J. Mol. Biol. (1975) 97, 225-235
Isolation, Characterization and Complementation of Salmonella typhimurium Chemotaxis Mutants DANA ASWADt AND D. E. KOS~rr.AwD, JR
Department of Biochemistry Univer, ity of California Berkeley, Calif. 94720, U.S.A. (Beceived 21 Oaober 1974) A novel technique has been developed for selecting non-chemotactic mutants of
Salmonella ~jphimurium that are defectivo in transmission of information from the chemoreceptor to the flagella. This technique utilizes a preformed gradient in a vertical column of liquid medium, which separates wild types, non-motile and motile but non-chemotactie into three regions. In the first application of the technique, 71 such mutants were obtained, of which 67 exhibited smooth motility and four had unco-ordinated motility. Complementation analysis, done by abortive transduction of recombination-defective recipients, indicates that there may be as many as six genetically distinct classes of ehe mutants in Salmonella. 1. I n t r o d u c t i o n One of the major advantages of using bacterial chemotaxis as a model for sensory reception is that one can perform sophisticated genetic manipulations of the molecular machinery that controls this behavior. Numerous mutants defective in chemotactic behavior have already been isolated in Escherichia coli and Salmonella typhimurium (Armstrong et al., 1967; Hazelbauer et al., 1969; Iino & 0guchi, 1972; Vary & Stocker, 1973; Parkinson, 1974). Adler and co-workers (Armstrong et al., 1967; Hazelbauer et al., 1969) have divided non-chemotactic mutants into two types: specific, which results in loss of ability to respond towards specific attractants or repellents; and genera], which cannot respond to any attractants or repellents. Armstrong & Adler (1969) studied 38 generally non-chemotactic mutants of E. coli and found three complementation groups, all of which mapped near the nw~ and fla region of the chromosome. Parkinson (1974) has recently reported a fourth complementation group. Chemotactie migration results from a selective suppression of spontaneous tumbling when the bacterium is swimming in a favorable direction (Macnab & Koshland, 1972; Berg & Brown, 1972). The control of tumbling frequency is therefore the means b y which the bacteria migrate up or down a gradient and the control t h a t is defective in generally non-chemotactic mutants. A total of 85% of the E. coli mutants t h a t were isolated by Parl~inson could not tumble spontaneously, while the remaining 15 ~ exhibited incessant tumbling with no periods of smooth motility. Similar mutants t Present address : Department of Biology, University of California, Los Angeles, Calif. 90024, U.S.A. 225
HfrB2 rectA22 recA1 strA
dhuA m o t
Wild type
metE1926
dhuA mot recA1
che-222(p) metE1926
hisF8786 thyA1981
che-222(p) recA1
trpA8 hisC527 (amber) che-lll (amber)
TA1859
ST1
ST2
ST8
ST20
ST23
ST25
SL4O41
Genotype
DB43
Strain
Derived from ST20 in two steps. First the mete was transduced with phage grown on wild type to make the strain rose +. The f e c a l was then introduced by conjugation with DB43 Vary & Stocker (1973).
Generally non-chemotactic, smooth motility. Recombination defective
Generally non-chemotactie. Unco-ordinated, frequent tumbling. Responds to a steep temporal gradient of attract.ant
ST20
che- mutants isolated in this study, except strain
Derived from a mutagenized culture of ST1 by penicillin selection (for his-) and trimethoprim selection (for thyA - )
Diethylsulfate mutagenesis of ST2, then screening of non-swarming isolates for motile cells (see Results)
TA1859 made thy- by trimethoprin selection, then f e c a l inserted by conjugation with strain DB43
Wild type ohemotaxis. This strain is the parent for all
Generally non-chemotaotio Smooth motility
Flagellated but non-motile Recombination defective
Diethylsulfate mutagenesis of ST1, then penicillin selection for met-
Selected for good serine taxis from LT2 strain of B. Ames
Normal chemotaxis and motility Parent for ST20
B. Ames.
B. Ames, Dept of Biochemistry, University of Calif., Berkeley
Source
Flagellated but non-motile
Used to transfer reoA mutation to other strains for use as recipients in abortive transduction
Relevant properties
Genotype, relevant properties a n d sources o f bacterial 8trains
TABLE 1
CHEMOTAXIS MUTANTS
OF E. T Y P H I M U R I U M
in ~ typhimurium h a v e been reported (Vary & Stocker, 1973; I i n o &
227
Oguehi,
1972). I n pursuing our studies on the transmission of information in the chemotactic system (Aswad & Koshland, 1974,1975)i it was desirable to isolate generally nonehemotactic m u t a n t s in the 8almoneUa system. Since the determination of the genes identified with the chemotactic machinery and their function requires the generation of a large n u m b e r o f mutants, a rapid and efficient isolation procedure was needed. A new method was devised utilizing a preformed a t t r a c t a n t gradient in a liquid column in order to separate generally non-chemotactic m u t a n t s from b o t h wild t y p e and non-motile bacteria. Using this method we were able to isolate a large n u m b e r of non-chemotactic m u t a n t s in a relatively short time. The m u t a n t s thus isolated were subjected to motility p a t t e r n and complementation analysis and fell into six distinct classes. 2. M a t e r i a l s a n d M e t h o d s
(a) Bacter/a The bacterial ~trains, all derived from ~. t/yphimurium LT2, that were used in developing the che mutant isolation and complementation procedures are described in Table 1. Media and growth conditions have been described elsewhere (Aswad & Koshland, 1974). (b) Oenetio techniques Diethylsulfate mutagenesis, penicillin selection and phage transduetion of ,.%lmonella were as described in detail by Roth (1970). P22-int3t, a non-lysogenizing mutant of phage P22, obtained from B. Ames, was used for all transduetions. Trimethoprin selection of thyA mutants has been described by Okada e~ a/. (1962). Transfer of ~ec.4 from strain DB43 into recipient strains was accomplished in the following manner. A 0.2-ml sample of thyA recipient (overnight nutrient broth culture) and 0.5 ml of DB43 (overnight nutrient broth culture) were inoculated into separate tubes containing 10 nd of nutrient broth and incubated for 3 to 4 h at 37°C. One ml of DB43 and 2 ml of recipient were then mixed and filtered and the filter was placed "face up" on a tryptone semisolid agar plate at 37°C for 90 rain, then into 1 ml of VBC medium (minimal salts plus citrate medium Of Vogel & Bonner, 1956) in a small tube and mixed vigorously for 1 rain to disengage mating bacteria. A 0-2-ml sample was then spread on minimal glucose plates and incubated for 24 h. Ten of the smallest colonies were streaked on nutrient agar to obtain single colony isolates. These were grown individually in nutrient broth and tested for sensitivity to u.v. light b y allowing 1 drop to run across the surface of a nutrient agar plate, then exposing half of the plate to a 15 W germicidal lamp (General Electric) at a distance of 3 3 c m for 7 s. After overnight incubation at 37°C recA- strains exhibited growth only on the non-irradiated half of the inoculum, while wild t y p e grew over the entire inoeulum.
3. Results (a) Attempts to isolate generaUy non.chemotadic mutants by negative swc~rmselectio~ A t t e m p t s .were m a d e to isolate generally non-chemotactic m u t a n t s of Salmonella b y the methods of Armstrong e~ al. (1967). A thick suspension of diethylsulfate mutagenized bacteria, strain ST2, was placed as a small spot in t h e center o f a t r y p t o n e semisolid agar plate. During an overnight incubation, the wild t y p e bacteria swarmed outwards from the inoculum, leaving a dense population at the center. A t Gene symbols in this paper are those used by Sanderson (1967).
