Multiple large DNA molecules of Azospirillum

PDF ( 2 MBs ) ( .pdf )

MULTIPLE LARGE DNA MOLECULES OF Azospirillum


BY

ALVIN G. WOOD
























A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL
OF THE UNIVERSITY OF FLORIDA IN
PARTIAL FULFILLMENT OF THE REQUIREMENTS
FOR THE DEGREE OF DOCTOR OF PHILOSOPHY



UNIVERSITY OF FLORIDA


1982











ACKNOWLEDGEMENTS

The author would like to thank his committee members,

Drs. Dennis E. Duggan, Francis C. Davis, Jr., Philip J.

Laipis, William W. Hauswirth and L.O. Ingram for their

valuable suggestions and criticisms.

He would also like to extend special thanks to

Dr. William B. Gurley for help with the hybridization

experiments and for the use of his laboratory equipment

and supplies.










TABLE OF CONTENTS


ACKNOWLEDGEMENTS


ABSTRACT

CHAPTER I


CHAPTER II


CHAPTER III


PAGE

ii

Siv


OCCURRENCE OF MULTIPLE LARGE DNA MOLECULES
IN Azospirillum AND A METHOD FOR ISOLAT-
ING THEM ON AGAROSE GELS


Introduction .
Materials and Methods
Results .
Discussion....


ACRIDINE ORANGE-INDUCED MUTATIONS
IN Azospirillum 31

Introduction 31
Materials and Methods . . . 32
Results. 33
Discussion. 37


Nif GENE HYBRIDIZATION STUDIES
OF Azospirillum DNA MOLECULES.


Introduction .
Materials and Methods . .
Results and Discussion. .

APPENDICES

A NORMALIZATION OF GEL MOBILITY DATA

B REGRESSION LINE CALCULATIONS

LITERATURE CITED .


BIOGRAPHICAL SKETCH


. 47
. 48
. 51



. 58

. 59

. 60


. 47


. 67












Abstract of Dissertation Presented to the Graduate Council
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy


MULTIPLE LARGE DNA MOLECULES OF Azospirillum

by
Alvin G. Wood

May 1982

Chairman: Dennis E. Duggan
Major Department: Microbiology and Cell Science



Six strains of Azospirillum brasilense and two of

A. lipoferum were found to harbor as many as eight different-

sized circular DNA molecules ranging from 45 to 1500 mega-

daltons. Identification and separation of these very large

molecules were achieved by gently lysing bacteria in the wells

of vertical agarose gels, subjecting the lysate to electro-

phoresis at 2 mA for 6 h, and then continuing electrophoresis

at 15-30 mA for an additional 12-48 h. Optimal recovery

required lysis at 40C in the presence of ribonuclease. The

technique has been used to isolate large DNAs from other

bacteria, including the chromosomes of Escherichia coli

and Agrobacterium tumefaciens.

Several types of mutants were isolated from acridine

orange-treated cultures of A. lipoferum and A. brasilense.

Mutants displaying increased sensitivity to cadmium and









unable to grow on carbon-free media or on ethanol were all

found to have lost a specific plasmid. One of these strains

was shown to have suffered deletions in most of its remaining

DNA molecules. A mutant unable to grow on N2 or reduce

acetylene was isolated from the multiply-deleted strain, but

its DNA molecules showed the same electrophoretic mobilities

as those of its parent strain. Methionine-requiring auxo-

trophs, isolated at a high frequency from A. lipoferum cul-

tures, also possessed DNA molecules with unaltered mobilities.

Attempts were made to determine which Azospirillum DNA

molecule includes the genes controlling nitrogen fixation by

hybridizing a labeled recombinant probe to Southern blots

of wild type and mutant DNA molecules. The limited success

acheived with this technique indicates that the structural

genes for nitrogenase are carried on the largest Azospirillum

DNA molecule.
















CHAPTER I
OCCURRENCE OF MULTIPLE LARGE DNA MOLECULES
IN Azospirillum
AND A METHOD FOR ISOLATING THEM ON
AGAROSE GELS

Introduction

The genus Azospirillum comprises Gram-negative, free-living,

nitrogen-fixing bacteria found in association with roots of

cereal crops and tropical forage grasses (19). Field experiments

conducted at the University of Florida showed higher yields

of dry matter in Azospirillum-inoculated pearl millet and

guinea grass than in uninoculated controls (64). More

recently, Azospirillum inoculation has been reported to

enhance corn yields in Israel (46).

The potential agronomic value of this association has

prompted studies of carbon and nitrogen metabolism in Azo-

spirillum (2, 21, 27, 28, 38, 43, 44, 45, 48, 49, 51), and

the scores of strains isolated from various parts of the

world have been grouped, on the basis of DNA homology and

biochemical characteristics, into two species, A. lipoferum

and A. brasilense (67). Very little is known, however,

concerning the genetics of Azospirillum (26, 41, 55). In

particular, no system of genetic transfer exists which would

permit location of the genes controlling nitrogen fixation

and facilitate studies of their expression.

1










Our interest in developing a system of genetic transfer

for Azospirillum led us to examine several wild type strains

for the presence of plasmids. Our initial attempt to identify

plasmid DNA in Azospirillum (9) involved dye bouyant density

ultracentrifugation of alkaline denatured lysates (61) and

direct visualization of plasmids in the satellite bands by

electron microscopy (18). Open circular (OC) DNAs of various

contour lengths were seen, but the apparent multiplicity of

molecules in each strain and their large sizes relative to

the plasmid chosen as a size standard (ColE1) made it difficult

to accurately assess the numbers and sizes of plasmids in

any given strain. However, plasmids with molecular weights

in excess of 300 megadaltons (Mdal) did appear to be present

in several Azospirillum strains.

This observation prompted us to try two electrophoretic

techniques specifically designed for the isolation of large

plasmids (10, 29), but neither of these permitted the isolation

of more than three plasmids from any Azospirillum strain.

We felt there was a strong possibility that very large

plasmids were present in these strains but that they were

being sheared during the mechanical manipulations, however

gentle, inherent in these procedures.

Therefore we adopted a method which is theoretically

the most gentle of all, that described by Eckhardt (22).

This technique differs fundamentally from other electrophoretic

techniques in that plasmid DNA is not extracted from cells









prior to electrophoresis. Rather, the bacteria are lysed

directly in the wells of the gel apparatus, resulting in

minimal nicking or breakage of covalently closed circular

(CCC) DNA. Using a modified version of this technique, we

have discovered the widespread occurrence of a multiplicity

of very large DNA molecules in strains of Azospirillum.

The present communication describes the electrophoretic

conditions necessary for the successful isolation of these

Azospirillum DNA molecules. Estimates of the sizes of the

molecules harbored by one Azospirillum strain are provided,

based on a comparison of their mobilities with those of

plasmids of known molecular weight. The very low mobilities

of some of the Azospirillum molecules suggest that they

represent small chromosomes rather than large plasmids.

Indeed, we have been able to isolate slowly-migrating DNA

bands from strains of Escherichia coli and Agrobacterium

tumefaciens, including two strains which do not harbor

plasmids. Evidence is provided indicating that the slowly-

migrating DNA bands isolated from Azospirillum represent CCC

DNA uncomplexed with protein.

Materials and Methods

Bacterial strains. Table 1-1 lists the Azospirillum

strains examined for plasmid content. Table 1-2 lists other

bacterial strains harboring plasmids of known molecular

weights used for construction of the standard size curve.









Growth conditions. All bacteria were grown to early

stationary phase prior to harvesting for electrophoresis.

Azospirillum and A. tumefaciens strains were grown in a

succinate/mineral salts medium (50) supplemented with 0.01%

yeast extract. Pseudomonas putida and E. coli strains were

grown in nutrient broth (Difco) supplemented with 0.01%

yeast extract. Growth temperatures were 35C for Azospirillum

and E. coli and 28C for A. tumefaciens and P. putida.

Plasmid isolation and agarose gel electrophoresis. We

used an electrophoretic method based on that described by

Eckhardt (22). The protocol outlined here includes modifications

found to be necessary for optimal, reproducible visualization

of the electrophoretic bands representing the largest DNA

molecules.

Vertical gels were cast with 0.6% or 0.7% molten agarose

(BioRad Standard Low -m ) in Tris/borate/EDTA electrophoresis

(E) buffer (40). The agarose was tempered at 420C for 20

min prior to casting the gel in order to minimize contracture

during solidification. The plastic comb used to form wells

had 16 teeth (13x9x1.5 mm). The gel was submerged in E

buffer and allowed to age at least 4 h at 4C prior to

removal of the comb. Spraying the comb lightly with PAM

(Boyle-Midway) prevented agarose from sticking to it during

removal.

Except where otherwise noted, wells were loaded in the fol-

lowing manner. One milliliter of cell culture was centrifuged

for one minute in a microcentrifuge (Fisher Model 235). The









supernatant solution was pipetted off using a vacuum aspirator

and the pellet was resuspended in 20-100 pl of 20% ficoll in

E buffer. Ten microliter aliquots of the cell suspensions were

added to wells preloaded with 15 1l of a solution containing

20% ficoll, 10 pg/ml lysozyme (Sigma), 100 pg/ml ribonuclease

A (Sigma), and 0.05% xylene cyanol FF (Kodak) in E buffer.

No attempt was made to mix the cell suspension with the

lysozyme solution inside the well. The cells were allowed

to interact with the lysozyme mixture at 40C for a minimum

of 30 min, and then 30 1 of 10% ficoll, 1% SDS in E buffer

was added, followed by 50 pl of 5% ficoll, 1% SDS in E buffer.

A current of 2 mA was applied for a minimum of 6 h, and

then the current was raised to 15-30 mA (50-100 V at 4C).

Electrophoresis was continued for 12-48 h depending on the

voltage used and the degree of molecular separation required.

The large capacity (2.5 1) of the buffer reservoirs of our

electrophoresis apparatus made recirculation of buffer un-

necessary.

Photography. Gels were stained for a minimum of 30 min

in 0.5 ug/ml ethidium bromide and visualized with either a

254 nm hand-held UV light (UV Products) or a 300 nm trans-

illuminator (Fotodyne). Photographs were taken through #4

and #29 Wratten filters, using Type 57 film (Polaroid).

Standard curve construction. Mobility data from 19 gels

were normalized to the mobility data of the gel illustrated

in Fig. 1-6 according to the method described by Hansen and

Olsen (29, Appendix A) except that absolute mobilities









(distances of plasmid migration from origin) rather than

relative mobilities (absolute mobilities divided by gel

lengths) were used in the calculations (see Appendix A).

The number of DNA bands in common between the normalized and

standard gels ranged from 7 to 15 with an average of 12.

The logarithm of the average normalized mobility of a given

standard plasmid was plotted against the logarithm of the

molecular weight of that plasmid, and a least squares regres-

sion line was calculated (see Appendix B). This regression

line was then used to estimate the molecular weights of the

Azospirillum molecules as well as the presumed chromosomes

of E. coli and A. tumefaciens.

Plasmid nomenclature. The molecule having the lowest

mobility in each Azospirillum strain has been designated

pAZ1; molecules are then numbered in order of increasing

mobility. All the molecules of a given strain are suffixed

with that strain designation. For example, the smallest

plasmid of Spl3t is pAZ6-Spl3t.

Results

Multiplicity of DNA molecules found in Azospirillum

strains. Figures 1-1 and 1-2 illustrate the electrophoretic

banding patterns obtained from eight geographically diverse

isolates of A. brasilense and A. lipoferum lysed in the

wells of vertical agarose gels. Each strain has a character-

istic array of DNA molecules of various mobilities, an

observation which can be used for purposes of identification.

Every strain harbors two molecules whose extremely low









mobilities suggest that they represent small chromosomes

rather than plasmids. Under optimal conditions, the recovery

of these DNAs is highly reproducible except for the small

plasmid which bands in the region of linear DNA in some gels

(Fig. 1-1, lane D; Fig. 1-2, lanes C and D).

A comparison of Figs. 1-1 and 1-2 indicates that very

long periods of electrophoresis are necessary in order to

achieve separation of all the DNA bands, in accordance with

what would be expected for very large DNAs. The increase in

resolution achieved by increasing electrophoresis time is,

unfortunately, accompanied by a tendency for the most slowly-

migrating bands to become faint or disappear altogether

(data not shown). This suggests, however, that the material

is those bands is fragile, presumably because of its high

molecular weight.

Effect of ultraviolet light on mobilities of JM125A,

molecules. Figure 1-3 compares the mobilities of UV-irradiated

and unirradiated molecules of JM125A2. The JM125A2 molecules

were isolated in the usual manner except that electrophoresis

was terminated after 4 h at 80 V. Blocks of agar were cut

from lanes of the gel, extending from the well to the position

of the tracking dye. One of the unstained agar blocks

(lane B) was subjected to a dose of approximately 3000 J/m2

of 254 nm UV, while the other (lane A) was untreated. This

dose should have been sufficient to introduce at least one

chain break into every CCC DNA molecule in the irradiated

gel (6, 7, 33, 56). Both treated and untreated agarose













blocks were then imbedded in a horizontal agarose gel and

subjected to electrophoresis for 8 h at 50 V. As indicated

in Fig. 1-3, UV irradiation converted the DNA molecules from

forms capable of movement through an agarose gel into forms

incapable of such movement. Presumably, this represents the

conversion of CCC DNA into OC DNA.

Effects of enzymatic treatments on DNA recovery. In

order to acquire information concerning the physical relation-

ships of the DNA molecules to other cellular components, the

roles of lysozyme, RNase, and protease (Sigma Type VI) in

optimal DNA recovery were assessed. Figure 1-4 shows that

neither the addition of protease to the standard cell

mixture nor the elimination of lysozyme had an appreciable

effect on DNA recovery or mobility. The elimination of

RNase, however, resulted in failure to recover pAZI and pAZ2

as well as poor recovery of pAZ3.

Effect of cell mass on DNA isolation. Figure 1-5 illus-

trates the result of an experiment in which cell suspensions

of JM125A2 and AT181 were serially diluted prior to loading

the wells. The smearing of the bands in lanes A and E

appears to be due to overloading. The use of only 107 cells

(lanes D and H) allowed visualization of all the DNA bands

in this experiment. However, in other experiments (data not

shown) using 107 cells resulted in very faint bands, particu-

larly for the smaller molecules. Optimal recovery was
usually achieved with 10 cells.
usually achieved with 10 cells.









Size estimates of DNA molecules. The mobilities of the

DNA molecules of JM125A2 in relation to plasmids of known

molecular weight and slowly-migrating DNAs of other bacteria

are illustrated in Figs. 1-6, 1-7, and 1-8. Based on Fig.

1-6, JM125A2 appears to harbor five molecules larger than

the largest standard molecule (pMG1, 312 Mdal). Recovery of

the larger JM125A2 molecules was poor in Figs. 1-6 and 1-7;

these are included primarily to show the slowly-migrating

DNAs isolated from E. coli (Fig. 1-7, lane F; Fig 1-3, lanes

E, G, and H) and A. tumefaciens (Fig. 1-7, lane A). The two

slowly-migrating DNAs recovered from A. tumefaciens have

been given the designations pXX1-AT181 and pXX2-AT181. The

isolation of these slowly-migrating DNAs is not completely

reproducible (Fig. 1-6, lanes E and F; Fig 1-7, lane E;

Fig. 1-8, lanes A and D). Indeed, the difficulty of isola-

ting the molecules appears, in our experience, to be inver-

sely related to mobility. Thus, our rate of success in

isolating slowly-migrating DNAs from Azospirillum is 90% or

better, but our success rate with the slowly-migrating

E. coli DNA has never exceeded 50%. We have never isolated

a slowly-migrating DNA band from either of the two Pseudo-

monas strains used in the present study. This failure may

be related to an observed tendency for these strains to lyse

prematurely.

