (24), lac gene expression is directly con- were spread on selective medium. Each plate
trolled by the tryptophan regulatory elements; was also seeded with 5 x 109 washed cells of
repression by exogenous tryptophan results in mutant B78AHis. The plates were incubated at
low-level synthesis of both tip- and lac-specific 370C for 4 days. Ten colonies from each selection
gene products (16, 23). Transcription initiated at plate were picked and screened for galactose
the trp promoter produces a compound trp-lac sensitivity by streaking on medium with 0.2%
message whose translation results in the coor- galactose. From this initial screening, we ob-
dinate synthesis of fully active tip- and lac-spe- tained a collection of phenylgalactoside (PG)-
cific enzymes (23). In a galE derivative of such resistant, galactose-sensitive derivatives of A8-
a fusion strain, resistance to f8-D-galactosides W205 and A7-W205. On the basis of fi-galacto-
should include tip-linked mutations, such as trp sidase activities and zone size upon exposure to
polar mutations and down mutations in the pro- PG, representative mutant stocks were pre-
moter, which reduce the efficiency of expression served for further study. fi-Galactosidase levels
of the trp operon. were precisely determined for each of the se-
In this study, a galE lesion was introduced lected isolates, as were galactose- and PG-in-
duced zones of lysis. It was evident that PG
t Journal paper no. 7675 of the Purdue University Agricul- resistance can result from several distinct mu-
tural Experiment Station.
t Present address: Department of Microbiology, Bur- tational events as follows.
roughs-Wellcome Co., Research Triangle Park, NC 27709. (i) lacY lesions prevent galactoside transport
VOL. 141, 1980 NOTES 423
trpE trpD trpC trpB trpA lcZ lacY lacA TABLE 1. Derivatives of trp-lac fusion strain W205
p, __ trp-loc mRNA
Bacterial strain Pertinent genetic markersa
CH20H CH20H W205 R- (tonB-lacIPO) trpR
cysB W205 A(tonB-lacIPO) cysB trpR
cysB W205 R+ A(tonB-lacIPO) cysB
OH OH thr W205 I(tonB-lacIPO) thr
D-goloctose A7-W205 A(tonB-lacIPO) galE
A8-W205 Q(tonB-lacIPO) trpR galE
UDPG - UDPGal - Gal-I-P PG3.lb A(tonB-lacIPO) trpR galE
E T K trpP0l1
FIG. 1. The W205 trp-lac fusion (23) and its rela- W205 R- (PG3-1)/F'tipc A(tonB-lacIPO) trpR trpPIOI/
tionship to selection for diminished exprcssion of the F'trp+ tonB+
W205 R+ (PG3-1)C A(tonB-lacIPO) trpPlOl
trp operon. The lac end of the deletion removes lacP W205 R- (PG3-1)c A(tonB-lacIPO) trpR trpPlOl
and lacO but leaves the lacZ gene intact. The trp
functions are indistinguishable from the wild type aW205 and all its derivatives carry (proAB-lac) deletion
(20). Depression of the trp operon, as in trpR mutants, XIII (24). trpPlOl is our allele number for the down mutation
results in a 12- to 13-fold increase in fl-galactosidase in the promoter carried by strain PG3-1 (see text).
b Selected in this study as PGr Gal mutants.
synthesis (16). Galactose liberated by the hydrolysis c
Constructed to characterize PG3-1 mutation.