228
D. ASWAD
AND D. E. KOSHLAND
JR
sample was taken from the center and subjected to a similar selection. After ten such serial selections, samples were grown in nutrient broth and plated at a density of about 50 cells per plate on Tryptone semisolid medium (0.35/o agar). Two hundred non-swarming colonies were picked, grown in nutrient broth and screened for motility by examination with a microscope. Only one culture with vigorous motility was found; the rest were probably mot and fla mutants. The single ohe mutant that was found, strain ST20, had smooth motility with no spontaneous tumbling. The dif~culty encountered with the swarm selection technique is probably due to the fact that the gradients that are necessary to draw the wild type away from the inoculum site do not exist prior to placing the inoculum on the plate. The gradients are established only after the bacteria at the inoculum site havo had time to metabolize a significant amount of certain nutrients in the tryptone that serve as attractants. Wild type, non-motile, and non-chemotactic bacteria will all multiply at equal rates before these gradients are established. By the time the gradients are established, the inoculum is so dense with wild type, that separation is difficult. Moreover, the use of semisolid medium inhibits movement of the bacteria and makes the separation rate much slower than it could be in a liquid medium. (b) The preformed liquid gradient technique To overcome the difficulties of the semisolid agar technique, it seemed reasonable to try a method (a) that used a preformed gradient; (b) that used a liquid rather than a semiaolid medium and (c) that created conditions that would separate non-motile from non-chemotactic bacteria, as well as non-motile from wild types. The approach used is shown in Figure 1. By placing the bacteria in a narrow band at the edge of a step gradient it was assumed that the wild type bacteria would diffuse only into the region of the attractant because they were under strong chemotactic pressure. The non-chemotactic bacteria would not respond to the gradient but would diffuse further than the non-motile ones in a symmetrical manner above and below the original band. Many non-chemotactic mutants could thus be found in the lower region where neither non-motile nor wild type individuals would be expected. To test this technique, cultures of a non-chemotactic mutant (strain ST20), a non-motile mutant (strain TA1859) and wild type (strain ST1) were independently tested in separate experiments using the abote apparatus. The distribution of the bacteria was measured at various time intervals after layering them in the gradient and the results are shown in Figure 2. As predicted, the non-motile strain did not show any significant change in distribution during the 60-minute incubation period. The wild type strain responded to the gradient by moving up the tube slightly. The che- strain behaved in a quite unexpected manner. Instead of spreading symmetrically in both directions from the band, it showed a marked preference for swimming down the tube. The asymmetrical diffusion of the the- smooth swimmer is apparently due to the influence of gravity. A similar phenomenon has been observed with sperm (Branham, 1966). Branham's explanation for this was that the sperm head had a higher specific gravity than the tail and this caused them to assume a head-down, tail-up orientation. An alternative explanation is that the slow sedimentation of the bacteria causes the flagella to swing behind the soma due to the hydrodynamic drag on the flagella. Once ~he bacterium is properly oriented, its forward velocity is sufficient to propel it downward at a rate of 0.17 em per minute (calculated from data given by Macnab & Koshland, 1972).
Pump
1
Grsdient maker
0-2
7
I
[Serine ] (raM) 0'4 0-6 0.8 I I
1.0
6Serine
F///A
///l.,f
Attractant
,////I / ///.4
N
\ Bacteria ~
4
\ ",,
Bacteria
3
~-,,,=,~ 2
\ ",
I
Glycerol '\
0
1.0
2.0 3.0 Glycerol % ( w / v )
Fzo. 1. Preformed liquid gradient used for isolation of ehemotaxis m u t a n t s . The drawing on t h e left indicates how t h e gradient w a s constructed. T h e drawing o n t h e r i g h t shows t h e dist r i b u t i o n of t h e selection t u b e contents a t t h e beginning of t h e e n r i c h m e n t period. All solutions were m a d e u p in minimal m e d i u m a n d p u m p e d t h r o u g h a stainless steel inlet pipe (no. 20 gauge hypodermic needle) t h a t e x t e n d e d to t h e b o t t o m of t h e 18 m m × 150 m m glass tube. F i r s t 2.0 ml of 1.0 mM-serine ( a t t r a c t a n t ) was p u m p e d in, followed b y 2.5 ml of 1.0 m~-serine, containing glycerol, The glycerol concentration was initially 0 % (w/v) a n d increased linearly to 1.0% a t t h e e n d of t h e 2.5-ml p u m p i n g segment. Next, 0.2 m l of bacterial suspension (4 × 101 b a c t e r i a / m l in 0.5 m l of a solution of 0-5 mM-serine, 1"5% glycerol) was p u m p e d in, followed b y ;2"5 m l glycerol, which increased linearly from 2-0~/o t o 3.0%. The glycerol gradient was necessary to stabilize t h e a t t r a c t a n t g r a d i e n t against convection. Top
w ÷
real
che -
-
(smooth)
== .F_
o
"*--..
30 rain
~
Bottom
min
rain
\ Relative concentration of bacteria
FIG. 2. Behavior of wild t y p e (w +, s t r a i n ST1), non-motile (mot-, s t r a i n TA1859) a n d motile b u t non-ohemotectic ( t h e - , s t r a i n ST20) b a c t e r i a in a vertical step gradient of a t t r a c t ~ n t . The selection a p p a r a t u s is described in Fig. 1. The distribution of t h e b a c t e r i a was m e a s u r e d b y lights c a t t e r i n g , using t h e a p p a r a t u s described b y D a h l q u i s t e~ a/. (1972). The distributions were recorded a t 6 min a n d 30 m i n after introducing t h e bacteria. 16
230
D. ASWAD
AND
D. E. KOSHLAND
JR
To establish that this explanation and not some repellent action of the gradient was responsible, the orientation of the step gradient was reversed, i.e. the attractant was placed in the bottom third of the tube and the bacteria layered at its top edge. With this configuration the non-chemotactic bacteria again diffused predominantly into the bottom region, showing that gravity and not repellent action was responsible. The gravity effect was an unexpected boon, since it meant that a more dramatic difference in migration would be obtained than that based solely on random diffusion of the non-chemotactic mutants. A reconstruction experiment was done in which a mixed culture with a ratio of one the- bacterium (ST20) per 1000 wild type bacteria was introduced into the liquid column apparatus. After 60 minutes of incubation, all but the bottom 1-0 ml of the tube contents were aspirated off and the walls of the tubes were rubbed with ethanol-soaked cotton to kill bacteria that might be adhering. A 2.0-ml volume of sterile VBC medium was added to the tube, which was then capped and incubated for 24 hours, a t which time turbidity appeared. The relative frequency of chebacteria increased 11-fold after one cycle of selection. This proeedure was then applied to a mutagenized culture of strain ST23 (wild type chemotaxis). After mutagenizing with diethylsulfate, the culture was enriched for 6he- mutants in the liquid column apparatus four times in succession, allowing for turbid growth of the culture between each enrichment. The final culture was plated out on tryptone semisolid medium to screen for non-swarming colonies. Since tryptone contains a mixture of many attractants, only generally non-chemotactie and nonmotile mutants will fail to swarm. Mutants defective in a specific class of at%ractants should not be isolated by this screening. Plate I shows that two distinct colony morphologies were present: tiny compact colonies, typical of 0/~- and non-motile strains, and large swarming colonies, typical of che- strains. Among the tiny colonies it was possible with the aid of a dissecting microscope to distinguish between nonmotile and non-chemotactic colonies, because the latter had less well-defined boundaries. The relative frequencies of the three colony types were 10% wild type (large colonies), 50% non-motile (small colonies with sharp boundaries) and 40% non-chemotactic (small colonies with fuzzy boundaries). (c) Motility and behavior of Salmonella che mutants A total of 71 che mutants were isolated from a single mutageuized culture as described above. These 71 mutants, together with the single mutant isolated as described in section (a), above, brings the total number of mutants isolated in this study to 72. Although all these mutants were highly motile, none of them exhibited a normal frequency of spontaneous tumbling. Sixty-eight exhibited only smooth co-ordinated motility with no evidence of tumbling when examined in either minimal medium or nutrient broth. The remaining four mutants had unto-ordinated motility indistinguishable from that of strain SIA041, an unee-ordinated mutant of Salmonella reported by Vary & Stocker (1973). Examples of these two motility types are shown in Plate II. We have previously demonstrated that derivatives of strain SL4041 exhibit smooth motility in response to a temporal gradient of 0 to 1 mM-serine (Aswad & Koshland, 1974). The four unco-ordinated mutants isolated in this study were subjected to a similar temporal gradient of serine to determine whether they were capable of exliibiting a smooth response. One of the muta~lts, che-223(e), showed a definite smooth
:PLATE I. Detection of ohemotaxis m u t a n t s on t r y p t o n e semisolid agar. Samples of 30/~1 or less containing 30 to 50 bacteria were spread on t r y p t o n e semisolid plates containing 0.5% agar. This agar concentration, which is twice t h a t normally used to observe chemotaetic swarms, prevents wild type colonies from spreading and overlapping the che- colonies. After spreading the bacteria, the plates were incubated at 37°C for 5 h to dry the surface of the agar. They were t h e n transferred to a 30°C room for overnight incubation. (a) Strain ST1, a wild type. (b) Strain ST20, a c h ~ - smooth-swimming m u t a n t . (c) :From a culture of ST23 t h a t h a d been mutagenized and selected 4 times in succession by the liquid column method. This plate shows a mixture of wild type, che- and non-motile cells and is typical of those plates used in screening for chemutants. The non-motile and c h e - colonies are not distinguishable in this photograph b u t we were able to accurately distinguish the two types by close examination of colony morphology under moderate magnification. [facing ~. 230
tp
g
t~
,
b
.i
PT.A~E I I . T y p i c a l m o t i l i t y t r a c k s of s m o o t h a n d u n e o - o r d i n a t e d t y p e m u t a n t s isolated in t h i s sl~udy. (a) Che-222 (p), a s m o o t h m u t a n t . (b} Che-221 (t), a n u n e o - o r d i n a t e d m u t a n t . T h e t r a c t s are p r o d u c e d b y u s i n g a s t r o b o s c o p i c l i g h t source in (lark field w i t h a t i m e e x p o s u r e o f 4 s ( M a e n a h & K o s h l a n d , 1972).