Mobility data from 20 agarose gels were normalized and

used to construct a standard curve relating electrophoretic

mobility and molecular weight (Fig. 1-9). Since reports in

the literature had suggested that CCC DNAs larger than 80









Mdal (72) or 140 Mdal (29) migrate faster than predicted

from linear extrapolations of standard curves based on

smaller CCC DNAs, we initially calculated a regression line

not including pMG1 and pMG5. When this line was used to

estimate the sizes of the Azospirillum molecules, however,

it seemed impossible that the values obtained could be

underestimates. We therefore recalculated regression data

with the large Pseudomonas plasmids included, and again with

the slowly-migrating band from E. coli included and assigned

a molecular weight of 2800 Mdal (8,14,28). The three sets

of regression line estimates are summarized in Table 1-3.

Figure 1-9 is a graph of regression line B, chosen because

it includes only those molecules measured by electron micro-

scopic contour length.

Effect of voltage gradient on regression estimates. If

the larger molecules examined in the present study were

really migrating faster than predicted, this effect should

be more pronounced at higher than at lower voltages (24).

We were particularly interested in this possibility since,

in our attempts to optimize electrophoresis conditions, we

had used voltages ranging from 50 V to 100 V and we wished

to include data from as many gels as possible in our regression

line calculations. Figure 1-10 compares the mobilities of

five standard plasmids run at 50 V for 42 h with mobilities

of the same plasmids run at 100 V for 24 h. The two curves

are nearly parallel and neither displays a convincing change









of slope above 140 Mdal. Table 1-4 indicates that, for each

standard plasmid, the ratio of mobility at 100 V to mobility

at 50V is a constant value.

Discussion

Eckhardt first described in situ lysis of bacteria in

agarose gels as a rapid method for plasmid isolation (22).

We have sacrificed the rapidity of the technique but exploited

its gentleness in order to isolate very high molecular

weight CCC DNAs from Azospirillum and other bacteria.

Our initial excitement in isolating slowly-migrating

DNAs on agarose gels was tempered with concern that the low

mobilities might reflect an open circular nature or some

protein interaction rather than large size. The electro-

phoretic behavior of the JM125A2 molecules subsequent to UV

irradiation, however, strongly indicates that they are

covalently closed and supercoiled. A DNA/protein interaction,

while not strictly ruled out by the failure of protease to

alter the electrophoretic mobilities of the molecules, seems

unlikely in view of this result. Furthermore, such an

association would have to be resistant to dissociation by

the SDS which quickly migrates from the upper ficoll layers

down through the DNA-containing region of the gel during

electrophoresis. Thus, the slowly-migrating bands do not

appear to represent relaxation complexes of the type isolated

from plasmid-bearing strains of E. coli, which dissociate

yielding OC DNA when exposed to SDS or protease (13).









A comparison of the DNA molecules isolated from

A. brasilense strains Sp7 and Spl3t provides further evidence

that the slowly-migrating DNAs are not simply isomeric forms

of smaller plasmids. Since these two strains were isolated

from the same region of Brazil and display nearly identical

electrophoretic banding patters, there is a good possibility

that they are isogenic except for the occurrence of pAZ6-Sp7

in one strain. If so, none of the slowly-migrating DNAs of

either strain could represent an isomeric form of this

relatively small molecule.

The requirement of RNase treatment for isolation of the

larger Azospirillum DNAs suggests that, in their native

forms, these molecules are attached via RNA to some cellular

component in a manner precluding entry into the gel matrix.

It is also possible that the larger Azospirillum molecules

are attached to one another via RNA. Assuming the likelihood

that essential genes are carried on the largest two or three

molecules, some mechanism to ensure cosegregation of newly-

replicated molecules into daughter cells would appear to be

necessary. Molecules as large as pAZ1 and pAZ2 might further

be expected to exist inside the cell in condensed, folded

states. These considerations lead us to postulate that pAZ1

and pAZ2 (and perhaps pAZ3) are arranged in a chromosomal

structure closely resembling that believed to occur in

E. coli (54, 73, 74). The only difference between the two

"nucleoid" structures would be that in E. coli the RNA-

stabilized domains comprise a single, continuous DNA molecule,









whereas in Azospirillum these domains are divided into two

or three continuous DNA molecules. The remaining Azospi-

rillum molecules might form nonintegrative associations with

the Azospirillum nucleoid analogous to those described

between other large, stringently controlled plasmids and

their host chromosomes (34, 35).

We have attempted to estimate the sizes of the DNA

molecules of one Azospirillum strain (JM125A2) by comparing
their mobilities to those of plasmids whose sizes have been

calculated from electron microscopic contour length measure-

ments. A problem with this effort arose in that five of the

JM125A2 molecules migrated more slowly than pMG1, the lar-

gest standard plasmid available. Their sizes, therefore,

had to be estimated from a linear extrapolation of our

standard curve (Fig. 1-9) and so must be considered only

approximate.

Some investigators have cautioned against standard

curve extrapolations on the grounds that CCC DNAs larger

than 80 Mdal (72) or 140 Mdal (29) migrate faster than

predicted. For several reasons, this appears not to be the

case under our electrophoresis conditions. First, it is

difficult to believe that the calculated values for the

larger Azospirillum molecules could be underestimates.

Second, these estimates change only modestly when the Pseu-

domonas plasmids are disallowed or when the chromosome of

E. coli is included in the regression line calculation

(Table 1-4). Third, the hypothetical nonlinearity of the

standard curve should have been greater for a gel run at 100









V than for a gel run at 50V, but Table 1-4 indicates that

the relative mobilities of the standard plasmids were nearly

identical at the two voltages. Finally, a theoretical

justification for nonlinearity of standard curves for CCC

DNAs in the high molecular weight range has not been advanced.

The explanation offered for the fast mobilities of high

molecular weight linear DNAs, i.e., "end-on" migration (1,

25), would seemingly not apply to high molecular weight CCC

DNA.

From the estimated sizes of the Azospirillum molecules

and assuming one copy of each per cell, the full genetic

complement of DNA for these bacteria appears to be approximately

4.3x109 daltons, some 50% greater than the corresponding

value for E. coli (8, 14, 28). At present, we can only

speculate as to the reason for this discrepancy. The large

complement of DNA may simply reflect the metabolic diversity

of these bacteria; Azospirillum species are capable of

carrying out most of the known nitrogen transformations (19,

44, 45), can grow heterotrophically (49) or autotrophically

(60), and tolerate the full range of oxygen tensions from

fully aerobic (49) to anaerobic with nitrate as terminal

electron acceptor (45). Alternatively, some of the Azospirillum

DNA may be redundant. This redundancy, if it does occur,

could provide a basis for recombination among Azospirillum

DNA molecules, underlying a potential mechanism for the

evolution of new strains.

The application of the modified Eckhardt technique to

bacteria harboring size standard plasmids led to the discovery









that slowly-migrating DNAs could be isolated from species

other than Azospirillum. Molecules with apparent molecular

weights of 500 and 1800 Mdal were isolated from A. tumefaciens

AT181 along with the two previously described plasmids.

These four molecules may well represent the full genetic

complement of this A. tumefaciens strain since the sum of

their estimated sizes is 2.6x109 daltons. Hence, pXX1-AT181

may, in fact, represent the Agrobacterium chromosome. The

slowly-migrating DNA isolated from E. coli strains appears

to represent the E. coli chromosome since it displayed an

appropriate mobility and could be recovered from both plasmid-

harboring and plasmidless strains. Isolation of intact

E. coli chromosomes by ultracentrifugation of gently lysed

cells through neutral sucrose gradients has been described

by others (66, 73).

In summary, we have demonstrated that strains of

Azospirillum harbor unique arrays of large DNA molecules.

The probability that these molecules comprise the full

genetic complement of their host bacteria suggests that

Azospirillum should be considered a multichromosomal

prokaryote. Arrangement of genetic material in this fashion

contrasts sharply with the situation in E. coli, in which

more than 90% of the DNA is carried on a single large molecule.

The common assumption that the DNA of most prokaryotes is

arranged as it is in E. coli may reflect, to a certain

extent, the previous lack of a suitable protocol for isolating

CCC DNAs larger than 500 Mdal.















Table 1-1. Geographic origins of Azospirillum strains


Species Strain Isolated bya Placea Countrya

A. brasilense Sp7 J. Dobereiner Rio de Janeiro Brazil
Spl3t "
Sp84 "
Cd D. L. Eskew Riverside, CA USA
JM82A1 J. R. Milam Venezuela

JM125A2 Gainesville, FL USA
A. lipoferum SpUSA5b J. Dobereiner Pullmann, WA "
SpRG6xx Passo Fundo Brazil


from Reference 67













Table 1-2. Strains containing plasmids of known molecular weight

Species Strain Source Plasmid MWa (ref.)


Escherichia coli K12 C600 this lab RP4 36 (40)
X1254 F14 202 (47)
II' F 61 (47)
S' F14 A F 141 (47)

' JC182 B. Bachmann none
Escherichia coli C W. Klipp none
Pseudomonas putida PpS1239 J. Shapiro pMGl 312 (29)
' PpS1240 pMG5 280 (29)

Aqrobacterium
tumefaciens AT181 W. Gurley pTi-AT181 114 (16)
S" pAT-AT181 158 (16)


ain megadaltons












Table 1-3. Linear regression estimates
of sizes of DNA molecules

DNA molecule Aa Bb Cc


pAZ1-JM125A2
pAZ2-JM125A2
pAZ3-JM125A2
pAZ4-JM125A2
pAZ5-JM125A2
pAZ6-JM125A2
pAZ7-JM125A2
pMG1
pMG5
pTi-AT181
pAT-AT181
pXX1-AT181
pXX2-AT181
RP4
F
F14
F14 cF
E. coli chromosome


138025
1130180
60056
38530
34526
1286
466
28946
26035
11317
15419
1650160
47342
336
6410
20516
14115
3150530


1510290
1230190
63762
40432
36230
1316
466
30249
27038
11418
15819
1810180
50045
326
6410
21217
14416
3500610


1330240
1090170
58753
37928
34225
1306
477
28746
25935
11417
15518
1580150
46540
346
6610
20516
14215
2970490


abased on mobilities
than 160 Mdal
based on mobilities
based on mobilities
E. coli chromosome


of standard plasmids smaller

of all standard plasmids
of all standard plasmids plus


NOTE: Sizes are in megadaltons, 1s.d.















Table 1-4. Comparison of mobilities of standard
plasmids run at 50 V and 100 V

Plasmid Gel Aa Mobility (cm) Gel Bb B/A


pMG1 3.10 3.80 1.23
pMG5 3.20 3.90 1.22
pTi-AT181 3.90 4.75 1.22
pAT-AT181 3.50 4.35 1.24
RP4 5.35 6.80 1.27


a50 V for 42 h
100 V for 24 h
























A B C D E F G H


Fig. 1-1. Agarose gel electrophoresis of large DNA
molecules from eight Azospirillum strains of diverse
geographic origin. Electrophoresis was for 24 h at 80 V.
A. brasilense strains: (A) Sp7; (B) Spl3t; (C) Sp84;
D) Cd; (E) JM82A ; (F) JM125A. A. lipoferum strains:
G) SpUSA5b; (H) SpRG6xx.
















A B C D E F G H


Fig. 1-2. Increased separation of large Azospirillum
DNA molecules in an agarose gel subjected to electrophoresis
for 36 h at 80 V. A. brasilense strains: (A) Sp7;
(B) Spl3t; (C) Sp84; (D) Cd; (E) JM82A; (F) JM125Ao.
A. lipoferum strains: (G) SpUSA5b; (H) SpRG6xx.






























A B





















Fig. 1-3. Agarose gel showing the effect of UV light
on mobilities of DNA molecules from JM125A Lanes from a
preparative vertical gel run at 80 V for 4 h were excised
and either (A) not treated or (B) irradiated with 3000 J/m
254 nm UV. The agarose blocks were then embedded in a
horizontal gel and subjected to electrophoresis at 50 V for
an additional 8 h.














A B C D


Fig. 1-4. Effects of various enzymatic treatments on
recovery of DNA molecules from JM125A. A cell suspension
was prepared as usual and 10 ul aliqugts were added to
wells preloaded with (A) and (B) standard reaction mixture;
(C) standard reaction mixture lacking RNase; (D) standard
reaction mixture lacking lysozyme. After 1 h, 30 ul
20 mg/ml protease, 20 O/o ficoll in E buffer was added to
well (B). The reaction mixtures were then allowed to
interact an additional 1 h at 40C. Electrophoresis was
for 42 h at 80 V.


















B C D E F G H

I r V VWI


Fig. 1-5. Effect of cell number on isolation of DNA
molecules from (A-D) JM125A2 and (E-H) AT181. Electro-
phoresis was at 80 V for 24 h. (A) and (E)
3xl8 cells; (B) and (F) 108 cells; (C) and (G) 3xl07
cells; (D) and (H) 107 cells.
















A B C D E F
v F


Fig. 1-6. Agarose gel electrophoresis showing
mobilities of JM125Ap DNA molecules in relation to
mobilities of size standard plasmids. Electrophoresis
was at 90 V for 28 h. (A) PpS1239 (pMG1); (B) PpS1240
pMG5); (C) JM125A2; (D) AT181 (pTi,pAT); (E) C600 (RP4)
(F) X125 (F).


















D E
-0 13


Fig. 1-7. Agarose gel electrophoresis showing
slowly-migrating DNAs isolated from A. tumefaciens
and E. coli. Electrophoresis was for 24 h at 75 V.
(A) AT181 (pTi,pAT); (B) PpS1239 (pMG1); (C) PpS1240
(pMG5); (D) JM125A2; (E) c600 (RP4); (F) JC182.


A B C


















A B C D E F G H


Fig. 1-8. Agarose gel electrophoresis showing
recovery of slowly-migrating DNA from three of four
E. coli strains. The strains run in lanes (G) and (H)
do not harbor plasmids. Electrophoresis was for 24 h
at 60 V. (A) X1254; (B) PpS1239 (pMG1); (C) JM125A ;
(D) AT181 (pTi,pAT); (E c600 (RP4); (F) PpS1240 (pGS);
(G) E. coli C; (H) JC182.










































Fig. 1-9. Standard curve relating electrophoretic mobilities of size marker
plasmids and JM125A2 DNA molecules to their molecular weights. Horizontal lines
represent one standard deviation to either side of the average normalized mobility.














chromosome (3)


pXX1 (7)


pAZ2 (18)


! (20)
pXX2 (14)
^zpAZ4 (20)


3))





1000




o
0

g 300





100
a:


pTi (15)


F (8)


pAZ7 (20)


RP4 (12)


I I I I


I I


0.4 0.5 0.6 0.7 0.8

LOG NORMALIZED MOBILITY


0.9 1.0


pAZI (12


-pMG5 (16)


30--





M
















500


pMG1l
VMG5



pM


100




50


100 V


0.5 O.t 0.7 0.8 0.9

LOG MOBILITY (cm)





Fig. 1-10. Comparison of mobilities of standard
plasmids in gels run at different voltages. The 50 V
gel was run for 42 h and the 100 V gel for 24 h.













CHAPTER II
ACRIDINE ORANGE-INDUCED MUTATIONS IN Azospirillum

Introduction

Genetic material in the genus Azospirillum appears to

be arranged in a manner unusual for bacteria. Rather than

possessing a single large chromosome with or without acces-

sory DNA in the form of plasmids, Azospirillum strains

harbor unique arrays of up to eight different-sized co-

valently closed circular (CCC) DNA molecules, the largest

estimated to be 1500 megadaltons (Mdal) (see Chapter I).

The occurrence of so many molecules intermediate in size

between what is normally considered to be very large for a

plasmid (300 Mdal) and what would be considered a small

chromosome (1500 Mdal) makes the usual plasmid/chromosome

dichotomy less than obvious in these bacteria.

Plasmid DNA is generally defined as encoding functions

not essential for cell growth under usual conditions.