of appropriate 3- D -galactosides enters central metab-
olism following phosphorylation catalyzed by galac-
tokinase (K), transglycosylation catalyzed by galac- TABLE 2. Miscellaneous bacterial strainsa
tose 1-phosphate uridyl transferase (T), and epimer- Strain Relevant genetic markers
ization of the 4'-hydroxyl of UDP-galactose, cata-
lyzed by epimerase (E). Cells unable to catalyze the W3110 Wild type
final reaction lose viability, owing to the accumula- YS57/F'trp Ftrp+ tonB/his pro A(trpE-A-
tion of UDP-galactose. Strain W205 lacks the tonB His- B78A F- galR lacY str leu thi his
locus, which specifies a protein involved in phage YS57/F'(PG3-1) his pro A(trpE-A-tonB)/
penetration and iron transport (29). This leads to a F'trpP101
requirement for Fe. To allow maximum growth, ALD102 &(tipE-D)
FeSO4 (25 ,uM) was routinely added. Our basal mini- SP-333 tna-2 trpR AtrpLD102
mal medium was the forrnulation of Vogel and Bon- W3110 (PG3-1)b trpPlOl
ner (28). The Fe present in Difco agar, tryptone, or W3110 (PG3-1)/F'lO1" trpPlOl/F'tipB
yeast extract is sufficient to allow normal growth (16). W3110 (PG3-1)/FVA2b trpPlOl/F'tipA
and are therefore unable to utilize melibiose at W3110/F'101 cysB/F'trpB
420C (18). 101/F'trp+ cysB t7pB/F'trp+
(ii) lacZ lesions prevent galactoside hydrolysis A2/F'trp+ cysB trpA2/F'trp+
and were resistant to PG- and lactose-induced a Unless otherwise specified, media are as described by
bacteriolysis but sensitive to galactose and mel- Miller (17). Vogel and Bonner (28) salts solution (medium E)
ibiose. Activity determinations directly verified supplemented with 0.2% glucose was used as minimal medium.
the absence of ,B-galactosidase. Normal expres- In some cases glucose was omitted and a neutral carbon
sion ofthe trpAB genes was observed. This source, 0.2% sodium succinate or glycerol, was substituted.
For cultivation of strain W205 and its derivatives (all of which
evidence, coupled with the demonstration of harbor a pro-lac deletion), L-proline (20 1&g/ml) and thiamine
wild-type 5-methyl-DL-tryptophan (5-MT) re- hydrochloride (1 jig/ml) were routinely added. Repression of
sistance (data not shown), proved that the ab- the tip operon was achieved at 20 jig of L-tryptophan per ml.
sence of lacZ gene expression was not a trp L-broth is described by Lennox (13). Selective minimal me-
dium was supplemented, for A8W205, with 2 mg of sodium
operon-specific phenomenon. succinate per ml, 1 mg of phenylthiogalactoside per ml, 50 Ag
To conclusively identify such lesions as lacZ of PG per ml, L-prolne, L-tryptophan, and thiamine hydro-
mutations, W205 R+ Lac- isolates were trans- chloride; for A7W205, medium was supplemented with 2 mg
duced to Lac' with X plac5. This phage carries of sodium succinate per ml, 0.5 mg of phenylthiogalactoside
per ml, 100 mg of PG per ml, L-proline, L-tryptophan, and
only the lacZ+ locus (17), and hence the appear- thiamine hydrochloride.
ance of Lac' transductants restricts the lac le- b Strains constructed in this study to characterize the trp
sion to a sequence intemal to the lacZ gene. mutation of the PG3-1 isolate.
(iii) Some cysB-linked, trp-specific lesions
were predicted to diminish gene expression mulates no pathway intermediates (determined
across the entire trp-lac fusion, thereby reducing by specific color tests for anthranilic acid, indole,
lac activities. Strain PG3-1, isolated in the W205 and indoleglycerolphosphate), it grows slowly in
trpR galE background, displays properties con- the absence of tryptophan. When the galE le-
sistent with a mutation of this type (Tables 3 sion was overcome by transduction with A gall8,
and 4). Although the strain is Trp+ and accu- the strain failed to grow on lactose as the sole