PLATE I I I . C o m p l e m e n t a t i o n o f mot a n d che m u t a n t s u s i n g r e c o m b i n a t i o n - d e f e c t i v e recipients. Six d r o p s of a full-grown n u t r i e n t b r o t h c u l t u r e of r e c i p i e n t b a c t e r i a anti 3 d r o p s of d o n o r p h a g e (1 × 101 °]ml) were m i x e d in a s m a l l t u b e a n d 20 to 50/zl o f t h i s m i x t u r e were a p p l i e d to a T r y p t o n e s e m i s o l i d p l a t e (0-3% agar). U s i n g a P a s t e u r p i p e t w i t h its t i p d r a w n o u t to a fine capillary, t h e i n o c u h t m w a s a p p l i e d as a t h i n vertical wall t h a t p e n e t r a t e d t h e e n t i r e d e p t h o f t h e agar. (a) S t r a i n ST8 (mot-recA-) in t h e a b s e n c e o f t r a n s d u c i n g p h a g e . R e v e r t a n t s w a r m s are e v i d e n t b u t no trails are p r e s e n t . (b) S a m e as (a) b u t m i x e d w i t h t r a n s d u c i n g p h a g e g r o w n on wild t y p e . Trails e x t e n d o u t 5 to 10 m m f r o m t h e i n o c u h u n . (c) S t r a i n ST20 (che-recA +) in t h e a b s e n c e of p h a g e . T h e g r o w t h is m u c h wider a r o u n d t h e i n o c u l u m b e c a u s e t h i s s t r a i n is motile. (d) S a m e as (c) b u t m i x e d w i t h wild t y p e p h a g e . A d e n s e b a c k g r o t m d o f g r o w t h h a s r e s u l t e d f r o m t h e m a n y c o m p l e t e t r a n s d u e t a n t s t h a t h a v e s w a r m e d a n d m u l t i p l i e d . (e) S t r a i n ST25 (che-recA-) in t h e a b s e n c e of p h a g e . (f) S a m e as (e) b u t m i x e d w i t h wild t y p e p h a g e . N u m e r o u s s h o r t diffuse trails are evident,.
CHEMOTAXIB MUTANTS O F S . T Y P H I M U R I U M
~31
response, similar in duration to t h a t of SL4041. The other three mutants gave no detectable response; however, a short response of 30 seconds or less would not have been detected in those experiments. (d) Complemen~ation of che mutants analyzed by abortive transduction
of recombination-defective recipients Abortive transduetion has been used for complementation analysis of Salmonella mot and fla mutants (Stoeker et al., 1953; Lederberg, 1956; Stoeker, 1956; Iino, 1958) and E. coli mot and the mutants (Armstrong & Adler, 1967,1969). Use of abortive transduction with mot and fla mutants is straightforward. An inoenlum consisting of a mixture of recipient (mutant 1) and donor phage (mutant 2) is streaked on Tryptone semisolid agar. The abortive transductants, i.e. those recipients containing non-integrated donor DNA, are motile if mutations 1 and 2 are in separate genes but non-motile ff they are in the same gene. I f mutations 1 and 2 are in separate genes, motile transductants will swim away from the inoculum, leaving behind them a dense trail of non-motile daughter cells. I f mutations 1 and 2 are in the same gene, no trails are formed. Lederberg (1956) showed t h a t trail formation requires chemotaetic migration of the motile transductants away from the inoculum site. This suggests that abortive transduction could also be used for complementation analysis of chemotaxis mutants; however, Armstrong & Adler (1969) encountered difficulties in trying to observe trails b y abortive transduction between E. coli chemotaxis mutants. These difficulties arose from two sources. (a) Since the recipient is motile, the inoeulum tends to spread out more t h a n in the case of a mot- f l a - recipient. This spreading is apparently sufficient to interfere with the formation or detection of trails. (b) Since the nonchemotactie daughter cells t h a t mark the trail axe motile, t h e trails are necessarily diffuse, unlike the dense compact trails arising from abortive transduction of nonmotile recipients. The difficulty in detecting these diffuse colonies is compounded by a background of growth resulting from complete transductants t h a t swarm rapidly throughout the agar, Armstrong et a[. attempted to circumvent these problems b y using che- mot- or the- fla- double mutants as recipients; however, this technique is necessarily limited to analyzing only the the genes t h a t lie within a single transducing fragment of the fla and mot genes. I n order to do complementation analysis of our Salmonella che mutants, an alternative method of abortive transduetion was used t h a t gives well-defined trails and is not limited to analyzing the genes t h a t are near the motility region. The technique is essentially the same as t h a t described for mot or fla mutants but with two important modifications. First of all the the- recipients were made recombination defective b y introducing a recA mutation. ~The recA mutation prevents complete transduetion from occurring and thus eliminates the background growth t h a t interferes with trail detection. As a second modification, the mixture of recipient and donor phage was imbedded in the agax rather than being spread along the surface. This minimizes spreading of the inoculum. The results of abortive transduction tests with these modifications are sho~a in Plate III. The first test was made with a mot- strain. I t was feared t h a t the recombination defect might result in extremely short, perhaps imperceptible trails, because recombination-defective strains are less viable than normal strains. As shown in Plate I I I (a) and (b), however, transdfiction ;of the mot-recA- strain g a v e long,
232
D. ASWAD AND D. E. K O S H L A N D J R
Recipients t~
W+
+
p 222 p51
p54
+
+
+
+
+
+
+
+
+
-
-
. . . .
+
+
+
+
+
-
-
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
-
+
+
+
+
+
+
. . . .
pTI
-
q62
+
+
q90
+
+
. . . .
+
+
q71
+
+
r 5,]
+
+
+
+
r 57
+
+
+
+
. -
q59
t
+
+
+
+
+
-
-
p56
o c o E3
. . . . +
p52
+
. . . .
+
+
+
+
+
+
.
+
+
+
+
t Ill
+
+
+
+
I 221
+
+
+
+
+
+
+
-
+
-
.
.
.
r 60
-
+ -
-
-
-
-
-
s 58
+
+
+
+
+
u70
+
+
+
+
+
+
+
u 205
+
÷
+
+
+
+
Fro. 3. Complementation of Salmonella mutants. The mutation the-Ill(t) in this Figure is the same mutation found in SL4041 isolated by Vary & Stocker (1973). A plus (+) sign indicates t h a t t r a i l s w e r e f o r m e d a n d a m i n u s ( - - ) sign indicates that no trails at all were evident.
well-defined trails against a clear background. The test was then tried on a chestrain. I f the recipient is recA +, Plate I I I (c) and (d), only a dense background of growth is observed. I f the recipient is recA - , Plate I I I (e) and (f), diffuse trails are easily detected against the clear background. I f transducing phage grown on a strain containing the same che mutation as the recipient are used, no trails are formed, thus indicating t h a t trail formation is not an artifact due to phage infection. Results of abortive transduction tests among 18 m u t a n t s representing each of the six m u t a n t types found in this study are given in Figure 3, and a s u m m a r y of the properties of each class is given in Table 2. Complementation among the p, q and r groups suggests t h a t each m a y represent a separate gene. Two cases of apparent intragenic complementation were found among the p m u t a n t s ; donor che-51 (p) × recipient che-222 (p) and donor the-54 (p) × recipient che-51 (p). I n both cases, however, the complementation was non-reciprocal. Non-reciprocal complementation was noted by Armstrong et al. (1967) among E. cell mot m u t a n t s and they explained it as a gene dosage effect. Abortive transductants have only one copy of the donor gene b u t as m a n y as three copies of the host gene. Tiffs explanation seems quite adequate. Mutants of class t appear to be a well-defined group, except for one anomolous case, donor che-57 (r) × recipient the-221 (t), which did not show complementation. This leaves open the possibility t h a t t and r m u t a n t s lie within the same cistron. I f this
CHEMOTAXIS
MUTANTS
TYPHIMURIUM
OF ~.