Growth of plasmid-harboring bacteria in the presence of

acridine orange (AO) often permits the isolation, at an in-

creased frequency, of strains which have been "cured" of the

plasmid (30). We therefore grew a representative strain of

A. lipoferum and one of A. brasilense in AO-containing

medium and tested individual colonies for a variety of

nonessential phenotypic traits. Since the capacity for









nitrogen fixation can be considered nonessential, the potentially

mutant colonies were tested for growth on nitrogen-free

medium. In addition, because of the blurring of the distinction

between plasmid and chromosome in these bacteria, all colonies

were screened for auxotrophy. Mutant colonies identified in

this manner were analyzed for DNA content by agarose gel

electrophoresis.

The present communication describes the successful cor-

relation of three phenotypic properties with the presence of

a specific plasmid in both Azospirillum strains examined.

Furthermore, we report the isolation of a mutant having

suffered multiple deletions in its DNA molecules and the

isolation of both Nif and auxotrophic strains for which no

obvious loss of DNA could be demonstrated.

Materials and Methods

Bacteria. A. brasilense JM125A2 and A. lipoferum

SpUSA5b were obtained from N. R. Krieg. Mutant strains

derived from them are listed in Table 2-1.

Media. The succinate/mineral salts medium described by

Okon et al. (50) was used for nitrogen-free growth. The

same medium supplemented with 0.5% (NH )2SO4 was used as

minimal agar (MA). Carbon-free agar (CFA) was MA lacking

succinate. Carbon source utilization tests were done on MA

with the appropriate carbon source replacing succinate at a

concentration of 1%. Sensitivities of bacteria to inhibitors

were determined on nutrient agar (Difco) supplemented with

0.01% yeast extract and filter-sterilized antibiotics or

separately-autoclaved heavy metal salts.










Curing. Bacteria were treated with AO as described by

Hirota (30), except that the pH was 8.0. The maximum concen-

tration of AO which consistently permitted growth was 2.5 g/ml.

Detection of DNA molecules. The electrophoretic method

used to detect the various DNA molecules present in wild

type and mutant strains has been extensively described

elsewhere (see Chapter I).

Results

Table 2-1 lists the types of mutants isolated from cul-

tures of A. brasilense JM125A2 and A. lipoferum SpUSA5b

grown in 2.5 ug/ml AO at pH 8.0 and the approximate fre-

quencies with which they were found to occur in such cul-

tures. No spontaneous mutants (from untreated cultures)

were found.

Isolation of JM125A, mutants with the T phenotype.

Originally, we screened JM125A2 colonies derived from AO-

treated cultures for a number of properties: 1) growth on

minimal agar plates containing arabinose, galactose, ethanol,

or butanol as sole carbon source; 2) sensitivity to tri-

methoprim (since wild type Azospirillum strains are resistant

(3)); and 3) growth on succinate/mineral salts media with and

without (NH4)2SO4. No auxotrophic, Nif Ara Gal or

trimethoprim-sensitive mutants were found. However, mutants

unable to grow on ethanol or butanol were isolated at a high

frequency (Table 2-1). Further testing showed that acetate

or acetaldehyde would support growth of the mutants.









Somewhat surprisingly, wild type colonies grew on the

medium (intended as a control) to which no carbon source had

been added. Growth on the carbon-free agar was slower and

less luxurious than growth on any of the media containing

added carbon. The mutant colonies failed to grow appreciably

on the carbon-free medium. Although it appeared from these

observations that Azospirillum was capable of autotrophic

growth and that the mutants had lost this ability, both wild

type and mutant strains produced a visible band of growth

when a small inoculum (100 cells) was introduced into the

carbon-free soft agar (0.02%) under a predominantly H2/CO2

atmosphere. Others have recently described autotrophic

growth of Azospirillum (60).

Since heavy metal ion resistance is a characteristic

often associated with plasmids in other bacteria (12, 42,

63, 71), we assessed the sensitivities of the mutant and

wild type strains to the salts of seven heavy metals. No

differential sensitivity to Ag, Cu2+, Co2+, Pb2+, Hg2+, or

Ni2+ was observed; however, 3x10-5 M Cd2+ permitted growth

of the wild type but not the mutant strains. Further analysis

showed the mutants to be 30-fold more sensitive than the

wild type to Cd2+ (Table 2-2). We have designated the "T"

phenotype to represent the triad of cadmium sensitivity,

inability to metabolize alcohols, and failure to grow on

carbon-free agar under ambient atmosphere.









Loss of pAZ4 determines the T phenotype in JM125A2.

Figure 2-1 shows the DNA molecules of wild type and mutant

JM125A2 strains fractionated on an agarose gel as described

(see Chapter I). All but one of the T mutants exhibited the

DNA banding pattern represented by 125-T2 (lane D), which is

identical to that of the wild type except for the absence of

pAZ4 (cf. lanes A and D). In contrast, mutant strain 125-T1

appeared to have lost not only pAZ4 but pAZ7 and parts of

pAZ2, pAZ3, pAZ5, and pAZ6 as well (lane B). Table 2-3

compares the estimated sizes of the 125-T1 DNA molecules to

those of wild type JM125A2.

We attempted to determine what functions, if any, could

be ascribed to the DNA missing in 125-T1 but not in 125-T2.

Wild type JM125A2 and its T derivatives displayed similar

sensitivities to UV light, and all reduced nitrate. As

indicated in Table 2-2, however, 125-T2 was 10-fold and

125-T1 was 100-fold more sensitive to kanamycin than was the

wild type. The sensitivities of the three strains to

streptomycin were identical.

Isolation of a Nif derivative of 125-T1. Despite re-

peated screening of colonies derived from nitrosoguanidine

(NG) or AO-treated JM125A2 cultures, we had never been able

to isolate a mutant with the Nif phenotype. The deletion

of approximately 800 Mdal of DNA in 125-T1 (Table 2-3)

suggested that at least 20% of the DNA in JM125A2 is non-

essential. This inference raised the possibility that our

failure to isolate Azospirillum Nif strains by conventional









mutagenesis might be due to occurrence of the nif genes in

more than one copy. Therefore, we screened 2000 colonies

derived from AO-treated 125-TI cultures for growth on

succinate/mineral salts medium with and without (NH4)2SO4.

One mutant was isolated which grew on the minimal medium

only when nitrogen was supplied. This strain (125-T1N1)

also failed to reduce acetylene even in the presence of low

concentrations of (NH4)2SO4 or yeast extract. The electro-

phoretic mobilities of its DNA molecules, however, could not

be distinguished from those of its parent strain (Fig. 2-1,

lanes B and C).

Mutants derived from AO-treated SpUSA5b cultures. We

wondered whether types of mutants similar to the ones iso-

lated from A. brasilense JM125A2 could be isolated from

A. lipoferum SpUSA5b. Only one of 2000 SpUSA5b colonies

derived from AO-treated cultures displayed the T phenotype;

thus, T mutants appear to occur at a somewhat lower fre-

quency in SpUSA5b than in JM125A2. An unexpected finding

was the high frequency of Met auxotrophs (Table 2-1); no

such mutants had been isolated from JM125A2.

Figure 2-2 illustrates the DNA banding patterns of SpUSA5b

and its mutant derivatives. As was the case with JM125A2,

the T phenotype correlated with loss of pAZ4 (lanes A and

B). However, we did not observe obvious differences in

electrophoretic mobility between the DNA molecules of wild

type SpUSA5b and those of a representative Met derivative

(lanes A and C).









125-T1N1 and USA-MI are nonreverting. To investigate

the possibility that 125-T1N1 and USA-MI arose from AO-induced

frameshift mutations, we attempted to isolate revertants of

these strains from untreated, AO-treated, and NG-treated

cultures. No true revertants were found. Slowly-growing
8
colonies did appear at a frequency of approximately 10 when

untreated 125-TIN1 was plated on nitrogen-free agar. However,

when restreaked on fresh nitrogen-free plates they failed to

flourish, and they were unable to reduce acetylene. Further-

more, their isolation frequency was enhanced to a much

greater extent by NG than by AO.

A similar situation occurred with potential Met revertants

of USA-Mi. Colonies arising on minimal agar grew when

restreaked on fresh minimal plates, but growth on all types

of media was slower than wild type growth. Like the pseudo-

revertants of 125-TIN1, their isolation frequency was in-

creased significantly by NG but only marginally by AO.

These Met strains also displayed a deep reddish pigmentation,

whereas colonies of our other SpUSA5b strains are peach

colored.

Discussion

The results show that pAZ4-JM125A2 and pAZ4-SpUSA5b
R +
determine, for their respective hosts, the CadR, Adh+ and

Cfa+ phenotypes. The two plasmids are not entirely homo-

logous, however, since the electrophoretic mobility of

pAZ4-SpUSA5b is slightly lower than that of pAZ4-JM125A2

(see Chapter I). Loss of pAZ4-JM125A2 at a high frequency









in AO-treated JM125A2 cultures is consistent with the known

curing effects of acridine dyes (30). On the other hand,

high-frequency AO-induced methionine auxotrophy and

AO-induced multiple deletions in DNA have not been

described.

The sensitivities of wild type and mutant Azospirillum

strains to cadmium are, interestingly enough, comparable to

the sensitivities exhibited by CadR and CadS Staphylococcus

aureus strains (11). Recently, cadmium resistance in

S. aureus has been shown to depend on an energy-dependent,

plasmid-encoded efflux system (53, 69). A similar system

may operate to confer cadmium resistance in Azospirillum.

Alternatively, the configuration of membrane proteins in

Azospirillum T mutants may result in a lesser degree of

thiol group shielding than occurs in the wild type (58).

Growth of the T mutants on acetate and acetaldehyde but

not on ethanol implies the existence of an alcohol dehydro-

genase encoded by pAZ4-JM125A2 and pAZ4-SpUSA5b. The basis

for the inability of the T mutants to flourish on carbon-free

agar under ambient atmosphere is less obvious. The defect

does not appear to be in the capacity for autotrophy per se,

as both mutant and wild type strains grow in carbon-free

soft agar under an autotrophic atmosphere. Biosynthesis of

ribulose diphosphate carboxylase in other bacteria has been

shown to be repressed under conditions of high oxygen tension

(37). Therefore, a single cell plated on carbon-free agar

under ambient atmosphere cannot immediately begin to grow









autotrophically; it must instead utilize some intracellular

carbon reserve to grow and divide several times, so that an

aerobic layer of cells is formed under which a micro-

aerophilic environment is created. Azospirillum strains

do, in fact, accumulate large amounts of poly-B-hydroxy-

butyrate (PHB), particularly (but not exclusively) when

grown in nitrogen-free media (49). Although the T mutants

as well as the wild type can be seen microscopically to

contain PHB granules, it is possible that they have

difficulty utilizing this storage polymer for growth.

The discovery that one of the T mutants (125-T1) had

suffered multiple deletions in its DNA molecules is an

observation which may be relevant to the proposed mode of

action of AO in plasmid elimination. Some investigators

have suggested that plasmids, owing to their relatively

small size, are more accessible to AO than is the bacterial

chromosome (52) or that AO selectively inhibits plasmid

replication (31, 32). In contrast, others have maintained

that AO causes nonspecific loss of DNA via inhibition of

polymerase I (4, 5); since only cells having lost nonessential

DNA survive, this nonspecific mode of action translates into

an apparent specificity for plasmid (i.e. nonessential) DNA.

Assuming that the generation of 125-T1 really was AO-mediated

and not merely a spontaneous event which happened to occur

in an AO-treated culture, our results tend to support the

idea of a nonspecific interaction of AO with DNA. If AO

specifically interacted with or inhibited replication of










plasmid DNA, one would not expect to observe partially

deleted plasmids unless those plasmids were capable of

dissociation into self-replicating component molecules. The

possibility of four DNA molecules undergoing dissociation of

this kind within the same cell seems remote.

Wild and mutant JM125A2 strains displayed three levels

of sensitivity to kanamycin, correlating with degree of DNA

loss. This tends to implicate two or more proteins or one

protein encoded by two or more loci (on at least two DNA

molecules) in the determination of kanamycin sensitivity

levels in Azospirillum. Furthermore, these proteins must

act specifically on kanamycin and not on aminoglycosides in

general, since the sensitivities of the various mutants to

streptomycin were identical. Several types of aminoglycoside

modifying enzymes have been described in other bacteria, but

the mechanism by which the modified antibiotic confers

resistance is not known (17). The mechanism of aminoglycoside

uptake by sensitive cells is also obscure (17).

The Met derivatives of SpUSA5b and the Nif derivative

of 125-T1 possess DNA molecules whose electrophoretic mobilities

cannot readily be distinguished from those of their parent

strains. These phenomena may be accounted for in one of

three ways. First, the mutants might have arisen via the

action of AO as a frameshift mutagen (15, 52). This possibil-

ity seems unlikely in view of the observed inability of AO

to promote reversions in the mutant strains. Furthermore,

the frameshift action of acridine dyes has been described










primarily in bacteriophage (15, 52); frameshift mutagenesis

of bacteria at the frequencies reported here has not been

described. A second possibility is that the mutants are

deleted for all or part of a molecule not identified by the

electrophoretic method used. This explanation also seems

rather unlikely since we have used the method to identify

molecules as large as the E. coli chromosome (see Chapter I).

The third, and most likely, possibility is that the mutants

carry deletions not large enough to lead to obvious differences

in mobility for the affected molecules. Thus, 125-T1N1

might have arisen by deletion of the entire nif cluster from

pAZ1-JM125A2 (1500 Mdal); the wild type and mutant pAZ1-

JM125A2 molecules, differing in molecular weight by only 1%,

would not be resolved by electrophoresis. We are currently

attempting to determine whether this is the case by hybridizing

a labeled nif probe to Southern blots (65) of fragmented

mutant and wild type molecules.
















Table 2-1. Azospirillum mutants isolated
after growth in acridine orange


Parent Strain Mutant Phenotypea Isolation
frequency


JM125A2

JM125A2
125-T1
SpUSA5b

SpUSA5b


125-T1 Adh-Cfa-CadSKanS2
125-T2 Adh-Cfa-CadSKanS1
125-T1N1 Adh-Cfa-CadSKanS2Nif-
USA-T100 Adh-Cfa-CadSKanS1

USA-M1 Met


5x10 4
5x10-3
5x10 4

5x10-4
5x10-3


aAbbreviations:
Adh = No growth on ethanol or butanol
Cfa = No growth on minimal medium lacking added carbon
CadS= Sensitive to 3x10 5 M CdCl2
KanS1= Sensitive to 3 mg/ml kanamycin
Kan 2= Sensitive to 0.3 mg/ml kanamycin
Nif = No growth on nitrogen-free medium; no acetylene
reduction
Met = Requires methionine for growth






43








Table 2-2. Sensitivities of
wild type and mutant Azospirillum strains
to cadmium, kanamycin, and streptomycin

Minimal inhibitory concentration

Strain [Cd2+],M [Kn], pg/ml [Sm],Ig/ml
-4 30 30
JM125A2 104 30 30
125-T1 3x10-6 0.3 30
125-T1N1 3x10-6 0.3 30
125-T2 3x10-6 3 30












Table 2-3. Estimated sizes
of DNA molecules
found in JM125A2 and 125-T1

Strain
Plasmid JM125A2 125-T1 deletion


pAZ1 1500 1500 -
pAZ2 1200 1100 100
pAZ3 640 430 210
pAZ4 400 400
pAZ5 360 320 40
pAZ6 130 95 35
pAZ7 46 46
total 4276 3445 831



Note: Sizes are in megadaltons. Sizes
of JM125A2 molecules were taken from
Table 1-3. 125-T1 sizes were deter-
mined from mobilities in Fig. 2-1.