424 NOTES J. BACTERIOL.
TABLE 3. Characterization of PG3-1 mutationa
Strain 16-Galactoaidase activity Tryptophan syn- Zone of growth inhibition (mm)'
0 +trp +IA a Gal PG 5-MTd IAe
A7W205 450 120 2,500 - - 46 41 33 0
A8W205 1,350 1,000 4,900 5.2 5.2 50 48 0 0
PG3-1 170 119 - 1.1 1.3 38 32 44 30
(PG3-1)/F'tip + 113 104 - - - 37 32 0 -
W205 R+ (PG3-1) 170 125 140 - - - - 58 36
W205 R- (PG3-1) 170 160 250 1.2 1.2 - - 45 30
W3110 - - - 2.0 2.0 - - 25 -
W3110 (PG3-1) - - - 0.9 0.9 - - 64 20
YS57/F'trp+ - - - 6.7 6.3 - - 25 -
YS57/F' (PG3-1) - - - 1.6 2.5 - - 46
a Earlier work (16) had demonstrated that the f?-galactosidase of stationary E. coli cells was metabolically
stable, thus allowing appreciable latitude in the choice of cell growth times before assay. To ensure that the
cultures had achieved balanced growth at the time of assay, the inoculum was always pregrown for 12 to 15 h
under the same conditions used in preparation of the cells for determination of enzymatic activity. The cells to
be assayed were cultivated for 12 to 15 h in minimal liquid medium containing 20,ug of FeSO4 per ml and
required nutrients or acid-hydrolyzed casein (0.1%). jB-Galactosidase was assayed by the procedure of Reznikoff
et al. (24).
bTryptophan synthetase a and 2 specific activities were assayed in the presence of excess complementing
subunit, according to the method of Smith and Yanofsky (25). A unit of tryptophan synthetase is that amount
of enzyme which will convert 0.1,umol of indole to tryptophan in 20 min at 370C. Crude extracts were prepared
by sonication or in a French pressure cell, and protein concentration was estimated by the method of Lowry et
Growth inhibition by 5-MT, a semiquantitative indication of the level of trp gene expression, was estimated
by diluting an overnight L-broth culture 1:10 or 1:100 in saline, then plating 0.1 ml on a minimal agar plate
containing required supplements. A sterile filter paper disk (13 mm) on the plate surface was saturated with 20
fd of a 2-mg/ml solution of 5-MT. The plate was incubated at 37°C for 15 to 20 h, or until confluent cell growth
around the perimeter of the plate was apparent. The diameter of any clear zone around the disk was measured.
The sensitivity of strains carrying a galE lesion was demonstrated in a similar way, except that the cells were
plated on a medium containing a neutral carbon source such as succinate or glycerol. The Penassay disk was
saturated with 10 or20 ul of a 2% sugar solution. Sensitivities to galactose (to verify the galE lesion), lactose, or
PG were used to assess the extent of expression of the lac-specific functions. A difference in zone diameter of 4
mm or greater was considered significant.
20-ul portion of a 1-mg/ml solution of 5-MT was used in the same manner as described above.
'A40-pl portion of a 4-mg/ml solution of indoleacrylic acid (IA) was used as described above.
carbon source. Both f8-galactosidase and tryp- strains) activities is apparent. This inability to
tophan synthetase activities were significantly respond to the tryptophan derepression signals
lower than the parental activities under condi- mediated by indoleacrylic acid results in in-
tions of tryptophan starvation. The PG3-1 lesion creased sensitivity to this compound and to
was transferred into a standard background both other tryptophan analogs (Table 3).
by cotransduction with cysB+ (supporting a lo- The location of the lesion in strain PG3-1 was
cation within the trp operon) and by transduc- established genetically using a test devised by
tion using strains harboring trp deletions (spe- Zurawski et al. (30) for attenuator mutants.