233
TABLe. 2
CharacSeristics of Salmonella chemotaxis mutants Number t found
Complementation group
D o m i n a n t or recessive
33 8 11 1 1 3
p q r s t u
Recessive Recessive Recessive Dominant Recessive Dominant
Motility
Responds to positive temporal gradient $
Smooth Smooth Smooth Smooth Unco-ordinated Unco-ord/nated
---Yes No
t This Table includes only 57 of the 72 m u t a n t s isolated in this study, because there were insufficient d a t a on the remaining 15 m u t a n t s to place t h e m accurately in a eomplementation group. Complementation d a t a for 18 of these 57 classified m u t a n t s is given in Fig. 3. The remaining 39 m u t a n t s were classified on the basis of partial complementation data. All 39 were first tested (as recipients) against phage lysates from two p m u t a n t s . Those giving no trails with either p lysate, b u t giving trails w h e n t r e a t e d with a t least one q lysate, were classified as t y p e p. The remaining m u t a n t s were t h e n tested against two q lysates, a n d so on, until t h e classification
of all 39 mutants was established. :~ A positive temporal gradient of 0 to 1 mM-serine was used for this test.
is the case, then the complementation that does occur between t and r mutants, e.g. donor che-52 (r) × eke-221 (t), could be due to intrageuic complementation. S and t mutants showed apparent dominance in crosses with all donors, including wild type. Because of the nature of the complementation test, however, we cannot distinguish between full or partial dominance for these mutants. It is possible that the dominance exhibited by the s and/or u mutants is due primarily to a gene dosage effect. S and u mutants do not necessarily define unique genes. They may simply represent a particular type of mutation in one of the other genes.
4. Discussion (a) The liquid column technique The search for generally non-ehemotactic bacterial mutants is aided by the existence of an efficient reproducible method for rapidly isolating this type of mutant. The effective separation of generally non-chemotactic mutants from both wild type and non-motile bacteria is achieved by floating a narrow band of mutageuized wild type bacteria in a vertical column of liquid medium. The separation is based on three effects. First, generally non-chemotactic mutants (especially smooth-swimming ones) show a marked tendency to swim towards the bottom of the separation tube. Second, non-motile bacteria show little or no migration away from the inoculum band. Third, wild type bacteria are trapped in the upper region of the column by an appropriately placed attractant gradient. The bottom portion of the column, enriched for chebacteria, is recovered, grown to turbidity and subjected to three additional enrichment cycles by the same procedure. Approxlmately 40~/o of the final culture consists of highly motile the- bacteria. The final separation of che- from the remaining nonmotile and wild type bacteria is facilitated by visualizing single colonies on tryptone
234
D. ASWAD
AND
D. E. KOSHLAND
JR
semisolid medium, since wild type give large swarming colonies, non-motile give small colonies with well-defined boundaries, and non-chemotactic give small colonies with less well-defined boundaries. A major advantage of the technique is that the separation is achieved in a relatively short time, about 30 to 60 minutes and does not require growth of the bacteria. Thus, this technique may be adaptable to isolating temperature-sensitive mutants in which the mutation simultaneously affects both chemotaxis and some vital metabolic function. This may be done by growing the bacteria at a permissive temperature, followed by an enrichment for che- mutants at a non-permissive temperature. Such mutants could not be detected by screening individual colonies in tryptone semisolid agar, because the ability to swarm on these plates depends on both chemotaxis and growth. The vertical column method does not require growth during the enrichment process. (b) Properties of the generally non.~hemotactio mutant8 A total of 72 generally non-chemotactic mutants of S. typhimurium were isolated in this study. Of these, 68 were smooth-swimming (exhibiting no spontaneous tumbring) and four were unco-ordinated (exhibiting constant tumbling). An unco-ordinated mutant ofS. tyThimuriura has been reported by Vary & Stocker (1973). In a study of 172 E. coli che mutants, Parl~in~on (1974) found 146 smooth mutants and 26 uncoordinated mutants. The fact that we obtained a much smaller proportion of unco-ordinated mutants than did Parldnson may be caused by our selection procedure, which is biased in favor of bacteria having a long mean free path. We did not obtain any J~e mutants with normal motility. If such mutants are possible, we probably would have found some, since their mean free path would be considerably longer than that of an unco-ordiuated mutant. The fact that none were isolated strongly suggests that any mutation that leads to loss of general chemotactic ability must necessarily do so by altering the spontaneous rate of tumbling. Parkinson (1974) has independently reached the same conclusion. Parkinson (1974) has postulated possible functions for the four identified the genes of E. coli. Two genes are involved with generating tumbles, a third is located at the base of the flagellum where it receives information from the tumble generator and a fourth is involved in transmitting information about serine and certain repellent gradients to the tumble generator. The finding of six types of Salmonel/a che mutants leaves open the possibility that the chemotactic machinery of that organism is more complex than that of E. eoli. This would not be too surprising in light of the fact that the fla system of Salmonella, which exhibits the unique property of phase variation, is more complex than in E. co//(Iino, 1959). The largest difference between our results and those of Parblnson (1974) has to do with the grouping of unco-ordinated mutants. Parlrln,on (1974) finds that all mutations producing the unco-ordinated phenotype fall within a single gene. Moreover, many of his smooth mutants fall within this same gene. I f the chemotactic machinery of Salmonel/a is similar to that of E. co//, then our t and u mutants should be in the same cistron as one of the smooth-swimmlng mutant classes. Our complementation data are not inconsistent with this possibility. A more extensive analysis of the the genes in Salmonella will be necessary before it is possible to determine the exact number and grouping of genes involved, or even if the limit has been reached.
CHEMOTAXIS MUTANTS O F S . T Y P H I M U R I U M
235
This research was supported by grants from the National Institutes of Health (AM90765) and the National Science Foundation (GB7057). REFERENCES Armstrong, J. & Adler, J. (1967). (~enet~cs, 56, 363-373. Armstrong, J. & Adler, J. (1969). Genetics, 61, 61-66. Armstrong, J., Adler, J. & Dahl, M. (1967). J . Bacteriol. 93, 390-398. Aswad, I). & Koshland, D. E., Jr (1974). J . Bacteriol. 118, 640-645. Aswad, I). & Koshland, D. E., Jr (1975). J . Mol. Biol. 97, 207-223. Berg, H. & Brown, D. (1972). Nature (London), 239, 500-504. Branham, J. (1966). Biol. Bull. 131, 251-260. Dahlquist, F., Lovely, P. & Koshland, D. E., Jr (1972). Nature N e w Biol. 236, 120-123. Hazelbauer, G., Mesibov, R. & Adler, J. (1969). Prec. Nat. Acad. Sci., U.S.A. 64, 13001307. Iino, T. (1958). A n n u . Rep. Nat. Inst. Genet. (Japan), 9, 96. Iino, T. (1959). Bacteriol. Rev. 33, 454-475. Iino, T. & Oguchi, T. (1972). A n n u . Nat. Inst. Genet. (Japan), 22, 13-14. Lederberg, J. (1956). Genetics, 41, 845-871. Macnab, R. & Koshland, D. E., Jr (1972). Prec. Nat. Acad. Sci., U.S.A. 69, 2509-2512. Okada, T., Homma, J. & Sonohara, H. (1962). J . Bacteriol. 84, 602-604. Parkinson, J. (1974). Nature (London), 252, 317-319. Roth, J. R. (1970). Methods i n Enzymology, 57, 3-35. Sanderson, K. (1967). Bacteriol. Rev. 31, 354-373. Stocker, B. (1956). J . Gen. Microbiol. 15, 575-598. Stocker, B., Zinder, N. & Lederberg, J. (1953). J . Gen. Microbiol. 9, 410-433. Vary, P. & Stocker, B. (1973). Genetics, 73, 229-245. Vogel, H. & Bonner, I). (1956). J. Biol. Chem. 218, 97-106. Note added in proof; Subsequent work with mapping and deletionmutants has indicated that the six complementation groups do represent separate genes and further mutant selection suggests that there are three additional genes making a total of nine (Warriek, H., Taylor, B. L. & Koshland, D. E., Jr, manuscript in preparation).