A B C D


Fig. 2-1. Agarose gel electrophoresis showing
DNA molecules recovered from wild type and mutant
JM125A strains: (A) JMI25A2; (B) 125-T1; (C) 125-TlNl;
(D) 12 -T2. Electrophoresis was for 23 h at 90 V.




















A B C


Fig. 2-2. Agarose gel electrophoresis showing
DNA molecules isolated from wild type and mutant
SpUSA5b strains: (A) SpUSA5b; (B) USA-T100;
(C) USA-Mi. Electrophoresis was at 80 V for 31 h.












CHAPTER III
Nif GENE HYBRIDIZATION STUDIES
OF Azospirillum DNA MOLECULES

Introduction

In previous chapters we showed that strains of

Azospirillum brasilense and A. lipoferum harbor unique

arrays of large circular DNA molecules and that phenotypically

altered strains, some exhibiting a change in plasmid array,

could be isolated at a high frequency from cultures treated

with acridine orange. We now address the question of which

of these molecules determines the phenotype of greatest

general interest, i.e. nitrogen fixation. We had hoped that

the Nif- derivative of JM125A2 would show a change in the

mobility of one or more of its DNA molecules, but such was

not the case. We therefore decided to try a different

approach to the problem, based on molecular hybridization.

The structural genes for nitrogenase are thought to

have been either introduced recently in evolutionary history

into the various nitrogen-fixing bacteria or conserved to a

greater extent than other translated prokaryotic genes (59).

The basis for this view is the observation that Klebsiella

pneumoniae nif structural genes can hybridize to DNA from

all types of nitrogen-fixing prokaryotes, including both

Gram-negative and Gram-positive bacteria, Actinomycetes, and

Cyanobacteria (59). This interspecies homology has been

exploited to study the organization of nif genes in









blue-green algae (39). Although Azospirillum DNA has not

been examined for homology to Klebsiella nif DIIA, there is

no reason to think that Azospirillum would behave differently

from other nitrogen-fixing prokaryotes in this regard,

especially since the individual Azospirillum nitrogenase

subunit proteins can form active complexes in vitro with

complementary proteins from other nitrogen-fixing bacteria

(23).

We obtained, from W. Klipp (36), an Escherichia coli

strain harboring pWK27, a recombinant plasmid which includes

an EcoRI fragment carrying the nifK, nifD, and nifH genes

from K. pneumoniae. This molecule was isolated in large

quantity, labeled to high specific activity by nick trans-

lation (57), and hybridized to Southern blots (65) of gels

containing DNA molecules from wild type and mutant Azospirillum

strains. The very limited success we have achieved with

this technique indicates that pAZ1, the largest of the

Azospirillum molecules, carries the nif structural genes in

Azospirillum.

Materials and Methods

Bacterial strains. HB101(pWK27) was supplied by W. Klipp

(36). Origins of Azospirillum strains are described in

Table 1-1 (Chapter I) and Table 2-1 (Chapter II).

Isolation of pWK27. To one liter of cell culture grown

at 370C to Klett 90 was added 170 mg chloramphenicol. The

cells were incubated at 370C for 20 h, harvested, washed

with 10 mM NaC1, and resuspended in 6 ml 0.02 M EDTA, 0.025









M Tris, 0.9% glucose (pH 8.0). Lysozyme (12 mg) was added

and the cells were chilled on ice for 30 min. Next, 12 ml

0.08% NaOH, 0.8% SDS was added and the lysate was gently

swirled for 5 min. Following addition, with gentle mixing,

of 9 ml KAc (pH 4.8), the lysate was incubated on ice for 2

h. The lysate was then centrifuged at 15,000 rpm for 30 min

and the supernatant solution was transferred to a fresh

tube. The addition of PEG 6000 to a final concentration of

10% followed by incubation on ice for 2 h caused the plasmid

to precipitate and it was pelleted at 2,500 rpm for 5 min.

The pellet was resuspended in 2 ml 50mM Tris, ImM EDTA

(pH 8.0). Ribonuclease was added to a concentration of

5 g/ml and the solution was incubated at 37C for 30 min.

Following adjustment of the volume to 10 ml with 50 mM Tris,

ImM EDTA (pH 8.0), the solution was extracted twice each

with phenol/chloroform, chloroform, and ether. Residual

ether was blown off with air. The plasmid was then ethanol

precipitated, washed in 70% ethanol, and further purified by

dye bouyant density ultracentrifugation. Plasmid bands,

visualized with UV light, were removed from the centrifuge

tubes with a plastic syringe. Ethidium bromide was removed

by extraction with isopropanol and CsC1 by dialysis against

three changes (2 1 each) of 25 mM Tris, 1 mM EDTA.

Blotting. Agarose gels containing separated Azospirillum

DNA molecules were prepared as described in Chapter I. DNA

from the gels was blotted onto strips of nitrocellulose

according to the method of Wahl et al. (70). This is a









slight modification of Southern's original technique (65).

The gels are treated with 0.25 M HCI as an initial step in

order to partially depurinate the DNA and fragment it for

more efficient transfer.

Nick translation of pWK27. The probe was labeled with
32
P according to the method described by Rigby et al. (57).
32
One hundred microCuries Pa-dCTP was dried under vacuum in a

1.5 ml microfuge tube. To this was added 5 1p 0.5 M Tris, 0.1 M

B-mercaptoethanol, 0.05 M MgCl2; 24 1 of a 1:1:1 dNTP mix (200

mM each); 1 g pWK27 DNA; and water to make 48 p 1 total

volume. After 5 min at 300C, 1 p 1 diluted, activated DNase

I was added. (DNase was activated by diluting a 1 mg/ml

stock in 10 mM HC1 1:2000 into 10 mM Tris, 5 mM MgCl2, i

mg/ml BSA (pH 7.6)). After 2 min at 300C, 0.5 p1 DNA

polymerase I (5 U/ml) was added and the reaction was held

at 150C for 1 h, whereupon it was terminated by the addition

of 5 P1 0.25 M EDTA. The reaction mixture was extracted

with phenol and the labeled plasmid was separated from the

unincorporated nucleotides by passage through a Sepharose 4B

or Sephadex G-100 column or by electrophoresis through a

0.7% low melting point agarose gel.

Hybridization. All hybridizations and prehybridizations

were carried out in Sears Seal-n-Save plastic bags. Filters

were prehybridized for a minimum of 12 h at 65C in a solu-

tion consisting of 0.5% SDS, 20 g/ml denatured salmon sperm

DNA, 10x Denhardt's solution (20), 2.5x SSC, and 0.05 M

Na2HPO4 (pH 8.0). Hybridizations were carried out under the









same conditions except that Ix Denhardt's solution was used

and denatured, labeled pWK27 was included. The washing

protocol of Thomashow et al. was followed (68). This consists

of one 30 min wash with 2.5x SSC at room temperature followed

by four washes (30 min each) with 2.5x SSC at 650C and one

wash at 650C with 0.1x SSC.

Autoradiography. The dried filters were exposed to

X-ray film at -70oC in the presence of an intensifying

screen. Exposure times were 2 h for the control hybridi-

zation of pWK27 to its restriction fragments and a minimum

of 1 week for the Azospirillum hybridizations.

Results and Discussion

Table 3-1 summarizes the variations in protocol and

outcomes of five hybridization experiments. Both Fig. 3-2

and Fig 3-4 were taken from Experiment 2. Despite the

somewhat disappointing results, we can conclude from Fig.

3-4 that the Azospirillum nif structural genes are carried

on pAZl, the molecule which probably represents the chromo-

some of these bacteria.

Since the control hybridization (Fig. 3-2) gave such a

strong signal it is doubtful that our difficulty in detect-

ing Azospirillum nif sequences reflects a serious flaw in

experimental procedure. Rather, the problem seems to be one

of sensitivity; the autoradiogram illustrated in Fig. 3-4

required 8 days exposure and shows only faint bands of

hybridization. There would appear to be two explanations

for this low level of hybridization. First, assuming that






52


the Azospirillum nif genes are carried on pAZ1, less than

0.5% of the DNA in the band representing pAZ1 is capable of

hybridizing to the probe. Second, since there is some

degree of divergence in the DNA sequences of nif structural

genes in various bacteria (59), the probe DNA is not entirely

homologous to that portion of pAZ1 which is capable of

hybridization. DNA preparations from different strains of

Rhizobium hybridize with differing intensities to Klebsiella

nif DNA (59). We do not think that the low level of hybrid-

ization exhibited by Azospirillum DNA reflects poor transfer

from gel to nitrocellulose. Gels were always restained and

examined for DNA after transfer; none was ever found.

















Table 3-1. Summary of results of hybridization experiments

Specific activity Labeled Purification Method of Hybridi-
Expt. of probea nucleotide of probe DNA transfer nation?

1 1.6x108 dCTP Sepharose 4B Southern no
2 9.1x107 dCTP Sepharose 4B Southern yes
3 2.2x107 dCTP Sepharose 4B electroblot no
4 1.8x107 dCTP Sephadex G-100 electroblot no
5 2.2x107 dCTP electrophoresis Southern no

acpm per jg DNA




















A B C D E F G H I J


Fig. 3-1. Agarose gel electrophoresis of pWK27
digested with HindIII (D-F) or EcoRI (G-I). Some
undigested OC and CCC plasmid DNA is visible. Lanes
(A-C) and (J) show lambda HindIII fragments used as
size standards. Volumes of DNA solutions added to
wells were as follows: (A) 1 ul; (B,J) 2 ul; (C) 4 ul;
(D,G) 3 ul; (E,H) 6 ul; (F,I) 9 ul. Electrophoresis
was at 60 V for 4 h.



























A B C D E F G H I J









cAgo


Fig. 3-2. Autoradiogram of gel shown in Fig. 3-1.
Gel was blotted and hybridized to labeled pWK27 as
described in the text. Exposure was for 2 h.
(A-C,J) HindIII-digested lambda; (D-F) HindIII-digested
pWK27; (G-I) EcoRI-digested pWK27.
















A B C D E F G H


Fig. 3-3. Agarose gel electrophoresis showing
DNA molecules recovered from Azospirillum strains
of diverse geographic origin. Electrophoresis was
for 30 h at 80 V. A. brasilense strains: (A) Sp7;
(B) Spl3t; (C) Sp84; (D) Cd; (E) JM82A ; (F JM125A2.
A. lipoferum strains: (G) SpUSA5b; (H) SpRG6xx.
























A B C D E F G H


Fig. 3-4. Autoradiogram of gel shown in Fig. 3-3.
The gel was trimmed of material above pAZ1, blotted
onto nitrocellulose, and hybridized to labeled pWK27
as described in the text. Exposure was for 8 days.
(A) Sp7; (B) Spl3t; (C) Sp84; (D) Cd; (E) JM82A1;
(F) JM125A 2; (G) SpUSA5b; (H) SpRG6xx.












APPENDIX A
NORMALIZATION OF GEL MOBILITY DATA

In order to calculate a regression line relating electro-

phoretic mobility to molecular weight, it was necessary

to pool mobility data from 20 gels run for various lengths

of time under slightly differing conditions. Therefore,

the mobility data from 19 of the gels had to be normalized

to the data from one gel chosen as a standard. The gel

illustrated in Fig. 1-6 was chosen as the standard because

it includes the greatest number (15) of DNA bands.

The standard gel was designated B and the mobilities

of its DNA bands were B1, B2,...., Bn. The mobilities

of the bands in gel A (which had in common with gel B n different

plasmid bands numbered 1,2,...,n) were designated

A1, A2,..., An. The constant Ka was calculated for

which the expression









I KaAi-Bil/(KaAi+Bi)

where i=1,2,...,n



was at a minimum.

Then the mobilities of the plasmids in gel A were multi-

plied by Ka. A different Ka was calculated for each gel

to be normalized. 58












APPENDIX B
REGRESSION LINE CALCULATIONS

If the logarithms of the average normalized mobilities

of the standard plasmids were designated as X1, X2,...Xn and

the logarithms of their molecular weights (in megadaltons)

were designated Y1, Y2'" .Yn then the logarithm of the

molecular weight (Yt) of the unknown molecule could be

calculated from the logarithm of its mobility (Xt) according

to the formula





Yt= A+BXt

where A=T-BX




B = EXiY (EXi)(zYi)

X 2 (Zxi)2
n



S= EXi/n


S= EYi/n













LITERATURE CITED


1. Aaij, C., and P. Borst. 1972. The gel electrophoresis
of DNA. Biochim. Biophys. Acta 269:192-200.

2. Ahmad, M. H. 1978. Influence of nitrogen on growth,
free amino acids and nitrogenase activity in Spirillum
lipoferum. J. Gen. Appl. Microbiol. 24:271-278.

3. Albrecht, S. L., and Y. Okon. 1980. Cultures of
Azospirillum, p. 746. In A. San Pietro (ed.), Methods
in enzymology, vol. 69. Academic Press, New York.

4. Barker, G. R. 1978. Genetic expression and its control
in naturally occurring bacterial plasmids, pp. 25-29.
In P. W. Kent (ed.), New approaches to genetics:
Developments in molecular genetics. Oriel Press, Boston.

5. Barker, G. R., and N. Hardman. 1978. The effects of
acridine orange on deoxyribonucleic acid in Escherichia
coli. Biochem. J. 171:567-573.

6. Brent, T. P. 1972. Repair enzyme suggested by mammalian
endonuclease activity specific for ultraviolet-
irradiated DNA. Nature (London), New Biol. 239:172-173.

7. Bujard, H. 1970. Electron microscopy of single-stranded
DNA. J. Mol. Biol. 49:125-137.

8. Cairns, J. 1963. The chromosome of Escherichia coli.
Cold Spring Harbor Symp. Quant. Biol. 28:43-45.

9. Carr, T. C. 1978. Cryptic plasmids in Azospirillum.
Master's Thesis, University of Florida, Gainesville.

10. Casse, F., C. Boucher, J. S. Julliot, M. Michel, and
J. Denarie. 1979. Identification and characterization
of large plasmids in Rhizobium meliloti using agarose
gel electrophoresis. J. Gen. Microbiol. 113:229-242.

11. Chopra, I. 1975. Mechanism of plasmid-mediated resist-
ance to cadmium in Staphylococcus aureus. Antimicrob.
Agents Chemother. 7:8-14.

12. Clark, D., A. A. Weiss, and S. Silver. 1977. Mercury
and organomercurial resistances determined by plasmids
in Pseudomonas. J. Bacteriol. 132:186-196.
60 -












13. Clewell, D. B., and D. R. Helenski. 1969. Supercoiled
circular DNA-protein complex in Escherichia coli:
Purification induced conversion to an open circular
DNA form. Proc. Natl. Acad. Sci. U.S.A. 62:1159-1166.

14. Cooper, S., and C. E. Helmstetter. 1968. Chromosome
replication and the division cycle of Escherichia
coli. B/r. J. Mol. Biol. 31:519-540.

15. Crick, F. H. C., L. Barnett, S. Brenner, and R. J.
Watts-Tobin. 1961. General nature of the genetic code
for proteins. Nature (London) 192:1227-1232.

16. Currier, T. C., and E. W. Nester. 1976. Evidence for
diverse types of large plasmids in tumor-inducing
strains of Agrobacterium. J. Bacteriol. 126:157-165.

17. Davis, J., and D. I. Smith. 1978. Plasmid-determined
resistance to antimicrobial agents. Ann. Rev.
Microbiol. 32:469-518.

18. Davis, R. W., M. Simon, and N. Davidson. 1971. Electron
microscope heteroduplex methods for mapping regions of
base sequence homology in nucleic acids. In L. Grossman
and K. Moldave (ed.), Methods in enzymology, vol. 21.
Academic Press. New York.

19. Day, J. M., and J. Dobereiner. 1976. Physiological
aspects of N -fixation by a Spirillum from Digitaria
roots. Soil Biol. Biochem. 8:45-50.

20. Denhardt, D. 1966. A membrane filter technique for
the detection of complementary DNA. Biochem. Biophys.
Res. Commun. 23:641-646.