cifically AtrpLD102 and AtrpEDCBA) as recipi- Phage P1 kc lysates prepared on PG3-1 were
ents. A more complete analysis (Table 3) of the used to transduce tna-2 trpR AttrpLD102 to pro-
enzyme activities in wild-type and trp-lac fusion totrophy. Among the trp+ transductants were
backgrounds, coupled with increased sensitivi- found recombinants fully resistant to high levels
ties to the tryptophan analogs 5-MT and indo- of 5-MT. The same recombinants were growth-
leacrylic acid, proved that the lesion in PG3-1 is inhibited on media containing indole (5,ug ml-')
a cis-dominant mutation which reduces the and 5-methylanthranilic acid (100 ,ug ml-1).
expression of the trp operon independently of Since the tipLD102 deletion removes most of
repression mediated by the trpR system. the trp leader sequence as well as trpE and trpD,
In strains harboring the galactoside-insensi- the wild-type allele of PG3-1 must lie outside
tive mutation PG3- 1, an approximately fivefold the region of DNA deleted in At7pLD102. PG3-
reduction in derepressed levels of tryptophan 1 cannot be a regulatory mutation in the leader
synthetase and ,B-galactosidase (in fusion region similar to those characterized by Zu-
VOL. 141, 1980 NOTES 425
TABLE 4. cis-dominance of PG3-1 mutation impose tryptophan auxotrophy, but also cause
severe transcription termination, thereby reduc-
Strain synthetase ing the synthesis of downstream gene products
(26). Nor did our procedures yield changes in
a , a the tip attenuator (2, 3, 11), similar to those
W3110/F' Olb 4.1 1.6 1.7 - found by Zurawski et al. (30), or in the tip
W3110/F'A2b 1.3 2.7 - 1.7 operator (7, 15). In an earlier study, we isolated
101/F'trp+b 3.4 2.5 2.9 - a number of E. coli strains displaying enhanced
A2/F'trp+b 2.6 3.2 - 1.5 sensitivity to 5-MT (12). All had partial defects
W3110 (PG3-1)/F'101C 9.7 4.2 2.3 -
in one of the trp structural genes. It is not clear
W3110 (PG3-1)F'A2c 2.6 9.0 - 2.5
101/F'(PG3-1)c 10.1 1.1 0.12 - at present why the aforementioned studies failed
A2/F'(PG3-1)c 0.65 10.9 - 0.06 to yield promoter-down mutations such as that
in strain PG3-1, but these results, taken together
aE/C values quantitate the amount of enzyme with ours, emphasize the desirability of employ-
made from episomal genes relative to that from chro- ing a variety of selective procedures in mutant
mosomal genes (for details see Stetson and Somerville
). By positioning the PG3-1 mutation at each hunts directed toward identifying regulatory ele-
location it is demonstrated above that E/C increases ments.
when PG3-1 is on the chromosome and decreases In selective schemes that rest upon galacto-
when on the episome. This phenomenon verifies the side sensitivity mediated by galE mutations, it
cis-dominance of the PG3-1 mutation. (Tryptophan is of paramount importance that one avoid con-
synthetase specific activities are defined in footnote b ditions that might lead to the accumulation in
of Table 3.) the selection media of free galactose, since this
b These values were determined for cells grown in compound is readily taken up and converted to
unsupplemented Vogel and Bonner minimal medium. UDP-galactose. Even moderate levels of galac-
See Stetson and Somerville (27).
These values were detennined in this study for tose generated in situ may eliminate or diminish
cells grown in medium supplemented with 0.1% acid- the survival of the galE Lac- cells that are the
hydrolyzed casein. object of a mutant hunt. Our procedures effec-
tively overcome the problem of nonspecific kill-
rawski et al. (30), and it is most probably a ing due to galactose accumulation in the selec-
change that diminishes the normal functioning tion media. Phenylthiogalactoside, a weak com-
of the trp promoter, based upon the following petitive inhibitor of ,B-galactosidase, also inhibits
the induction of the gal operon (4). Both of
lines of evidence. First, it is a cis-dominant mu- these effects "tighten up" the selection proce-
tation which severely restricts the expression of dure. The inclusion of a nongrowing scavenger
adjacent trp (and lac, in the case of fusion strain constitutive for the uptake and catabolism
strains) genes (Table 4). Second, it lies outside of galactose (in our case, strain B78A his) further
of the trp leader region, because the wild allele reduces the likelihood that galE Lac- cells will
of PG3-1 is carried by the attenuator deletion perish during the period of exposure to selection.