Isolation, Characterization and Complementation of Salmonella typhimurium Chemotaxis Mutants DANA ASWADt AND D. E. KOS~rr.AwD, JR
Department of Biochemistry Univer, ity of California Berkeley, Calif. 94720, U.S.A. (Beceived 21 Oaober 1974) A novel technique has been developed for selecting non-chemotactic mutants of
Salmonella ~jphimurium that are defectivo in transmission of information from the chemoreceptor to the flagella. This technique utilizes a preformed gradient in a vertical column of liquid medium, which separates wild types, non-motile and motile but non-chemotactie into three regions. In the first application of the technique, 71 such mutants were obtained, of which 67 exhibited smooth motility and four had unco-ordinated motility. Complementation analysis, done by abortive transduction of recombination-defective recipients, indicates that there may be as many as six genetically distinct classes of ehe mutants in Salmonella. 1. I n t r o d u c t i o n One of the major advantages of using bacterial chemotaxis as a model for sensory reception is that one can perform sophisticated genetic manipulations of the molecular machinery that controls this behavior. Numerous mutants defective in chemotactic behavior have already been isolated in Escherichia coli and Salmonella typhimurium (Armstrong et al., 1967; Hazelbauer et al., 1969; Iino & 0guchi, 1972; Vary & Stocker, 1973; Parkinson, 1974). Adler and co-workers (Armstrong et al., 1967; Hazelbauer et al., 1969) have divided non-chemotactic mutants into two types: specific, which results in loss of ability to respond towards specific attractants or repellents; and genera], which cannot respond to any attractants or repellents. Armstrong & Adler (1969) studied 38 generally non-chemotactic mutants of E. coli and found three complementation groups, all of which mapped near the nw~ and fla region of the chromosome. Parkinson (1974) has recently reported a fourth complementation group. Chemotactie migration results from a selective suppression of spontaneous tumbling when the bacterium is swimming in a favorable direction (Macnab & Koshland, 1972; Berg & Brown, 1972). The control of tumbling frequency is therefore the means b y which the bacteria migrate up or down a gradient and the control t h a t is defective in generally non-chemotactic mutants. A total of 85% of the E. coli mutants t h a t were isolated by Parl~inson could not tumble spontaneously, while the remaining 15 ~ exhibited incessant tumbling with no periods of smooth motility. Similar mutants t Present address : Department of Biology, University of California, Los Angeles, Calif. 90024, U.S.A. 225
HfrB2 rectA22 recA1 strA
dhuA m o t
Wild type
metE1926
dhuA mot recA1
che-222(p) metE1926
hisF8786 thyA1981
che-222(p) recA1
trpA8 hisC527 (amber) che-lll (amber)
TA1859
ST1
ST2
ST8
ST20
ST23
ST25
SL4O41
Genotype
DB43
Strain
Derived from ST20 in two steps. First the mete was transduced with phage grown on wild type to make the strain rose +. The f e c a l was then introduced by conjugation with DB43 Vary & Stocker (1973).
Generally non-chemotactic, smooth motility. Recombination defective
Generally non-chemotactie. Unco-ordinated, frequent tumbling. Responds to a steep temporal gradient of attract.ant
ST20
che- mutants isolated in this study, except strain
Derived from a mutagenized culture of ST1 by penicillin selection (for his-) and trimethoprim selection (for thyA - )
Diethylsulfate mutagenesis of ST2, then screening of non-swarming isolates for motile cells (see Results)
TA1859 made thy- by trimethoprin selection, then f e c a l inserted by conjugation with strain DB43
Wild type ohemotaxis. This strain is the parent for all
Generally non-chemotaotio Smooth motility
Flagellated but non-motile Recombination defective
Diethylsulfate mutagenesis of ST1, then penicillin selection for met-
Selected for good serine taxis from LT2 strain of B. Ames
Normal chemotaxis and motility Parent for ST20
B. Ames.
B. Ames, Dept of Biochemistry, University of Calif., Berkeley
Source
Flagellated but non-motile
Used to transfer reoA mutation to other strains for use as recipients in abortive transduction
Relevant properties
Genotype, relevant properties a n d sources o f bacterial 8trains
TABLE 1
CHEMOTAXIS MUTANTS
OF E. T Y P H I M U R I U M
in ~ typhimurium h a v e been reported (Vary & Stocker, 1973; I i n o &
227
Oguehi,
1972). I n pursuing our studies on the transmission of information in the chemotactic system (Aswad & Koshland, 1974,1975)i it was desirable to isolate generally nonehemotactic m u t a n t s in the 8almoneUa system. Since the determination of the genes identified with the chemotactic machinery and their function requires the generation of a large n u m b e r o f mutants, a rapid and efficient isolation procedure was needed. A new method was devised utilizing a preformed a t t r a c t a n t gradient in a liquid column in order to separate generally non-chemotactic m u t a n t s from b o t h wild t y p e and non-motile bacteria. Using this method we were able to isolate a large n u m b e r of non-chemotactic m u t a n t s in a relatively short time. The m u t a n t s thus isolated were subjected to motility p a t t e r n and complementation analysis and fell into six distinct classes. 2. M a t e r i a l s a n d M e t h o d s
(a) Bacter/a The bacterial ~trains, all derived from ~. t/yphimurium LT2, that were used in developing the che mutant isolation and complementation procedures are described in Table 1. Media and growth conditions have been described elsewhere (Aswad & Koshland, 1974). (b) Oenetio techniques Diethylsulfate mutagenesis, penicillin selection and phage transduetion of ,.%lmonella were as described in detail by Roth (1970). P22-int3t, a non-lysogenizing mutant of phage P22, obtained from B. Ames, was used for all transduetions. Trimethoprin selection of thyA mutants has been described by Okada e~ a/. (1962). Transfer of ~ec.4 from strain DB43 into recipient strains was accomplished in the following manner. A 0.2-ml sample of thyA recipient (overnight nutrient broth culture) and 0.5 ml of DB43 (overnight nutrient broth culture) were inoculated into separate tubes containing 10 nd of nutrient broth and incubated for 3 to 4 h at 37°C. One ml of DB43 and 2 ml of recipient were then mixed and filtered and the filter was placed "face up" on a tryptone semisolid agar plate at 37°C for 90 rain, then into 1 ml of VBC medium (minimal salts plus citrate medium Of Vogel & Bonner, 1956) in a small tube and mixed vigorously for 1 rain to disengage mating bacteria. A 0-2-ml sample was then spread on minimal glucose plates and incubated for 24 h. Ten of the smallest colonies were streaked on nutrient agar to obtain single colony isolates. These were grown individually in nutrient broth and tested for sensitivity to u.v. light b y allowing 1 drop to run across the surface of a nutrient agar plate, then exposing half of the plate to a 15 W germicidal lamp (General Electric) at a distance of 3 3 c m for 7 s. After overnight incubation at 37°C recA- strains exhibited growth only on the non-irradiated half of the inoculum, while wild t y p e grew over the entire inoeulum.
3. Results (a) Attempts to isolate generaUy non.chemotadic mutants by negative swc~rmselectio~ A t t e m p t s .were m a d e to isolate generally non-chemotactic m u t a n t s of Salmonella b y the methods of Armstrong e~ al. (1967). A thick suspension of diethylsulfate mutagenized bacteria, strain ST2, was placed as a small spot in t h e center o f a t r y p t o n e semisolid agar plate. During an overnight incubation, the wild t y p e bacteria swarmed outwards from the inoculum, leaving a dense population at the center. A t Gene symbols in this paper are those used by Sanderson (1967).