21. Dobereiner, J., I. E. Marriel, and M. Nery. 1976.
Ecological distribution of Spirillum lipoferum
Beijerinck. Can. J. Microbiol. 22:1464-1473.

22. Eckhardt, T. 1978. A rapid method for the identification
of plasmid desoxyribonucleic acid in bacteria.
Plasmid 1:584-588.

23. Emerich, D. W., and R. H. Burris. 1978. Complementary
functioning of the component proteins of nitrogenase
from several bacteria. J. Bacteriol. 134:936-943.

24. Fangman, W. L. 1978. Separation of very large DNA
molecules by gel electrophoresis. Nucleic Acids
Res. 5:653-665.

25. Fisher, M. P., and C. W. Dingman. 1971. Role of mole-
lecular conformation in determining the electrophoretic
properties of polynucleotides in agarose-acrylamide
composite gels. Biochemistry 10:1895-1899.









26. Franche, C., and C. Elmerich. 1981. Physiological
properties and plasmid content of several strains
of Azospirillum brasilense and A. lipoferum. Ann.
Microbiol. (Inst. Pasteur) 132A:3-18.

27. Gauthier, D., and C. Elemrich. 1977. Relationship
between glutamine synthetase and nitrogenase in
Spirillum lipoferum. FEMS Microbiol. Letters 2:101-104.

28. Gillis, M., J. De Ley, and M. De Cleene. 1970. The
determination of molecular weight of bacterial genome
DNA from renaturation rates. Eur. J. Biochem. 12:143-153.

29. Hansen, J. B., and R. H. Olsen. 1978. Isolation of large
bacterial plasmids and characterization of the P2
incompatibility group plasmids pMG1 and pMG5. J.
Bacteriol. 135:227-238.

30. Hirota, Y. 1960. The effect of acridine dyes on mating
type factors in Escherichia coli. Genetics 46:57-64.

31. Hohn, B., and D. Korn. 1969. Cosegregation of a sex
factor with the Escherichia coli chromosome during
curing by acridine orange. J. Mol. Biol. 45:385-395.

32. Jacob, F., S. Brenner, and F. Curzin. 1963. On the
regulation of DNA replication in bacteria. Cold Spring
Harbor Symp. Quant. Biol. 28:329-348.

33. Kato, A. C., and M. J. Fraser. 1973. Action of a single-
strand specific Neurospora crassa endonuclease on
ultraviolet light-irradiated native DNA. Biochem.
Biophys. Acta 312:645-655.

34. Kline, B. C., and J. R. Miller. 1975. Detection of non-
integrated plasmid deoxyribonucleic acid in the folded
chromosome of Escherichia coli: Physicochemical approach
to studying the unit of segregation. J. Bacteriol. 121:
165-172.

35. Kline, B. C., J. R. Miller, D. E. Cress, M. Wlodarczyk,
J. J. Manis, and M. R. Otten. 1976. Nonintegrated
plasmid-chromosome complexes in Escherichia coli.
J. Bacteriol. 127:881-889.

36. Klipp, W. 1980. Personal communication.

37. Kuehn, G. D., and B. A. McFadden. 1968. Factors
affecting the synthesis and degradation of ribulose-
1, 5-diphosphate carboxylase in Hydrogenomonas facilis
and Hydrogenomonas eutropha. J. Bacteriol. 95:937-946.









38. Magalhaes, L. M. S., C. A. Neyra, and J. Dobereiner.
1978. Nitrate and nitrite reductase negative mutants
of N -fixing Azospirillum spp. Arch. Microbiol.
117:247-252.

39. Mazur, B. J., D. Rice, and R. Haselkorn. 1980. Inter-
species homology of nitrogenase genes. Proc. Natl.
Acad. Sci. USA 77: 191-195.

40. Meyers, J. A., D. Sanchez, L. P. Elwell, and S. Falkow.
1976. Simple agarose gel electrophoretic method for
the identification and characterization of plasmid
deoxyribonucleic acid. J. Bacteriol. 127:1529-1537.
(Erratum, J. Bacteriol. 129:1171, 1977.)

41. Mishra, A. K., P. Roy, and S. Bhattacharya. 1980.
Deoxyribonucleic acid-mediated transformation of
Spirillum lipoferum. J. Bacteriol. 137:1425-1427.

42. Nakahara, H., T. Ishikawa, S. Sarai, I. Kondo, and S.
Mitshuhashi. 1977. Frequency of heavy metal resistance
in bacteria from inpatients in Japan, Nature 266:165-167.

43. Nelson, L. M., and R. Knowles. 1978. Effect of oxygen
and nitrate on nitrogen fixation and denitrification
by Azospirillum brasilense grown in continuous culture
Can. J. Microbiol. 24:1395-1403.

44. Neyra, C. A., J. Dobereiner, R. Lalande, and R. Knowles.
1977. Denitrification by N -fixing Spirillum lipoferum.
Can. J. Microbiol. 23:300-305.

45. Neyra, C. A., and P. Van Berkum. 1977. Nitrate re-
duction and nitrogenase activity in Spirillum lipoferum.
Can. J. Micbriol. 23: 306-310.

46. Nur, I., Y. Okon, and Y. Henis. 1980. An increase in
nitrogen content of Setaria Italica and Zea mays
inoculated with Azospirillum. Can. J. Microbiol.
26:482-485.

47. Ohtsubo, E., R. C. Deonier, H. J. Lee, and N. Davidson.
Electron microscope heteroduplex studies of sequence
relations among plasmids of Escherichia coli. IV. The
sequence in F14. J. Mol. Biol. 89:565-584.

48. Okon, Y., S. L. Albrecht, and R. H. Burris. 1976.
Factors affecting growth and nitrogen fixation of
Spirillum lipoferum. J. Bacteriol. 127:1248-1254.

49. Okon, Y., S. L. Albrecht, and R. H. Burris. 1976.
Carbon and ammonia metabolism of Spirillum lipoferum.
J. Bacteriol. 128: 592-597.












50. Okon, Y., S. L. Albrecht, and R. H. Burris. 1977. Methods
for growing Spirillum lipoferum and for counting it
in pure culture and in association with plants. Appl.
Env. Microbiol. 33:85-88.

51. Okon, Y., J. P. Houchins, S. L. Albrecht, and R. H.
Burris. 1977. Growth of Spirillum lipoferum at constant
partial pressures of oxygen, and the properties of its
nitrogenase in cell-free extracts. J. Gen. Microbiol.
98:87-93.

52. Orgel, A., and S. Brenner. 1961. Mutagenesis of bacterio-
phage T4 by acridines. J. Mol. Biol. 3:762-768.

53. Perry, R. D., and S. Silver. 1981. Transport studies
on cadmium resistance in whole cells and subcellular
membranes of Staphylococcus aureus, p. 159. In Abstracts
of the annual meeting of the American Society for
Microbiology, 1981. ASM Publications, Washington, D.C.

54. Pettijohn, D. E., and R. Hecht. 1973. RNA molecules
bound to the folded bacterial genome stabilize DNA
folds and segregate domains of supercoiling. Cold
Spring Harbor Symp. Quant. Biol. 38:31-41.

55. Polsinelli, M., E. Baldanzi, M. Bazzicalupo, and E.
Gallori. 1980. Transfer of plasmid pRD1 from Escherichia
coli to Azospirillum brasilense. Mol. Gen. Genet.
178:709-711.

56. Rainbow, A. J., and S. Mak. 1973. DNA damage and bio-
logical function of human adenovirus after U.V.-
irradiation. Int. J. Radiat. Biol. 24:59-72.

57. Rigby, P. W. J., M. Dieckmann, C. Rhodes, and P. Berg.
1977. Labeling deoxyribonucleic acid to high specific
activity in vitro by nick translation with DNA poly-
merase. J. Mol. Biol. 113:237-251.

58. Rothstein, A. 1959. Cell membrane as site of action of
heavy metals. Fed. Proc. 18:1026-1035.

59. Ruvkun, G. B., and F. M. Ausubel. 1980. Interspecies
homology of nitrogenase genes. Proc. Natl. Acad. Sci.
USA 77:191-195.

60. Sampaio, J. A. M., E. M. R. daSilva, J. Dobereiner,
M. G. Yates, and F. O. Pedrosa. 1981. Autotrophy and
methylotrophy in Derxia gummosa, Azospirillum brasilense,
and A. lipoferum, p. 444. In A. H. Gibson and W. E.
Newton (ed.), Current perspectives in nitrogen fixation.
Australian Academy of Science, Canberra City.









61. Sharp, P.A., M Hsu, E. Ohtsubo, and N. Davidson.
1972. Electron microscope heteroduplex studies of se-
quence relations among plasmids of Escherichia coli.
I. Structure of F-prime factors. J. Mol. Biol. 71:471-487.

62. Silver, S., E. Levine, and P. M. Spielman. 1968.
Acridine binding by Escherichia coli: pH dependency
and strain differences. J. Bacteriol. 95:333-339.

63. Smith, D. H. 1967. R factors mediate resistance to mercury,
nickel, and cobalt. Science 156:1114-1116.

64. Smith, R. L., J. H. Bouton, S. C. Schank, K. H.
Quesenberry, M. E. Tyler, J. R. Milam, M. H. Gaskins,
and R. C. Littell. 1976. Nitrogen fixation in grasses
inoculated with Spirillum lipoferum. Science 193:1003-
1005.

65. Southern, E. M. 1975. Detection of specific sequences
among DNA fragments separated by gel electrophoresis.
J. Mol. Biol. 98:503-517.

66. Stonington, O. G., and D. E. Pettijohn. 1971. The folded
genome of Escherichia coli isolated in a protein-DNA-RNA
complex. Proc. Natl. Acad. Sci. USA 68:6-9.

67. Tarrand, J. J., N. R. Krieg, and J. Dobereiner. 1978.
A taxonomic study of the Spirillum lipoferum group,
with descriptions of a new genus, Azospirillum gen. nov.
and two species, Azospirillum lipoferum (Beijerinck)
comb. nov. and Azospirillum brasilense sp. nov. Can. J.
Microbiol. 24:967-980.

68. Thomashow, M. F., R. Nutter, A. L. Montoya, M. P. Gordon,
and E. W. Nester. 1980. Integration and organization
of Ti plasmid sequences in crown gall tumors. Cell 19:
729-739.

69. Tynecka, Z., Z. Gos, and J. Zajac. 1981. Plasmid-
determined Cd2 resistance in Staphylococcus aureus:
Accelerated efflux and reduced net uptake, p. 159.
In Abstracts of the annual meeting of the American
Society for Microbiology, 1981. ASM Publications, Washing-
ton, D.C.

70. Wahl, G. M., M. Stern, and G. R. Stark. 1979. Efficient
transfer of large DNA fragments from agarose gels to
diazobenzylozymethlyl-paper and rapid hybridization by
using dextran sulfate. Proc. Natl. Acad. Sci. USA 76:
3683-3687.

71. Weiss, A. A., S. D. Murphy, and S. Silver. 1977. Mercury
and organomercurial resistances determined by plasmids
in Staphylococcus aureus. J. Bacteriol. 132:197-208.





66



72. Willshaw, G. A., H. R. Smith, and E. S. Anderson. 1979.
Application of agarose gel electrophoresis to the
characterization of plasmid DNA in drug-resistant
Enterobacteria. J. Gen. Microbiol. 114:15-25.

73. Worcel, A., and E. Burgi. 1972. On the structure of
the folded chromosome of Escherichia coli. J. Mol.
Biol. 71:127-147.

74. Worcel, A., E. Burgi, J. Robinton, and C. L. Carlson.
1973. Studies on the folded chromosome of Escherichia
coli. Cold Spring Harbor Symp. Quant. Biol. 38:43-51.














BIOGRAPHICAL SKETCH


The author was born Alvin Gleave Wood on October 5,

1951, in St. Petersburg, Florida. He is the only son of

Mary Gleave Harris and the late Rowland Emery Wood. Follow-

ing graduation from Northeast High School in St. Petersburg,

he attended the University of Chicago and received a B.A. in

biology in June 1973. He received his M.S. degree from the

Department of Microbiology and Cell Science, University of

Florida, 1978. He is currently a candidate for the degree

of Doctor of Philosophy, also in the Department of Microbiology

and Cell Science.












I certify that I have read this study and that in my
opinion it conforms to acceptable stand rds of scholarly
presentation and is fully adequate, in/scope and quality,
as a dissertation for the degree of Doctor-of -Philosophy.

/ // 9: / // (_
Dennis. E. Duggan,,Chairman
Associate Professor of
Microbiology and Cell Science

I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy)

; .- t""/ ,//
Francis C. Davis, Jr./
Associate Professor of
Microbiology and Cell Science

I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy.

- . ., '..
Lonnie 0. Ingram
Associate Professor of
Microbiology and Cell Science

I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of P ilosophy. 4
/ !/ / ','

William W. Hauswirth
Associate Professor of Immunology
and Medical Microbiology













This dissertation was submitted to the Graduate Faculty of
the College of Agriculture and to the Graduate Council, and
was accepted as partial fulfillment of he requirements for
the degree of Doctor of Philosophy./ ,

May 1982 X 1
Deal//College of Agri-ulture


Dean for Graduate Studies and
Research

PAGE 1

MULTIPLE LARGE DNA MOLECULES OF Azospirillum BY ALVIN G. WOOD A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 1982

PAGE 2

ACKNOWLEDGEMENTS The author would like to thank his committee members, Drs. Dennis E. Duggan, Francis C. Davis, Jr., Philip J. Laipis, William W. Hauswirth and L.O. Ingram for their valuable suggestions and criticisms. He would also like to extend special thanks to Dr. William B. Gurley for help with the hybridization experiments and for the use of his laboratory equipment and supplies.

PAGE 3

TABLE OF CONTENTS PAGE ACKNOWLEDGEMENTS ii ABSTRACT iv CHAPTER I OCCURRENCE OF MULTIPLE LARGE DNA MOLECULES IN Azospirillum AND A METHOD FOR ISOLATING THEM ON AGAROSE GELS 1 Introduction 1 Materials and Methods 3 Results 6 Discussion 11 CHAPTER II ACRIDINE ORANGE-INDUCED MUTATIONS IN Azospirillum 31 Introduction 31 Materials and Methods 32 Results 33 Discussion 37 CHAPTER III Nif GENE HYBRIDIZATION STUDIES OF Azospirillum DNA MOLECULES. ... 47 Introduction 47 Materials and Methods 4 8 Results and Discussion 51 APPENDICES A NORMALIZATION OF GEL MOBILITY DATA . . 58 B REGRESSION LINE CALCULATIONS .... 59 LITERATURE CITED 60 BIOGRAPHICAL SKETCH 67

PAGE 4

Abstract of Dissertation Presented to the Graduate Council of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy MULTIPLE LARGE DNA MOLECULES OF Azospirillum by Alvin G. Wood May 19 82 Chairman: Dennis E. Duggan Major Department: Microbiology and Cell Science Six strains of Azospirillum brasilense and two of A. lipoferum were found to harbor as many as eight differentsized circular DNA molecules ranging from 45 to 1500 megadaltons. Identification and separation of these very large molecules were achieved by gently lysing bacteria in the wells of vertical agarose gels, subjecting the lysate to electrophoresis at 2 mA for 6 h, and then continuing electrophoresis at 15-30 mA for an additional 12-48 h. Optimal recovery required lysis at 4°C in the presence of ribonuclease. The technique has been used to isolate large DNAs from other bacteria, including the chromosomes of Escherichia coli and Agrobacterium tumefaciens . Several types of mutants were isolated from acridine orange-treated cultures of A. l ipoferum and A. brasilense . Mutants displaying increased sensitivity to cadmium and

PAGE 5

unable to grow on carbon-free media or on ethanol were all found to have lost a specific plasmid. One of these strains was shown to have suffered deletions in most of its remaining DNA molecules. A mutant unable to grow on N 2 or reduce acetylene was isolated from the multiply-deleted strain, but its DNA molecules showed the same electrophoretic mobilities as those of its parent strain. Methionine-requiring auxotrophs, isolated at a high frequency from A. lipoferum cultures, also possessed DNA molecules with unaltered mobilities. Attempts were made to determine which Azospirillum DNA molecule includes the genes controlling nitrogen fixation by hybridizing a labeled recombinant probe to Southern blots of wild type and mutant DNA molecules. The limited success acheived with this technique indicates that the structural genes for nitrogenase are carried on the largest Azospirillum DNA molecule.