AtrpLD102 (8, 9). We have designated the mu-
tation as trpPlOl. This work was supported by Public Health Service grant
In many respects the PG3-1 lesion resembles GM-22131 from the National Institute of General Medical
a naturally occurring promoter-down mutation Sciences.
We thank C. Yanofsky, W. Reznikoff, M. Feiss, and H.
discovered in Shigella dysenteriae 16 by Mioz- Echols for bacterial and phage strains. The capable technical
zari and Yanofsky (19). In strain PG3-1, the trp assistance of Linda Eades and Barbara Nicodemus is grate-
or trp-lac operon functions at a low constitutive fully acknowledged.
level (about 10% of normal) and is not subject to
further reduction by tip holorepressor. Because LITERATURE CITED
a small but significant increase in downstream 1. Beckwith, J. R., E. R. Signer, and W. Epstein. 1966.
expression is seen following tryptophan starva- Transposition of the Lac region of E. coli. Cold Spring
Harbor Symp. Quant. Biol. 31:393-401.
tion (Table 4, line 6), we argue that the normal 2. Bertrand, K., L. Korn, F. Lee, T. Platt, C. L. Squires,
trp attenuation mechanism remains functional. C. Squires, and C. Yanofsky. 1975. New features of
The relative inability of strain PG3-1 to undergo the regulation of the tryptophan operon. Science 189:
derepression is accompanied by increased sen- 22-26.
3. Bertrand, K., C. Squires, and C. Yanofsky. 1976. Tran-
sitivity to 5-MT and indoleacrylic acid, as seen scription termination in vivo in the leader region of the
by enhanced zones of inhibition on solid medium tryptophan operon of Escherichia coli. J. Mol. Biol.
(Table 3). 103:319-337.
Notably absent from our collection (over 100 4. Buttin, G. 1963. Mechanismes regulateurs dans la bio-
synthese des enzymes du metabolisme du galactose chez
in all) of galactoside-resistant strains are polar Escherichia coli K,2. II. Le determinisme genetique de
tip mutants. These pleiotropic lesions not only la regulation. J. Mol. Biol. 7:183-205.
426 NOTES J. BACTERIOL.
5. Fredericq, P. 1969. The recombination of colicinogenic Ippen, E. R. Signer, and J. R. Beckwith. 1970.
factors with other episomes and plasmids, p. 163-178. Fusions of the lac and trp regions of the Escherichia
In G. E. W. Wolstenholme and M. O'Connor (ed), coli chromosome. J. Bacteriol. 104:1273-1279.
Bacterial episomes and plasmids. Little, Brown and Co., 19. Miozzari, F., and C. Yanofsky. 1978. Naturally occur-
Boston. ring promoter down mutation: nucleotide sequence of
6. Fukaswa, T., and H. Nilkaido. 1961. Galactose sensi- the trp promoter/operator/leader region of Shigella
tive mutants of SalnoneUa. H. Bacteriolysis induced dysenteriae 16. Proc. Natl. Acad. Sci. U.S.A. 75:5580-
by galactose. Biochim. Biophys. Acta 48:470-482. 5584.
7. Hiraga, S. 1969. Operator mutants of the tryptophan 20. Mitchell, D. H., W. S. Reznikoff, and J. R. Beckwith.
operon in Escherichia coli. J. Mol. Biol. 39:159-179. 1975. Genetic fusions defining trp and lac operon regu-
8. Jackson, E. N., and C. Yanofsky. 1972. Internal pro- latory elements. J. Mol. Biol. 93:331-350.
moter of the tryptophan operon of Escherichia coli is 21. Nikaido, H. 1961. Galactose sensitive mutants of Sabno-
located in a structural gene. J. Mol. Biol. 69:307-313. neUa. I. Metabolism of galactose. Biochim. Biophys.