228
D. ASWAD
AND D. E. KOSHLAND
JR
sample was taken from the center and subjected to a similar selection. After ten such serial selections, samples were grown in nutrient broth and plated at a density of about 50 cells per plate on Tryptone semisolid medium (0.35/o agar). Two hundred non-swarming colonies were picked, grown in nutrient broth and screened for motility by examination with a microscope. Only one culture with vigorous motility was found; the rest were probably mot and fla mutants. The single ohe mutant that was found, strain ST20, had smooth motility with no spontaneous tumbling. The dif~culty encountered with the swarm selection technique is probably due to the fact that the gradients that are necessary to draw the wild type away from the inoculum site do not exist prior to placing the inoculum on the plate. The gradients are established only after the bacteria at the inoculum site havo had time to metabolize a significant amount of certain nutrients in the tryptone that serve as attractants. Wild type, non-motile, and non-chemotactic bacteria will all multiply at equal rates before these gradients are established. By the time the gradients are established, the inoculum is so dense with wild type, that separation is difficult. Moreover, the use of semisolid medium inhibits movement of the bacteria and makes the separation rate much slower than it could be in a liquid medium. (b) The preformed liquid gradient technique To overcome the difficulties of the semisolid agar technique, it seemed reasonable to try a method (a) that used a preformed gradient; (b) that used a liquid rather than a semiaolid medium and (c) that created conditions that would separate non-motile from non-chemotactic bacteria, as well as non-motile from wild types. The approach used is shown in Figure 1. By placing the bacteria in a narrow band at the edge of a step gradient it was assumed that the wild type bacteria would diffuse only into the region of the attractant because they were under strong chemotactic pressure. The non-chemotactic bacteria would not respond to the gradient but would diffuse further than the non-motile ones in a symmetrical manner above and below the original band. Many non-chemotactic mutants could thus be found in the lower region where neither non-motile nor wild type individuals would be expected. To test this technique, cultures of a non-chemotactic mutant (strain ST20), a non-motile mutant (strain TA1859) and wild type (strain ST1) were independently tested in separate experiments using the abote apparatus. The distribution of the bacteria was measured at various time intervals after layering them in the gradient and the results are shown in Figure 2. As predicted, the non-motile strain did not show any significant change in distribution during the 60-minute incubation period. The wild type strain responded to the gradient by moving up the tube slightly. The che- strain behaved in a quite unexpected manner. Instead of spreading symmetrically in both directions from the band, it showed a marked preference for swimming down the tube. The asymmetrical diffusion of the the- smooth swimmer is apparently due to the influence of gravity. A similar phenomenon has been observed with sperm (Branham, 1966). Branham's explanation for this was that the sperm head had a higher specific gravity than the tail and this caused them to assume a head-down, tail-up orientation. An alternative explanation is that the slow sedimentation of the bacteria causes the flagella to swing behind the soma due to the hydrodynamic drag on the flagella. Once ~he bacterium is properly oriented, its forward velocity is sufficient to propel it downward at a rate of 0.17 em per minute (calculated from data given by Macnab & Koshland, 1972).
Pump
1
Grsdient maker
0-2
7
I
[Serine ] (raM) 0'4 0-6 0.8 I I
1.0
6Serine
F///A
///l.,f
Attractant
,////I / ///.4
N
\ Bacteria ~
4
\ ",,
Bacteria
3
~-,,,=,~ 2
\ ",
I
Glycerol '\
0
1.0
2.0 3.0 Glycerol % ( w / v )
Fzo. 1. Preformed liquid gradient used for isolation of ehemotaxis m u t a n t s . The drawing on t h e left indicates how t h e gradient w a s constructed. T h e drawing o n t h e r i g h t shows t h e dist r i b u t i o n of t h e selection t u b e contents a t t h e beginning of t h e e n r i c h m e n t period. All solutions were m a d e u p in minimal m e d i u m a n d p u m p e d t h r o u g h a stainless steel inlet pipe (no. 20 gauge hypodermic needle) t h a t e x t e n d e d to t h e b o t t o m of t h e 18 m m × 150 m m glass tube. F i r s t 2.0 ml of 1.0 mM-serine ( a t t r a c t a n t ) was p u m p e d in, followed b y 2.5 ml of 1.0 m~-serine, containing glycerol, The glycerol concentration was initially 0 % (w/v) a n d increased linearly to 1.0% a t t h e e n d of t h e 2.5-ml p u m p i n g segment. Next, 0.2 m l of bacterial suspension (4 × 101 b a c t e r i a / m l in 0.5 m l of a solution of 0-5 mM-serine, 1"5% glycerol) was p u m p e d in, followed b y ;2"5 m l glycerol, which increased linearly from 2-0~/o t o 3.0%. The glycerol gradient was necessary to stabilize t h e a t t r a c t a n t g r a d i e n t against convection. Top
w ÷
real
che -
-
(smooth)
== .F_
o
"*--..
30 rain
~
Bottom
min
rain
\ Relative concentration of bacteria
FIG. 2. Behavior of wild t y p e (w +, s t r a i n ST1), non-motile (mot-, s t r a i n TA1859) a n d motile b u t non-ohemotectic ( t h e - , s t r a i n ST20) b a c t e r i a in a vertical step gradient of a t t r a c t ~ n t . The selection a p p a r a t u s is described in Fig. 1. The distribution of t h e b a c t e r i a was m e a s u r e d b y lights c a t t e r i n g , using t h e a p p a r a t u s described b y D a h l q u i s t e~ a/. (1972). The distributions were recorded a t 6 min a n d 30 m i n after introducing t h e bacteria. 16
230
D. ASWAD
AND
D. E. KOSHLAND
JR
To establish that this explanation and not some repellent action of the gradient was responsible, the orientation of the step gradient was reversed, i.e. the attractant was placed in the bottom third of the tube and the bacteria layered at its top edge. With this configuration the non-chemotactic bacteria again diffused predominantly into the bottom region, showing that gravity and not repellent action was responsible. The gravity effect was an unexpected boon, since it meant that a more dramatic difference in migration would be obtained than that based solely on random diffusion of the non-chemotactic mutants. A reconstruction experiment was done in which a mixed culture with a ratio of one the- bacterium (ST20) per 1000 wild type bacteria was introduced into the liquid column apparatus. After 60 minutes of incubation, all but the bottom 1-0 ml of the tube contents were aspirated off and the walls of the tubes were rubbed with ethanol-soaked cotton to kill bacteria that might be adhering. A 2.0-ml volume of sterile VBC medium was added to the tube, which was then capped and incubated for 24 hours, a t which time turbidity appeared. The relative frequency of chebacteria increased 11-fold after one cycle of selection. This proeedure was then applied to a mutagenized culture of strain ST23 (wild type chemotaxis). After mutagenizing with diethylsulfate, the culture was enriched for 6he- mutants in the liquid column apparatus four times in succession, allowing for turbid growth of the culture between each enrichment. The final culture was plated out on tryptone semisolid medium to screen for non-swarming colonies. Since tryptone contains a mixture of many attractants, only generally non-chemotactie and nonmotile mutants will fail to swarm. Mutants defective in a specific class of at%ractants should not be isolated by this screening. Plate I shows that two distinct colony morphologies were present: tiny compact colonies, typical of 0/~- and non-motile strains, and large swarming colonies, typical of che- strains. Among the tiny colonies it was possible with the aid of a dissecting microscope to distinguish between nonmotile and non-chemotactic colonies, because the latter had less well-defined boundaries. The relative frequencies of the three colony types were 10% wild type (large colonies), 50% non-motile (small colonies with sharp boundaries) and 40% non-chemotactic (small colonies with fuzzy boundaries). (c) Motility and behavior of Salmonella che mutants A total of 71 che mutants were isolated from a single mutageuized culture as described above. These 71 mutants, together with the single mutant isolated as described in section (a), above, brings the total number of mutants isolated in this study to 72. Although all these mutants were highly motile, none of them exhibited a normal frequency of spontaneous tumbling. Sixty-eight exhibited only smooth co-ordinated motility with no evidence of tumbling when examined in either minimal medium or nutrient broth. The remaining four mutants had unto-ordinated motility indistinguishable from that of strain SIA041, an unee-ordinated mutant of Salmonella reported by Vary & Stocker (1973). Examples of these two motility types are shown in Plate II. We have previously demonstrated that derivatives of strain SL4041 exhibit smooth motility in response to a temporal gradient of 0 to 1 mM-serine (Aswad & Koshland, 1974). The four unco-ordinated mutants isolated in this study were subjected to a similar temporal gradient of serine to determine whether they were capable of exliibiting a smooth response. One of the muta~lts, che-223(e), showed a definite smooth
:PLATE I. Detection of ohemotaxis m u t a n t s on t r y p t o n e semisolid agar. Samples of 30/~1 or less containing 30 to 50 bacteria were spread on t r y p t o n e semisolid plates containing 0.5% agar. This agar concentration, which is twice t h a t normally used to observe chemotaetic swarms, prevents wild type colonies from spreading and overlapping the che- colonies. After spreading the bacteria, the plates were incubated at 37°C for 5 h to dry the surface of the agar. They were t h e n transferred to a 30°C room for overnight incubation. (a) Strain ST1, a wild type. (b) Strain ST20, a c h ~ - smooth-swimming m u t a n t . (c) :From a culture of ST23 t h a t h a d been mutagenized and selected 4 times in succession by the liquid column method. This plate shows a mixture of wild type, che- and non-motile cells and is typical of those plates used in screening for chemutants. The non-motile and c h e - colonies are not distinguishable in this photograph b u t we were able to accurately distinguish the two types by close examination of colony morphology under moderate magnification. [facing ~. 230
tp
g
t~
,
b
.i
PT.A~E I I . T y p i c a l m o t i l i t y t r a c k s of s m o o t h a n d u n e o - o r d i n a t e d t y p e m u t a n t s isolated in t h i s sl~udy. (a) Che-222 (p), a s m o o t h m u t a n t . (b} Che-221 (t), a n u n e o - o r d i n a t e d m u t a n t . T h e t r a c t s are p r o d u c e d b y u s i n g a s t r o b o s c o p i c l i g h t source in (lark field w i t h a t i m e e x p o s u r e o f 4 s ( M a e n a h & K o s h l a n d , 1972).