PAGE 6

CHAPTER I OCCURRENCE OF MULTIPLE LARGE DNA MOLECULES IN Azospirillum AND A METHOD FOR ISOLATING THEM ON AGAROSE GELS Introduction. The genus Azospirillum comprises Gram-negative, free-living, nitrogen-fixing bacteria found in association with roots of cereal crops and tropical forage grasses (19). Field experiments conducted at the University of Florida showed higher yields of dry matter in Azospirillum -inoculated pearl millet and guinea grass than in uninoculated controls (64). More recently, Azospirillum inoculation has been reported to enhance corn yields in Israel (46). The potential agronomic value of this association has prompted studies of carbon and nitrogen metabolism in Azospirillum (2, 21, 27, 28, 38, 43, 44, 45, 48, 49, 51), and the scores of strains isolated from various parts of the world have been grouped, on the basis of DNA homology and biochemical characteristics, into two species, A. lipoferum and A. brasilense (67). Very little is known, however, concerning the genetics of Azospirillum (26, 41, 55). In particular, no system of genetic transfer exists which would permit location of the genes controlling nitrogen fixation and facilitate studies of their expression. 1

PAGE 7

Our interest in developing a system of genetic transfer for Azospirillum led us to examine several wild type strains for the presence of plasmids. Our initial attempt to identify plasmid DNA in Azospirillum (9) involved dye bouyant density ultracentrifugation of alkaline denatured lysates (61) and direct visualization of plasmids in the satellite bands by electron microscopy (18). Open circular (OC) DNAs of various contour lengths were seen, but the apparent multiplicity of molecules in each strain and their large sizes relative to the plasmid chosen as a size standard (ColE. ) made it difficult to accurately assess the numbers and sizes of plasmids in any given strain. However, plasmids with molecular weights in excess of 300 megadaltons (Mdal) did appear to be present in several Azospirillum strains. This observation prompted us to try two electrophoretic techniques specifically designed for the isolation of large plasmids (10, 29), but neither of these permitted the isolation of more than three plasmids from any Azospirillum strain. We felt there was a strong possibility that very large plasmids were present in these strains but that they were being sheared during the mechanical manipulations, however gentle, inherent in these procedures. Therefore we adopted a method which is theoretically the most gentle of all, that described by Eckhardt (22). This technique differs fundamentally from other electrophoretic techniques in that plasmid DNA is not extracted from cells

PAGE 8

prior to electrophoresis. Rather, the bacteria are lysed directly in the wells of the gel apparatus, resulting in minimal nicking or breakage of covalently closed circular (CCC) DNA. Using a modified version of this technique, we have discovered the widespread occurrence of a multiplicity of very large DNA molecules in strains of Azospirillum . The present communication describes the electrophoretic conditions necessary for the successful isolation of these Azospirillum DNA molecules. Estimates of the sizes of the molecules harbored by one Azospirillum strain are provided, based on a comparison of their mobilities with those of plasmids of known molecular weight. The very low mobilities of some of the Azospirillum molecules suggest that they represent small chromosomes rather than large plasmids. Indeed, we have been able to isolate slowly-migrating DNA bands from strains of Escherichia coli and Agrobacterium tumef aciens , including two strains which do not harbor plasmids. Evidence is provided indicating that the slowlymigrating DNA bands isolated from Azospirillum represent CCC DNA uncomplexed with protein. Materials and Methods Bacterial strains . Table 1-1 lists the Azospirillum strains examined for plasmid content. Table 1-2 lists other bacterial strains harboring plasmids of known molecular weights used for construction of the standard size curve.

PAGE 9

Growth conditions . All bacteria were grown to early stationary phase prior to harvesting for electrophoresis. Azospirillum and A. tumefaciens strains were grown in a succinate/mineral salts medium (50) supplemented with 0.01% yeast extract. Pseudomonas putida and E. coli strains were grown in nutrient broth (Difco) supplemented with 0.01% yeast extract. Growth temperatures were 35°C for Azospirillum and E. coli and 28°C for A. tumefaciens and P. putida . Plasmid isolation and agarose gel electrophoresis . We used an electrophoretic method based on that described by Eckhardt (22). The protocol outlined here includes modifications found to be necessary for optimal, reproducible visualization of the electrophoretic bands representing the largest DNA molecules. Vertical gels were cast with 0.6% or 0.7% molten agarose (BioRad Standard Low -m ) in Tris/borate/EDTA electrophoresis (E) buffer (40). The agarose was tempered at 42°C for 20 min prior to casting the gel in order to minimize contracture during solidification. The plastic comb used to form wells had 16 teeth (13x9x1.5 mm). The gel was submerged in E buffer and allowed to age at least 4 h at 4°C prior to removal of the comb. Spraying the comb lightly with PAM (3oyle-Midway ) prevented agarose from sticking to it during removal. Except where otherwise noted, wells were loaded in the following manner. One milliliter of cell culture was centrifuged for one minute in a microcentrifuge (Fisher Model 235). The

PAGE 10

supernatant solution was pipetted off using a vacuum aspirator and the pellet was resuspended in 20-100 pi of 20% ficoll in E buffer. Ten microliter aliquots of the cell suspensions were added to wells preloaded with 15 y 1 of a solution containing 20% ficoll, 10 yg/ml lysozyme (Sigma), 100 pg/ml ribonuclease A (Sigma), and 0.05% xylene cyanol FF (Kodak) in E buffer. No attempt was made to mix the cell suspension with the lysozyme solution inside the well. The cells were allowed to interact with the lysozyme mixture at 4°C for a minimum of 30 min, and then 30 u 1 of 10% ficoll, 1% SDS in E buffer was added, followed by 50 y 1 of 5% ficoll, 1% SDS in E buffer. A current of 2 mA was applied for a minimum of 6 h, and then the current was raised to 15-30 mA (50-100 V at 4°C). Electrophoresis was continued for 12-48 h depending on the voltage used and the degree of molecular separation required. The large capacity (2.5 1) of the buffer reservoirs of our electrophoresis apparatus made recirculation of buffer unnecessary. Photography . Gels were stained for a minimum of 30 min in 0.5 ug/ml ethidium bromide and visualized with either a 254 nm hand-held UV light (UV Products) or a 300 nm transilluminator (Fotodyne). Photographs were taken through #4 and #29 Wratten filters, using Type 57 film (Polaroid). Standard curve construction . Mobility data from 19 gels were normalized to the mobility data of the gel illustrated in Fig. 1-6 according to the method described by Hansen and Olsen (29, Appendix A) except that absolute mobilities

PAGE 11

(distances of plasmid migration from origin) rather than relative mobilities (absolute mobilities divided by gel lengths) were used in the calculations (see Appendix A). The number of DNA bands in common between the normalized and standard gels ranged from 7 to 15 with an average of 12. The logarithm of the average normalized mobility of a given standard plasmid was plotted against the logarithm of the molecular weight of that plasmid, and a least squares regression line was calculated (see Appendix B). This regression line was then used to estimate the molecular weights of the Azospirillum molecules as well as the presumed chromosomes of E. coli and A. tumefaciens . Plasmid nomenclature . The molecule having the lowest mobility in each Azospirillum strain has been designated pAZl; molecules are then numbered in order of increasing mobility. All the molecules of a given strain are suffixed with that strain designation. For example, the smallest plasmid of Spl3t is pAZ6-Spl3t. Results Mult i plicity of DNA molecules found in Azospirillum strains . Figures 1-1 and 1-2 illustrate the electrophoretic banding patterns obtained from eight geographically diverse isolates of A. brasilense and A. lipoferum lysed in the wells of vertical agarose gels. Each strain has a characteristic array of DNA molecules of various mobilities, an observation which can be used for purposes of identification. Every strain harbors two molecules whose extremely low

PAGE 12

7 mobilities suggest that they represent small chromosomes rather than plasmids. Under optimal conditions, the recovery of these DMAs is highly reproducible except for the small plasmid which bands in the region of linear DNA in some gels (Fig. 1-1, lane D; Fig. 1-2, lanes C and D) . A comparison of Figs. 1-1 and 1-2 indicates that very long periods of electrophoresis are necessary in order to achieve separation of all the DNA bands, in accordance with what would be expected for very large DNAs. The increase in resolution achieved by increasing electrophoresis time is, unfortunately, accompanied by a tendency for the most slowlymigrating bands to become faint or disappear altogether (data not shown). This suggests, however, that the material is those bands is fragile, presumably because of its high molecular weight. Effect of ultraviolet light on mobilities of JM125A 2 molecules . Figure 1-3 compares the mobilities of UV-irradiated and unirradiated molecules of JM125A 2 . The JM125A 2 molecules were isolated in the usual manner except that electrophoresis was terminated after 4 h at 80 V. Blocks of agar were cut from lanes of the gel, extending from the well to the position of the tracking dye. One of the unstained agar blocks 2 (lane B) was subjected to a dose of approximately 3000 J/m of 25 4 nm UV, while the other (lane A) was untreated. This dose should have been sufficient to introduce at least one chain break into every CCC DNA molecule in the irradiated gel (6, 7, 33, 56). Both treated and untreated agarose

PAGE 13

blocks were then imbedded in a horizontal agarose gel and subjected to electrophoresis for 8 h at 50 V. As indicated in Fig. 1-3, UV irradiation converted the DNA molecules from forms capable of movement through an agarose gel into forms incapable of such movement. Presumably, this represents the conversion of CCC DNA into OC DNA. Effects of enzymatic treatments on DNA recovery . In order to acquire information concerning the physical relationships of the DNA molecules to other cellular components, the roles of lysozyme, RNase, and protease (Sigma Type VI) in optimal DNA recovery were assessed. Figure 1-4 shows that neither the addition of protease to the standard cell mixture nor the elimination of lysozyme had an appreciable effect on DNA recovery or mobility. The elimination of RNase, however, resulted in failure to recover pAZl and pAZ2 as well as poor recovery of pAZ3. Effect of cell mass on DNA isolation . Figure 1-5 illustrates the result of an experiment in which cell suspensions of JM125Aand AT181 were serially diluted prior to loading the wells. The smearing of the bands in lanes A and E 7 appears to be due to overloading. The use of only 10 cells (lanes D and H) allowed visualization of all the DNA bands in this experiment. However, in other experiments (data not 7 shown) using 10 cells resulted in very faint bands, particularly for the smaller molecules. Optimal recovery was p usually achieved with 10 cells.

PAGE 14

Size estimates of DNA molecules . The mobilities of the DNA molecules of JM125A ? in relation to plasmids of known molecular weight and slowly-migrating DNAs of other bacteria are illustrated in Figs. 1-6, 1-7, and 1-8. Based on Fig. 1-6, JM125&2 appears to harbor five molecules larger than the largest standard molecule (pMGl, 312 Mdal ) . Recovery of the larger JM125Amolecules was poor in Figs. 1-6 and 1-7; these are included primarily to show the slowly-migrating DNAs isolated from E. coli (Fig. 1-7, lane F ; Fig 1-3, lanes E, G, and H) and A. tumefaciens (Fig. 1-7, lane A). The two slowly-migrating DNAs recovered from A. tumefaciens have been given the designations pXXl-AT181 and pXX2-ATl81. The isolation of these slowly-migrating DNAs is not completely reproducible (Fig. 1-6, lanes E and F ; Fig 1-7, lane E; Fig. 1-8, lanes A and D). Indeed, the difficulty of isolating the molecules appears, in our experience, to be inversely related to mobility. Thus, our rate of success in isolating slowly-migrating DNAs from Azospirillum is 90% or better, but our success rate with the slowly-migrating E. coli DNA has never exceeded 5 0%. We have never isolated a slowly-migrating DNA band from either of the two Pseudomonas strains used in the present study. This failure may be related to an observed tendency for these strains to lyse prematurely. Mobility data from 20 agarose gels were normalized and used to construct a standard curve relating electrophoretic mobility and molecular weight (Fig. 1-9). Since reports in the literature had suggested that CCC DNAs larger than 80

PAGE 15

10 Mdal (72) or 140 Mdal (29) migrate faster than predicted from linear extrapolations of standard curves based on smaller CCC DNAs, we initially calculated a regression line not including pMGl and pMG5. When this line was used to estimate the sizes of the Azospirillum molecules, however, it seemed impossible that the values obtained could be underestimates. We therefore recalculated regression data with the large Pseudomonas plasmids included, and again with the slowly-migrating band from E. coli included and assigned a molecular weight of 2800 Mdal (8,14,28). The three sets of regression line estimates are summarized in Table 1-3. Figure 1-9 is a graph of regression line B, chosen because it includes only those molecules measured by electron microscopic contour length. Effect of voltage gradient on regression estimates . If the larger molecules examined in the present study were really migrating faster than predicted, this effect should be more pronounced at higher than at lower voltages (24). We were particularly interested in this possibility since, in our attempts to optimize electrophoresis conditions, we had used voltages ranging from 50 V to 100 V and we wished to include data from as many gels as possible in our regression line calculations. Figure 1-10 compares the mobilities of five standard plasmids run at 50 V for 42 h with mobilities of the same plasmids run at 100 V for 24 h. The two curves are nearly parallel and neither displays a convincing change

PAGE 16

11 of slope above 140 Mdal. Table 1-4 indicates that, for each standard plasmid, the ratio of mobility at 100 V to mobility at 50V is a constant value. Discussion Eckhardt first described in situ lysis of bacteria in agarose gels as a rapid method for plasmid isolation (22). We have sacrificed the rapidity of the technique but exploited its gentleness in order to isolate very high molecular weight CCC DNAs from Azospirillum and other bacteria. Our initial excitement in isolating slowly-migrating DNAs on agarose gels was tempered with concern that the low mobilities might reflect an open circular nature or some protein interaction rather than large size. The electrophoretic behavior of the JM125A„ molecules subsequent to UV irradiation, however, strongly indicates that they are covalently closed and supercoiled. A DNA/protein interaction, while not strictly ruled out by the failure of protease to alter the electrophoretic mobilities of the molecules, seems unlikely in view of this result. Furthermore, such an association would have to be resistant to dissociation by the SDS which quickly migrates from the upper ficoll layers down through the DNA-containing region of the gel during electrophoresis. Thus, the slowly-migrating bands do not appear to represent relaxation complexes of the type isolated from plasmid-bearing strains of E. coli , which dissociate yielding OC DNA when exposed to SDS or protease (13).