9. Jackson, E. N., and C. Yanofsky. 1973. The region Acta 48:460-469.
between the operator and first structural gene of the 22. Reznikoff, W. S., and J. R. Beckwith. 1969. Genetic
tryptophan operon of Escherichia coli may have a evidence that the operator locus is distinct from the Z
regulatory function. J. Mol. Biol. 76:89-101. gene in the lac operon of Escherichia coli. J. Mol. Biol.
10. Kuhn, J. C., M. J. Pabst, and RK L Somerville. 1972. 43:215-218.
Mutant strains of Escherichia coli K-12 exhibiting en- 23. Reznikoff, W. S., C. A. Michels, T. G. Cooper, A. E.
hanced sensitivity to 5-methyltryptophan. J. Bacteriol. Silverstone, and B. Magnik 1974. Inhibition of
112:93-101. lacZ gene translation initiation in trp-lac fusion strains.
11. Lee, F., C. L Squires, and C. Yanofsky. 1976. Termi- J. Bacteriol. 117:1231-1239.
nation of tamnscription in vitro in the Escherichia coli 24. Reznikoff, W. S., J. Miller, J. Scalfe, and J. R. Beck-
tryptophan operon leader region. J. Mol. Biol. 103:383- with. 1969. A mechanism for repressor action. J. Mol.
393. Biol. 43:201-213.
12. Lennox, E. S. 1955. Transduction of linked genetic char- 25. Smith, OH., and C. Yanofsky. 1972. Enzymes involved
acters of the host by bacteriophage P1. Virology 1:190- in the biosynthesis of tryptophan. Methods Enzynol. 5:
13. Lowry, 0. H., N. J. Rosebrough, A. L Farr, and R. J. 26. Somerville, R. L, and C. Yanofsky. 1965. Studies on
Randall. 1951. Protein measurement with the Folin the regulation of tryptophan biosynthesis in Esche-
phenol reagent. J. Biol. Chem. 193:265-275. richia coli. J. Mol. Biol. 11:747-759.
14. Malamy, M. H. 1966. Frameshift mutations in the lactose 27. Stetson, H., and R. L Somerville. 1971. Expression of
operon of E. coli. Cold Spring Harbor Symp. Quant. the tryptophan operon in merodiploids of Escherichia
Biol. 31:189-201. coli. Mol. Gen. Genet. 111:342-351.
15. Matsushiro, A., K. Sato, J. Ito, S. Kida, and F. Ima- 28. Vogel, H. J., and D. M. Bonner. 1956. Acetylomithinase
moto. 1965. On the transcription of the tryptophan of Escherichia coli: partial purification and some prop-
operon in Escherichia coli. I. The tryptophan operator. erties. J. Biol. Chem. 218:97-106.
J. Mol. Biol. 11:54-63. 29. Wang, C. C., and N. A. Newton. 1969. Iron transport in
16. McPartland, A., and R. L Somerville. 1976. Isolation Escherichia coli: relationship between chromium sen-
and characterization of mutations creating high-effi- sitivity and high iron requirements in mutants of Esch-
ciency trncription initiation signals within the tip erichia coli. J. Bacteriol. 98:1135-1141.
operon of Escherichia coli. J. Bacteriol. 128:557-572. 30. Zurawski, G., D. Elseviers, G. V. Stauffer, and C.
17. Miller, J. H. 1972. Experiments in molecular genetics. Yanof8ky. 1978. Translational control of transcription
Cold Spring Harbor Laboratory, Cold Spring Harbor, termination at the attenuator of the Escherichia coli
N.Y. tryptophan operon. Proc. Natl. Acad. Sci. U.S.A. 75:
18. Miller, J. H., W. S. Reznikoff, A. E. Silverstone, K. 5988-5992.