PLATE I I I . C o m p l e m e n t a t i o n o f mot a n d che m u t a n t s u s i n g r e c o m b i n a t i o n - d e f e c t i v e recipients. Six d r o p s of a full-grown n u t r i e n t b r o t h c u l t u r e of r e c i p i e n t b a c t e r i a anti 3 d r o p s of d o n o r p h a g e (1 × 101 °]ml) were m i x e d in a s m a l l t u b e a n d 20 to 50/zl o f t h i s m i x t u r e were a p p l i e d to a T r y p t o n e s e m i s o l i d p l a t e (0-3% agar). U s i n g a P a s t e u r p i p e t w i t h its t i p d r a w n o u t to a fine capillary, t h e i n o c u h t m w a s a p p l i e d as a t h i n vertical wall t h a t p e n e t r a t e d t h e e n t i r e d e p t h o f t h e agar. (a) S t r a i n ST8 (mot-recA-) in t h e a b s e n c e o f t r a n s d u c i n g p h a g e . R e v e r t a n t s w a r m s are e v i d e n t b u t no trails are p r e s e n t . (b) S a m e as (a) b u t m i x e d w i t h t r a n s d u c i n g p h a g e g r o w n on wild t y p e . Trails e x t e n d o u t 5 to 10 m m f r o m t h e i n o c u h u n . (c) S t r a i n ST20 (che-recA +) in t h e a b s e n c e of p h a g e . T h e g r o w t h is m u c h wider a r o u n d t h e i n o c u l u m b e c a u s e t h i s s t r a i n is motile. (d) S a m e as (c) b u t m i x e d w i t h wild t y p e p h a g e . A d e n s e b a c k g r o t m d o f g r o w t h h a s r e s u l t e d f r o m t h e m a n y c o m p l e t e t r a n s d u e t a n t s t h a t h a v e s w a r m e d a n d m u l t i p l i e d . (e) S t r a i n ST25 (che-recA-) in t h e a b s e n c e of p h a g e . (f) S a m e as (e) b u t m i x e d w i t h wild t y p e p h a g e . N u m e r o u s s h o r t diffuse trails are evident,.
CHEMOTAXIB MUTANTS O F S . T Y P H I M U R I U M
~31
response, similar in duration to t h a t of SL4041. The other three mutants gave no detectable response; however, a short response of 30 seconds or less would not have been detected in those experiments. (d) Complemen~ation of che mutants analyzed by abortive transduction
of recombination-defective recipients Abortive transduetion has been used for complementation analysis of Salmonella mot and fla mutants (Stoeker et al., 1953; Lederberg, 1956; Stoeker, 1956; Iino, 1958) and E. coli mot and the mutants (Armstrong & Adler, 1967,1969). Use of abortive transduction with mot and fla mutants is straightforward. An inoenlum consisting of a mixture of recipient (mutant 1) and donor phage (mutant 2) is streaked on Tryptone semisolid agar. The abortive transductants, i.e. those recipients containing non-integrated donor DNA, are motile if mutations 1 and 2 are in separate genes but non-motile ff they are in the same gene. I f mutations 1 and 2 are in separate genes, motile transductants will swim away from the inoculum, leaving behind them a dense trail of non-motile daughter cells. I f mutations 1 and 2 are in the same gene, no trails are formed. Lederberg (1956) showed t h a t trail formation requires chemotaetic migration of the motile transductants away from the inoculum site. This suggests that abortive transduction could also be used for complementation analysis of chemotaxis mutants; however, Armstrong & Adler (1969) encountered difficulties in trying to observe trails b y abortive transduction between E. coli chemotaxis mutants. These difficulties arose from two sources. (a) Since the recipient is motile, the inoeulum tends to spread out more t h a n in the case of a mot- f l a - recipient. This spreading is apparently sufficient to interfere with the formation or detection of trails. (b) Since the nonchemotactie daughter cells t h a t mark the trail axe motile, t h e trails are necessarily diffuse, unlike the dense compact trails arising from abortive transduction of nonmotile recipients. The difficulty in detecting these diffuse colonies is compounded by a background of growth resulting from complete transductants t h a t swarm rapidly throughout the agar, Armstrong et a[. attempted to circumvent these problems b y using che- mot- or the- fla- double mutants as recipients; however, this technique is necessarily limited to analyzing only the the genes t h a t lie within a single transducing fragment of the fla and mot genes. I n order to do complementation analysis of our Salmonella che mutants, an alternative method of abortive transduetion was used t h a t gives well-defined trails and is not limited to analyzing the genes t h a t are near the motility region. The technique is essentially the same as t h a t described for mot or fla mutants but with two important modifications. First of all the the- recipients were made recombination defective b y introducing a recA mutation. ~The recA mutation prevents complete transduetion from occurring and thus eliminates the background growth t h a t interferes with trail detection. As a second modification, the mixture of recipient and donor phage was imbedded in the agax rather than being spread along the surface. This minimizes spreading of the inoculum. The results of abortive transduction tests with these modifications are sho~a in Plate III. The first test was made with a mot- strain. I t was feared t h a t the recombination defect might result in extremely short, perhaps imperceptible trails, because recombination-defective strains are less viable than normal strains. As shown in Plate I I I (a) and (b), however, transdfiction ;of the mot-recA- strain g a v e long,
232
D. ASWAD AND D. E. K O S H L A N D J R
Recipients t~
W+
+
p 222 p51
p54
+
+
+
+
+
+
+
+
+
-
-
. . . .
+
+
+
+
+
-
-
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
-
+
+
+
+
+
+
. . . .
pTI
-
q62
+
+
q90
+
+
. . . .
+
+
q71
+
+
r 5,]
+
+
+
+
r 57
+
+
+
+
. -
q59
t
+
+
+
+
+
-
-
p56
o c o E3
. . . . +
p52
+
. . . .
+
+
+
+
+
+
.
+
+
+
+
t Ill
+
+
+
+
I 221
+
+
+
+
+
+
+
-
+
-
.
.
.
r 60
-
+ -
-
-
-
-
-
s 58
+
+
+
+
+
u70
+
+
+
+
+
+
+
u 205
+
÷
+
+
+
+
Fro. 3. Complementation of Salmonella mutants. The mutation the-Ill(t) in this Figure is the same mutation found in SL4041 isolated by Vary & Stocker (1973). A plus (+) sign indicates t h a t t r a i l s w e r e f o r m e d a n d a m i n u s ( - - ) sign indicates that no trails at all were evident.
well-defined trails against a clear background. The test was then tried on a chestrain. I f the recipient is recA +, Plate I I I (c) and (d), only a dense background of growth is observed. I f the recipient is recA - , Plate I I I (e) and (f), diffuse trails are easily detected against the clear background. I f transducing phage grown on a strain containing the same che mutation as the recipient are used, no trails are formed, thus indicating t h a t trail formation is not an artifact due to phage infection. Results of abortive transduction tests among 18 m u t a n t s representing each of the six m u t a n t types found in this study are given in Figure 3, and a s u m m a r y of the properties of each class is given in Table 2. Complementation among the p, q and r groups suggests t h a t each m a y represent a separate gene. Two cases of apparent intragenic complementation were found among the p m u t a n t s ; donor che-51 (p) × recipient che-222 (p) and donor the-54 (p) × recipient che-51 (p). I n both cases, however, the complementation was non-reciprocal. Non-reciprocal complementation was noted by Armstrong et al. (1967) among E. cell mot m u t a n t s and they explained it as a gene dosage effect. Abortive transductants have only one copy of the donor gene b u t as m a n y as three copies of the host gene. Tiffs explanation seems quite adequate. Mutants of class t appear to be a well-defined group, except for one anomolous case, donor che-57 (r) × recipient the-221 (t), which did not show complementation. This leaves open the possibility t h a t t and r m u t a n t s lie within the same cistron. I f this
CHEMOTAXIS
MUTANTS
TYPHIMURIUM
OF ~.