PAGE 17

12 A comparison of the DNA molecules isolated from A. brasilense strains Sp7 and Spl3t provides further evidence that the slowly-migrating DNAs are not simply isomeric forms of smaller plasmids. Since these two strains were isolated from the same region of Brazil and display nearly identical electrophoretic banding patters, there is a good possibility that they are isogenic except for the occurrence of pAZ6-Sp7 in one strain. If so, none of the slowly-migrating DNAs of either strain could represent an isomeric form of this relatively small molecule. The requirement of RNase treatment for isolation of the larger Azospirillum DNAs suggests that, in their native forms, these molecules are attached via RNA to some cellular component in a manner precluding entry into the gel matrix. It is also possible that the larger Azospirillum molecules are attached to one another via RNA. Assuming the likelihood that essential genes are carried on the largest two or three molecules, some mechanism to ensure cosegregation of newlyreplicated molecules into daughter cells would appear to be necessary. Molecules as large as pAZl and pAZ2 might further be expected to exist inside the cell in condensed, folded states. These considerations lead us to postulate that pAZl and pAZ2 (and perhaps pAZ3) are arranged in a chromosomal structure closely resembling that believed to occur in E. coli (54, 73, 74). The only difference between the two "nucleoid" structures would be that in E. coli the RNAstabilized domains comprise a single, continuous DNA molecule,

PAGE 18

13 whereas in Azospirillum these domains are divided into two or three continuous DNA molecules. The remaining Azospirillum molecules might form nonintegrative associations with the Azospirillum nucleoid analogous to those described between other large, stringently controlled plasmids and their host chromosomes (34, 35). We have attempted to estimate the sizes of the DNA molecules of one Azospirillum strain (JM125A2* b ^ comparing their mobilities to those of plasmids whose sizes have been calculated from electron microscopic contour length measurements. A problem with this effort arose in that five of the JM12 5Ao molecules migrated more slowly than pMGl, the largest standard plasmid available. Their sizes, therefore, had to be estimated from a linear extrapolation of our standard curve (Fig. 1-9) and so must be considered only approximate. Some investigators have cautioned against standard curve extrapolations on the grounds that CCC DNAs larger than 80 Mdal (72) or 140 Mdal (29) migrate faster than predicted. For several reasons, this appears not to be the case under our electrophoresis conditions. First, it is difficult to believe that the calculated values for the larger Azospirillum molecules could be underestimates. Second, these estimates change only modestly when the Pseudomonas plasmids are disallowed or when the chromosome of E. coli is included in the regression line calculation (Table 1-4). Third, the hypothetical nonlinearity of the standard curve should have been greater for a gel run at 100

PAGE 19

14 V than for a gel run at 50V, but Table 1-4 indicates that the relative mobilities of the standard plasmids were nearly identical at the two voltages. Finally, a theoretical justification for nonlinearity of standard curves for CCC DNAs in the high molecular weight range has not been advanced. The explanation offered for the fast mobilities of high molecular weight linear DNAs, i.e., "end-on" migration (1, 25), would seemingly not apply to high molecular weight CCC DNA. From the estimated sizes of the Azospirillum molecules and assuming one copy of each per cell, the full genetic complement of DNA for these bacteria appears to be approximately 9 4.3x10 daltons, some 50% greater than the corresponding value for E. coli (8, 14, 28). At present, we can only speculate as to the reason for this discrepancy. The large complement of DNA may simply reflect the metabolic diversity of these bacteria; Azospirillum species are capable of carrying out most of the known nitrogen transformations (19, 44, 45), can grow heterotrophically (49) or autotrophically (60), and tolerate the full range of oxygen tensions from fully aerobic (49) to anaerobic with nitrate as terminal electron acceptor (45). Alternatively, some of the Azospirillum DNA may be redundant. This redundancy, if it does occur, could provide a basis for recombination among Azospirillum DNA molecules, underlying a potential mechanism for the evolution of new strains. The application of the modified Eckhardt technique to bacteria harboring size standard plasmids led to the discovery

PAGE 20

15 that slowly-migrating DNAs could be isolated from species other than Azospirillum . Molecules with apparent molecular weights of 500 and 1800 Mdal were isolated from A. tumefaciens AT181 along with the two previously described plasmids. These four molecules may well represent the full genetic complement of this A. tumefaciens strain since the sum of Q their estimated sizes is 2.6x10 daltons. Hence, pXXl-AT181 may, in fact, represent the Agrobacterium chromosome. The slowly-migrating DNA isolated from E. coli strains appears to represent the E. coli chromosome since it displayed an appropriate mobility and could be recovered from both plasmidharboring and plasmidless strains. Isolation of intact E. coli chromosomes by ultracentrif ugation of gently lysed cells through neutral sucrose gradients has been described by others (66, 73) . In summary, we have demonstrated that strains of Azospirillum harbor unique arrays of large DNA molecules. The probability that these molecules comprise the full genetic complement of their host bacteria suggests that Azospirillum should be considered a multichromosomal prokaryote. Arrangement of genetic material in this fashion contrasts sharply with the situation in E. coli , in which more than 90% of the DNA is carried on a single large molecule. The common assumption that the DNA of most prokaryotes is arranged as it is in E. coli may reflect, to a certain extent, the previous lack of a suitable protocol for isolating CCC DNAs larger than 500 Mdal.

PAGE 21

16

PAGE 22

17 4J

PAGE 23

13 Table 1-3. Linear regression estimates of sizes of DNA molecules DNA molecule B pAZl-JM125A pAZ2-JM125A pAZ3-JM125A pAZ4-JM125A pAZ5-JM12 5A„ pAZ6-JMl25A pAZ7-JM125A 2 pMGl pMG5 pTi-AT181 pAT-AT181 pXXl-AT181 PXX2-AT181 RP4 F F14 F14 AF E. coli chromosome 1380+25 1130±180 600+56 385±30 345±26 128±6 46±6 289±46 260±35 113+17 154±19 1650+160 473+42 33 + 6 64±10 205+16 141+15 3150±530 1510±290 1230±190 637±62 404±32 362±30 131 + 6 46±6 302±49 270±38 114±18 158+19 18101180 500±45 32±6 64±10 212±17 144±16 3500±610 1330+240 1090±170 587+53 379±28 342±25 130±6 47±7 287+46 259+35 114±17 155±13 1580+150 465+40 34±6 66±10 205116 142115 29701490 based on mobilities of standard plasmids smaller than 160 Mdal based on mobilities of all standard plasmids 'based on mobilities of all standard plasmids plus E. coli chromosome NOTE: Sizes are in megadaltons, + ls.d.

PAGE 24

19 Table 1-4. Comparison of mobilities of standard plasmids run at 50 V and 100 V Plasmid Gel A a Mobility (cm) Gel B° B/A pMGl

PAGE 25

20 Fig. 1-1. Agarose gel electrophoresis of large DNA molecules from eight Azospirillum strains of diverse geographic origin. Electrophoresis was for 24 h at 80 V. A . brasilense strains: (A) Sp7; (b) Spl3t; (c) Sp84; D) Cd; (E) JM82A n ; (f) JM125A . A. lipoferum strains: SpUSA5b E) JM82A,; (F) JM125A 2 . 5b; (H) SpRG6xx. ''

PAGE 26

21 Fig. 1-2* Increased separation of large Azospirillum DNA molecules in an agarose gel subjected to electrophoresis for 36 h at 80 V. A. brasilense strains: (A) Sp7; (B) Spl3t; (C) Sp84; (D) Cd; (E) JM82A 1 ; (f) JM125A 2 . A. lipoferum strains: (g) SpUSA5bj (h) SpRGbxx.

PAGE 27

22 Fig. 1-3Agarose gel showing the effect of UV light on mobilities of DNA molecules from JMl25Ap. Lanes from a preparative vertical gel run at 80 V for 4 h were excised ? and either (A) not treated or (B) irradiated with 3000 J/m 25^ nm UV. The agarose blocks were then embedded in a horizontal gel and subjected to electrophoresis at 50 V for an additional 8 h.

PAGE 28

23 I

PAGE 29

24 A B CD EFGH Fig. 1-5 . Effect of cell number on isolation of DNA molecules from (A-D) JM125A 2 and (E-H) ATl8l. Electrophoresis was at 80 V for 24 h. (A) and (e) q ft 3x10 cells; (B) and (F) 10 cells; (c) and (G) 3x10 cells; (D) and (h) 10^ cells. 7

PAGE 30

25 Fig. 1-6. Agarose gel electrophoresis showing mobilities of JM125A, mobilities of size s\ was at 90 V for 28 h. (pMG5); (C) JM125A P ; (F) X125^ (F). d DNA molecules in relation to andard plasmids. Electrophoresis (A) P P S1239 (pMGl); (B) PpSl2^0 (D) AT181 (pTi,pAT); (E) C600 (RP4)

PAGE 31

26 Fig. 1-7. Agarose gel electrophoresis showing slowly-migrating DNAs isolated from A. tumefaciens and E. coli . Electrophoresis was for 2)\ h at 75 v (A) AT181 (pTi^pAT); (B) PpSl239 (pMGl)j (c) PpS1240 (pMG5); (D) JM125A 2 ; (E) C600 (RP 1 !); (F) JCl82.

PAGE 32

27 Fig. 1-8. Agarose gel electrophoresis showing recovery of slowly-migrating DNA from three of four E. coli strains. The strains run in lanes (g) and (h) do not harbor plasmids. Electrophoresis was for 24 h at 60 V. (A) X1254; (b) PpSl239 (pMGl); (c) JM125A (S) D) AT181 (pTi,pAT); (E) C600 (RP4); (f) PpSl240 (pMG5); G) E. coli C; (H) JC182.

PAGE 34

29 (SNoxivavoaw) xhoism avinoaiow

PAGE 35

30 500 pMGl 100 50 J. 1 1 0.5 0.6 ,7 LOG MOBILITY (cm) 0.8 o.q Fig. 1-10. Comparison of mobilities of standard plasmids in gels run at different voltages. The 50 V gel was run for 42 h and the 100 V gel for 24 h.

PAGE 36

CHAPTER II ACRIDINE ORANGE-INDUCED MUTATIONS IN Azospirillum Introduction Genetic material in the genus Azospirillum appears to be arranged in a manner unusual for bacteria. Rather than possessing a single large chromosome with or without accessory DNA in the form of plasmids, Azospirillum strains harbor unique arrays of up to eight different-sized covalently closed circular (CCC) DNA molecules, the largest estimated to be 1500 megadaltons (Mdal) (see Chapter I). The occurrence of so many molecules intermediate in size between what is normally considered to be very large for a plasmid (300 Mdal) and what would be considered a small chromosome (1500 Mdal) makes the usual plasmid/chromosome dichotomy less than obvious in these bacteria. Plasmid DNA is generally defined as encoding functions not essential for cell growth under usual conditions. Growth of plasmid-harboring bacteria in the presence of acridine orange (AO) often permits the isolation, at an increased frequency, of strains which have been "cured" of the plasmid (30). We therefore grew a representative strain of A. lipoferum and one of A. brasilense in AO-containing medium and tested individual colonies for a variety of nonessential phenotypic traits. Since the capacity for 31

PAGE 37

32 nitrogen fixation can be considered nonessential, the potentially mutant colonies were tested for growth on nitrogen-free medium. In addition, because of the blurring of the distinction between plasmid and chromosome in these bacteria, all colonies were screened for auxotrophy. Mutant colonies identified in this manner were analyzed for DNA content by agarose gel electrophoresis. The present communication describes the successful correlation of three phenotypic properties with the presence of a specific plasmid in both Azospirillum strains examined. Furthermore, we report the isolation of a mutant having suffered multiple deletions in its DNA molecules and the isolation of both Nif" and auxotrophic strains for which no obvious loss of DNA could be demonstrated. Materials and Methods Bacteria . A. brasilense JM125A 2 and A. lipoferum SpUSA5b were obtained from N. R. Krieg. Mutant strains derived from them are listed in Table 2-1. Media. The succinate/mineral salts medium described by Okon et al . (50) was used for nitrogen-free growth. The same medium supplemented with 0.5% (NH.KSO, was used as minimal agar (MA). Carbon-free agar (CFA) was MA lacking succinate. Carbon source utilization tests were done on MA with the appropriate carbon source replacing succinate at a concentration of 1%. Sensitivities of bacteria to inhibitors were determined on nutrient agar (Difco) supplemented with 0.01% yeast extract and filter-sterilized antibiotics or separately-autoclaved heavy metal salts.

PAGE 38

33 Curing . Bacteria were treated with AO as described by Hirota (30), except that the pH was 8.0. The maximum concentration of AO which consistently permitted growth was 2.5 p g/ral Detection of DNA molecules . The electrophoretic method used to detect the various DNA molecules present in wild type and mutant strains has been extensively described elsewhere (see Chapter I). Results Table 2-1 lists the types of mutants isolated from cultures of A. brasilense JM125Aand A. lipoferum SpUSA5b grown in 2.5 P g/ral AO at pH 8.0 and the approximate frequencies with which they were found to occur in such cultures. No spontaneous mutants (from untreated cultures) were found. Isolat i on of JM125A mutants with the T phenotype . Originally, we screened JM125A„ colonies derived from AOtreated cultures for a number of properties: 1) growth on minimal agar plates containing arabinose, galactose, ethanol, or butanol as sole carbon source; 2) sensitivity to trimethoprim (since wild type Azospirillum strains are resistant (3)); and 3) growth on succinate/mineral salts media with and without (NH 4 ) 2 S0 4 . No auxotrophic, Nif~, Ara~, Gal", or trimethoprim-sensitive mutants were found. However, mutants unable to grow on ethanol or butanol were isolated at a high frequency (Table 2-1). Further testing showed that acetate or acetaldehyde would support growth of the mutants.

PAGE 39

34 Somewhat surprisingly, wild type colonies grew on the medium (intended as a control) to which no carbon source had been added. Growth on the carbon-free agar was slower and less luxurious than growth on any of the media containing added carbon. The mutant colonies failed to grow appreciably on the carbon-free medium. Although it appeared from these observations that Azospirillum was capable of autotrophic growth and that the mutants had lost this ability, both wild type and mutant strains produced a visible band of growth when a small inoculum (100 cells) was introduced into the carbon-free soft agar (0.02%) under a predominantly H^/CO^ atmosphere. Others have recently described autotrophic growth of Azospirillum (60). Since heavy metal ion resistance is a characteristic often associated with plasmids in other bacteria (12, 42, 63, 71), we assessed the sensitivities of the mutant and wild type strains to the salts of seven heavy metals. No differential sensitivity to Ag , Cu , Co , Pb , Hg , or Ni was observed; however, 3x10 M Cd permitted growth of the wild type but not the mutant strains. Further analysis showed the mutants to be 3 0-fold more sensitive than the wild type to Cd + (Table 2-2). We have designated the "T" phenotype to represent the triad of cadmium sensitivity, inability to metabolize alcohols, and failure to grow on carbon-free agar under ambient atmosphere.

PAGE 40

35 Loss of pAZ4 determines the T phenotype in JM125A ? . Figure 2-1 shows the DNA molecules of wild type and mutant JM125A„ strains fractionated on an agarose gel as described (see Chapter I). All but one of the T mutants exhibited the DNA banding pattern represented by 125-T2 (lane D), which is identical to that of the wild type except for the absence of pAZ4 (cf. lanes A and D). In contrast, mutant strain 125-T1 appeared to have lost not only pAZ4 but pAZ7 and parts of pAZ2, pAZ3, pAZ5, and pAZ6 as well (lane B). Table 2-3 compares the estimated sizes of the 125-Tl DNA molecules to those of wild type JM125A„. We attempted to determine what functions, if any, could be ascribed to the DNA missing in 125-Tl but not in 125-T2. Wild type JM125A and its T derivatives displayed similar sensitivities to UV light, and all reduced nitrate. As indicated in Table 2-2, however, 125-T2 was 10-fold and 125-Tl was 100-fold more sensitive to kanamycin than was the wild type. The sensitivities of the three strains to streptomycin were identical. Isolation of a Nif derivative of 125-Tl . Despite repeated screening of colonies derived from nitrosoguanidine (NG) or AO-treated JM125Acultures, we had never been able to isolate a mutant with the Nif phenotype. The deletion of approximately 800 Mdal of DNA in 125-Tl (Table 2-3) suggested that at least 20% of the DNA in JM125Ais nonessential. This inference raised the possibility that our failure to isolate Azospirillum Nif strains by conventional

PAGE 41

36 mutagenesis might be due to occurrence of the nif genes in more than one copy. Therefore, we screened 2000 colonies derived from AO-treated 125-Tl cultures for growth on succinate/mineral salts medium with and without (NH.)„SO.. One mutant was isolated which grew on the minimal medium only when nitrogen was supplied. This strain (125-T1N1) also failed to reduce acetylene even in the presence of low concentrations of (NH.) 2 SO. or yeast extract. The electrophoretic mobilities of its DNA molecules, however, could not be distinguished from those of its parent strain (Fig. 2-1, lanes B and C ) . Mutants derived from AO-treated SpUSA5b cultures . We wondered whether types of mutants similar to the ones isolated from A. brasilense JM125A ? could be isolated from A. lipoferum SpUSA5b. Only one of 2000 SpUSA5b colonies derived from AO-treated cultures displayed the T phenotype; thus, T mutants appear to occur at a somewhat lower frequency in SpUSA5b than in JM125A-. An unexpected finding was the high frequency of Met auxotrophs (Table 2-1); no such mutants had been isolated from JM125A 2 . Figure 2-2 illustrates the DNA banding patterns of SpUSA5b and its mutant derivatives. As was the case with JM125A-, the T phenotype correlated with loss of pAZ4 (lanes A and B). However, we did not observe obvious differences in electrophoretic mobility between the DNA molecules of wild type SpUSA5b and those of a representative Met derivative ( lanes A and C) .