233
TABLe. 2
CharacSeristics of Salmonella chemotaxis mutants Number t found
Complementation group
D o m i n a n t or recessive
33 8 11 1 1 3
p q r s t u
Recessive Recessive Recessive Dominant Recessive Dominant
Motility
Responds to positive temporal gradient $
Smooth Smooth Smooth Smooth Unco-ordinated Unco-ord/nated
---Yes No
t This Table includes only 57 of the 72 m u t a n t s isolated in this study, because there were insufficient d a t a on the remaining 15 m u t a n t s to place t h e m accurately in a eomplementation group. Complementation d a t a for 18 of these 57 classified m u t a n t s is given in Fig. 3. The remaining 39 m u t a n t s were classified on the basis of partial complementation data. All 39 were first tested (as recipients) against phage lysates from two p m u t a n t s . Those giving no trails with either p lysate, b u t giving trails w h e n t r e a t e d with a t least one q lysate, were classified as t y p e p. The remaining m u t a n t s were t h e n tested against two q lysates, a n d so on, until t h e classification
of all 39 mutants was established. :~ A positive temporal gradient of 0 to 1 mM-serine was used for this test.
is the case, then the complementation that does occur between t and r mutants, e.g. donor che-52 (r) × eke-221 (t), could be due to intrageuic complementation. S and t mutants showed apparent dominance in crosses with all donors, including wild type. Because of the nature of the complementation test, however, we cannot distinguish between full or partial dominance for these mutants. It is possible that the dominance exhibited by the s and/or u mutants is due primarily to a gene dosage effect. S and u mutants do not necessarily define unique genes. They may simply represent a particular type of mutation in one of the other genes.
4. Discussion (a) The liquid column technique The search for generally non-ehemotactic bacterial mutants is aided by the existence of an efficient reproducible method for rapidly isolating this type of mutant. The effective separation of generally non-chemotactic mutants from both wild type and non-motile bacteria is achieved by floating a narrow band of mutageuized wild type bacteria in a vertical column of liquid medium. The separation is based on three effects. First, generally non-chemotactic mutants (especially smooth-swimming ones) show a marked tendency to swim towards the bottom of the separation tube. Second, non-motile bacteria show little or no migration away from the inoculum band. Third, wild type bacteria are trapped in the upper region of the column by an appropriately placed attractant gradient. The bottom portion of the column, enriched for chebacteria, is recovered, grown to turbidity and subjected to three additional enrichment cycles by the same procedure. Approxlmately 40~/o of the final culture consists of highly motile the- bacteria. The final separation of che- from the remaining nonmotile and wild type bacteria is facilitated by visualizing single colonies on tryptone
234
D. ASWAD
AND
D. E. KOSHLAND
JR
semisolid medium, since wild type give large swarming colonies, non-motile give small colonies with well-defined boundaries, and non-chemotactic give small colonies with less well-defined boundaries. A major advantage of the technique is that the separation is achieved in a relatively short time, about 30 to 60 minutes and does not require growth of the bacteria. Thus, this technique may be adaptable to isolating temperature-sensitive mutants in which the mutation simultaneously affects both chemotaxis and some vital metabolic function. This may be done by growing the bacteria at a permissive temperature, followed by an enrichment for che- mutants at a non-permissive temperature. Such mutants could not be detected by screening individual colonies in tryptone semisolid agar, because the ability to swarm on these plates depends on both chemotaxis and growth. The vertical column method does not require growth during the enrichment process. (b) Properties of the generally non.~hemotactio mutant8 A total of 72 generally non-chemotactic mutants of S. typhimurium were isolated in this study. Of these, 68 were smooth-swimming (exhibiting no spontaneous tumbring) and four were unco-ordinated (exhibiting constant tumbling). An unco-ordinated mutant ofS. tyThimuriura has been reported by Vary & Stocker (1973). In a study of 172 E. coli che mutants, Parl~in~on (1974) found 146 smooth mutants and 26 uncoordinated mutants. The fact that we obtained a much smaller proportion of unco-ordinated mutants than did Parldnson may be caused by our selection procedure, which is biased in favor of bacteria having a long mean free path. We did not obtain any J~e mutants with normal motility. If such mutants are possible, we probably would have found some, since their mean free path would be considerably longer than that of an unco-ordiuated mutant. The fact that none were isolated strongly suggests that any mutation that leads to loss of general chemotactic ability must necessarily do so by altering the spontaneous rate of tumbling. Parkinson (1974) has independently reached the same conclusion. Parkinson (1974) has postulated possible functions for the four identified the genes of E. coli. Two genes are involved with generating tumbles, a third is located at the base of the flagellum where it receives information from the tumble generator and a fourth is involved in transmitting information about serine and certain repellent gradients to the tumble generator. The finding of six types of Salmonel/a che mutants leaves open the possibility that the chemotactic machinery of that organism is more complex than that of E. eoli. This would not be too surprising in light of the fact that the fla system of Salmonella, which exhibits the unique property of phase variation, is more complex than in E. co//(Iino, 1959). The largest difference between our results and those of Parblnson (1974) has to do with the grouping of unco-ordinated mutants. Parlrln,on (1974) finds that all mutations producing the unco-ordinated phenotype fall within a single gene. Moreover, many of his smooth mutants fall within this same gene. I f the chemotactic machinery of Salmonel/a is similar to that of E. co//, then our t and u mutants should be in the same cistron as one of the smooth-swimmlng mutant classes. Our complementation data are not inconsistent with this possibility. A more extensive analysis of the the genes in Salmonella will be necessary before it is possible to determine the exact number and grouping of genes involved, or even if the limit has been reached.
CHEMOTAXIS MUTANTS O F S . T Y P H I M U R I U M
235
This research was supported by grants from the National Institutes of Health (AM90765) and the National Science Foundation (GB7057). REFERENCES Armstrong, J. & Adler, J. (1967). (~enet~cs, 56, 363-373. Armstrong, J. & Adler, J. (1969). Genetics, 61, 61-66. Armstrong, J., Adler, J. & Dahl, M. (1967). J . Bacteriol. 93, 390-398. Aswad, I). & Koshland, D. E., Jr (1974). J . Bacteriol. 118, 640-645. Aswad, I). & Koshland, D. E., Jr (1975). J . Mol. Biol. 97, 207-223. Berg, H. & Brown, D. (1972). Nature (London), 239, 500-504. Branham, J. (1966). Biol. Bull. 131, 251-260. Dahlquist, F., Lovely, P. & Koshland, D. E., Jr (1972). Nature N e w Biol. 236, 120-123. Hazelbauer, G., Mesibov, R. & Adler, J. (1969). Prec. Nat. Acad. Sci., U.S.A. 64, 13001307. Iino, T. (1958). A n n u . Rep. Nat. Inst. Genet. (Japan), 9, 96. Iino, T. (1959). Bacteriol. Rev. 33, 454-475. Iino, T. & Oguchi, T. (1972). A n n u . Nat. Inst. Genet. (Japan), 22, 13-14. Lederberg, J. (1956). Genetics, 41, 845-871. Macnab, R. & Koshland, D. E., Jr (1972). Prec. Nat. Acad. Sci., U.S.A. 69, 2509-2512. Okada, T., Homma, J. & Sonohara, H. (1962). J . Bacteriol. 84, 602-604. Parkinson, J. (1974). Nature (London), 252, 317-319. Roth, J. R. (1970). Methods i n Enzymology, 57, 3-35. Sanderson, K. (1967). Bacteriol. Rev. 31, 354-373. Stocker, B. (1956). J . Gen. Microbiol. 15, 575-598. Stocker, B., Zinder, N. & Lederberg, J. (1953). J . Gen. Microbiol. 9, 410-433. Vary, P. & Stocker, B. (1973). Genetics, 73, 229-245. Vogel, H. & Bonner, I). (1956). J. Biol. Chem. 218, 97-106. Note added in proof; Subsequent work with mapping and deletionmutants has indicated that the six complementation groups do represent separate genes and further mutant selection suggests that there are three additional genes making a total of nine (Warriek, H., Taylor, B. L. & Koshland, D. E., Jr, manuscript in preparation).