PAGE 42

37 125-T1N1 and USA-MI are nonreverting . To investigate the possibility that 125-T1N1 and USA-MI arose from AO-induced frameshift mutations, we attempted to isolate revertants of these strains from untreated, AO-treated, and NGtreated cultures. No true revertants were found. Slowly-growing — 8 colonies did appear at a frequency of approximately 10 when untreated 125-T1N1 was plated on nitrogen-free agar. However, when restreaked on fresh nitrogen-free plates they failed to flourish, and they were unable to reduce acetylene. Furthermore, their isolation frequency was enhanced to a much greater extent by NG than by AO. A similar situation occurred with potential Met revertants of USA-MI. Colonies arising on minimal agar grew when restreaked on fresh minimal plates, but growth on all types of media was slower than wild type growth. Like the pseudorevertants of 125-T1N1, their isolation frequency was increased significantly by NG but only marginally by AO. These Met strains also displayed a deep reddish pigmentation, whereas colonies of our other SpUSA5b strains are peach colored. Discus s ion The results show that pAZ4-JM125A 2 and pAZ4-SpUSA5b R + determine, for their respective hosts, the Cad , Adh , and Cfa phenotypes. The two plasmids are not entirely homologous, however, since the electrophoretic mobility of pAZ4-SpUSA5b is slightly lower than that of pAZ4-JMl25A 2 (see Chapter I). Loss of pAZ4-JM125A 2 at a high frequency

PAGE 43

33 in AO-treated JM125Acultures is consistent with the known curing effects of acridine dyes (30). On the other hand, high-frequency AO-induced methionine auxotrophy and AO-induced multiple deletions in DNA have not been described. The sensitivities of wild type and mutant Azospirillum strains to cadmium are, interestingly enough, comparable to R S the sensitivities exhibited by Cad and Cad Staphylococcus aureus strains (11). Recently, cadmium resistance in S. aureus has been shown to depend on an energy-dependent, plasmid-encoded efflux system (53, 69). A similar system may operate to confer cadmium resistance in Azospirillum . Alternatively, the configuration of membrane proteins in Azospirillum T mutants may result in a lesser degree of thiol group shielding than occurs in the wild type (58). Growth of the T mutants on acetate and acetaldehyde but not on ethanol implies the existance of an alcohol dehydrogenase encoded by pAZ4-JM125A 2 and pAZ4-SpUSA5b. The basis for the inability of the T mutants to flourish on carbon-free agar under ambient atmosphere is less obvious. The defect does not appear to be in the capacity for autotrophy per se , as both mutant and wild type strains grow in carbon-free soft agar under an autotrophic atmosphere. Biosynthesis of ribulose diphosphate carboxylase in other bacteria has been shown to be repressed under conditions of high oxygen tension (37). Therefore, a single cell plated on carbon-free agar under ambient atmosphere cannot immediately begin to grow

PAGE 44

39 autotrophically ; it must instead utilize some intracellular carbon reserve to grow and divide several times, so that an aerobic layer of cells is formed under which a microaerophilic environment is created. Azospirillum strains do, in fact, accumulate large amounts of poly-B-hydroxybutyrate (PHB), particularly (but not exclusively) when grown in nitrogen-free media (49). Although the T mutants as well as the wild type can be seen microscopically to contain PHB granules, it is possible that they have difficulty utilizing this storage polymer for growth. The discovery that one of the T mutants (125-T1) had suffered multiple deletions in its DNA molecules is an observation which may be relevant to the proposed mode of action of AO in plasmid elimination. Some investigators have suggested that plasmids, owing to their relatively small size, are more accessible to AO than is the bacterial chromosome (52) or that AO selectively inhibits plasmid replication (31, 32). In contrast, others have maintained that AO causes nonspecific loss of DNA via inhibition of polymerase I (4, 5); since only cells having lost nonessential DNA survive, this nonspecific mode of action translates into an apparent specificity for plasmid (i.e. nonessential) DNA. Assuming that the generation of 125-T1 really was AO-mediated and not merely a spontaneous event which happened to occur in an AO-treated culture, our results tend to support the idea of a nonspecific interaction of AO with DNA. If AO specifically interacted with or inhibited replication of

PAGE 45

40 plasmid DNA, one would not expect to observe partially deleted plasmids unless those plasmids were capable of dissociation into self-replicating component molecules. The possibility of four DNA molecules undergoing dissociation of this kind within the same cell seems remote. Wild and mutant JM125Astrains displayed three levels of sensitivity to kanamycin, correlating with degree of DNA loss. This tends to implicate two or more proteins or one protein encoded by two or more loci (on at least two DNA molecules) in the determination of kanamycin sensitivity levels in Azospiril lum . Furthermore, these proteins must act specifically on kanamycin and not on aminoglycosides in general, since the sensitivities of the various mutants to streptomycin were identical. Several types of aminoglycoside modifying enzymes have been described in other bacteria, but the mechanism by which the modified antibiotic confers resistance is not known (17). The mechanism of aminoglycoside uptake by sensitive cells is also obscure (17). The Met" derivatives of SpUSA5b and the Nif derivative of 125-T1 possess DNA molecules whose electrophoretic mobilities cannot readily be distinguished from those of their parent strains. These phenomena may be accounted for in one of three ways. First, the mutants might have arisen via the action of AO as a frameshift mutagen (15, 52). This possibility seems unlikely in view of the observed inability of AO to promote reversions in the mutant strains. Furthermore, the frameshift action of acridine dyes has been described

PAGE 46

41 primarily in bacteriophage (15, 52); frameshift mutagenesis of bacteria at the frequencies reported here has not been described. A second possibility is that the mutants are deleted for all or part of a molecule not identified by the electrophoretic method used. This explanation also seems rather unlikely since we have used the method to identify molecules as large as the E. coli chromosome (see Chapter I). The third, and most likely, possibility is that the mutants carry deletions not large enough to lead to obvious differences in mobility for the affected molecules. Thus, 125-T1N1 might have arisen by deletion of the entire nif cluster from pAZl-JM125A 2 (1500 Mdal); the wild type and mutant pAZlJM125A_ molecules, differing in molecular weight by only 1%, would not be resolved by electrophoresis. We are currently attempting to determine whether this is the case by hybridizing a labeled nif probe to Southern blots (65) of fragmented mutant and wild type molecules.

PAGE 47

42 Table 2-1. Azospirillum mutants isolated after growth in acridine orange ™ j. r.j_ ^«j-j. t-.i *. a Isolation Parent Strain Mutant Pnenotype _ frequency JM125A 2 125-T1 Adh~Cf a~Cad S Kan S 2 5xl0~4 JM125A 2 125-T2 Adh~Cf a"Cad S Kan S l 5xl0~3 125-T1 125-T1N1 Adh"Cf a~Cad S Kan S 2Nif " 5xl0~4 SpUSASb USA-T100 AdhCf a Ca d S Kan S l 5xl0"4 SpUSA5b USA-MI Met" 5xl0"3 Abbreviations : Adh_= No growth on ethanol or butanol Cfa = No growth on minimal medium lacking added carbon Cad^= Sensitive to 3xl0~5 M CdCl 2 Kan q l= Sensitive to 3 mg/ml kanarnycin Kan 2= Sensitive to 0.3 rng/ml kanarnycin Nif~= No growth on nitrogen-free medium; no acetylene reduction Met = Requires methionine for growth

PAGE 48

4 3 Table 2-2. Sensitivities of wild type and mutant Azospirillum strains to cadmium, kanamycin^ and streptomycin Minimal inhibitory concentration Strain [Cd2+] ,M [Km], yg/ral [Sm],yg/ml jmi25a 2

PAGE 49

44 Table 2-3. Estimated sizes of DNA molecules found in JM125A 2 an( j 125-T1

PAGE 50

45 B C Fig. 2-1. Agarose gel electrophoresis showing DNA molecules recovered from wild type and mutant JMl25Ap strains: (A) JM125A 2 ; (b) 125-Tl; (c) 125-TlNl; (D) 125-T2. Electrophoresis was for 23 h at 90 V.

PAGE 51

K6 Fig. 2-2. Agarose gel electrophoresis showing DNA molecules isolated from wild type and mutant SpUSA5b strains: (A) SpUSA5b; (b) USA-TlOO; (C) USA-Mi. Electrophoresis was at 80 V for 31 h.

PAGE 52

CHAPTER III Nif GENE HYBRIDIZATION STUDIES OF Azospirillum DNA MOLECULES Introduction In previous chapters we showed that strains of Azospirillum brasilense and A. lipoferum harbor unique arrays of large circular DNA molecules and that phenotypically altered strains, some exhibiting a change in plasmid array, could be isolated at a high frequency from cultures treated with acridine orange. We now address the question of which of these molecules determines the phenotype of greatest general interest, i.e. nitrogen fixation. We had hoped that the Nif" derivative of JM12 5A 2 would show a change in the mobility of one or more of its DNA molecules, but such was not the case. We therefore decided to try a different approach to the problem, based on molecular hybridization. The structural genes for nitrogenase are thought to have been either introduced recently in evolutionary history into the various nitrogen-fixing bacteria or conserved to a greater extent than other translated prokaryotic genes (59). The basis for this view is the observation that Klebsiella pneumoniae nif structural genes can hybridize to DNA from all types of nitrogen-fixing prokaryotes, including both Gram-negative and Gram-positive bacteria, Actinomycetes, and Cyanobacteria (59). This interspecies homology has been exploited to study the organization of nif genes in 47

PAGE 53

48 blue-green algae (39). Although Azospirillum DNA has not been examined for homology to Klebsiella nif DUA, there is no reason to think that Azospirillum would behave differently from other nitrogen-fixing prokaryotes in this regard, especially since the individual Azospirillum nitrogenase subunit proteins can form active complexes in vitro with complementary proteins from other nitrogen-fixing bacteria (23). We obtained, from W. Klipp (36), an Escherichia coli strain harboring pWK27, a recombinant plasmid which includes an EcoRI fragment carrying the nif K, nif D, and nifH genes from K. pneumoniae . This molecule was isolated in large quantity, labeled to high specific activity by nick translation (57), and hybridized to Southern blots (65) of gels containing DNA molecules from wild type and mutant Azospirillum strains. The very limited success we have achieved with this technique indicates that pAZl, the largest of the Azospirillum molecules, carries the nif structural genes in Azospirillum . Materials and Methods Bacterial strains . HB101(pWK27) was supplied by W. Klipp (36). Origins of Azospirillum strains are described in Table 1-1 (Chapter I) and Table 2-1 (Chapter II). Isolation of pWK27 . To one liter of cell culture grown at 37°C to Klett 90 was added 170 mg chloramphenicol. The cells were incubated at 37°C for 20 h, harvested, washed with 10 mM NaCl, and resuspended in 6 ml 0.02 M EDTA, 0.025

PAGE 54

4 9 M Tris, 0.9% glucose (pH 8.0). Lysozyme (12 mg) was added and the cells were chilled on ice for 30 min. Next, 12 ml 0.08% NaOH, 0.8% SDS was added and the lysate was gently swirled for 5 min. Following addition, with gentle mixing, of 9 ml KAc (pH 4.8), the lysate was incubated on ice for 2 h. The lysate was then centrifuged at 15,000 rpm for 30 min and the supernatant solution was transferred to a fresh tube. The addition of PEG 6000 to a final concentration of 10% followed by incubation on ice for 2 h caused the plasmid to precipitate and it was pelleted at 2,500 rpm for 5 min. The pellet was resuspended in 2 ml 50mM Tris, ImM EDTA (pH 8.0). Ribonuclease was added to a concentration of 5 \i g/ml and the solution was incubated at 37 °C for 30 min. Following adjustment of the volume to 10 ml with 50 mM Tris, ImM EDTA (pH 8.0), the solution was extracted twice each with phenol /chloroform, chloroform, and ether. Residual ether was blown off with air. The plasmid was then ethanol precipitated, washed in 70% ethanol, and further purified by dye bouyant density ultracentrifugation. Plasmid bands, visualized with UV light, were removed from the centrifuge tubes with a plastic syringe. Ethidium bromide was removed by extraction with isopropanol and CsCl by dialysis against three changes (2 1 each) of 2 5 mM Tris, 1 mM EDTA. Blotting . Agarose gels containing separated Azospirillum DNA. molecules were prepared as described in Chapter I. DNA from the gels was blotted onto strips of nitrocellulose according to the method of Wahl et al. (70). This is a

PAGE 55

50 slight modification of Southern's original technique (65). The gels are treated with 0.25 M HC1 as an initial step in order to partially depurinate the DNA and fragment it for more efficient transfer. Nick translation of pWK2 7 . The probe was labeled with 32 P according to the method described by Rigby et al . (57). 32 One hundred microCuries Pa-dCTP was dried under vacuum in a 1.5 ml microfuge tube. To this was added 5 pi 0.5 M Tris, 0.1 M 3-mercaptoethanol, 0.05 M MgCl 2 ; 24 p 1 of a 1:1:1 dNTP mix (200 mM each); IP g pWK27 DNA; and water to make 48 p 1 total volume. After 5 min at 30°C, 1 pi diluted, activated DNase I was added. (DNase was activated by diluting a 1 mg/ml stock in 10 mM HC1 1:2000 into 10 mM Tris, 5 mM MgCl 2 , 1 mg/ml BSA (pH 7.6)). After 2 min at 30°C, 0.5 pi DNA polymerase I (5 U/ml ) was added and the reaction was held at 15°C for 1 h, whereupon it was terminated by the addition of 5 Pi 0.25 M EDTA. The reaction mixture was extracted with phenol and the labeled plasmid was separated from the unincorporated nucleotides by passage through a Sepharose 4B or Sephadex G-100 column or by electrophoresis through a 0.7% low melting point agarose gel. Hyb ridi zation . All hybridizations and prehybridizations were carried out in Sears Seal-n-Save plastic bags. Filters were prehybridized for a minimum of 12 h at 65°C in a solution consisting of 0.5% SDS, 20 p g/ml denatured salmon sperm DNA, lOx Denhardt's solution (20), 2 . 5x SSC, and 0.05 M Na^HPO. (pH 8.0). Hybridizations were carried out under the

PAGE 56

51 same conditions except that lx Denhardt's solution was used and denatured, labeled pWK27 was included. The washing protocol of Thomashow et al. v/as followed (68). This consists of one 30 min wash with 2.5x SSC at room temperature followed by four washes (30 min each) with 2 . 5x SSC at 65°C and one wash at 65°C with O.lx SSC. Autoradiography . The dried filters were exposed to X-ray film at -70°C in the presence of an intensifying screen. Exposure times were 2 h for the control hybridization of pWK27 to its restriction fragments and a minimum of 1 week for the Azospirillum hybridizations. Results and Discussion Table 3-1 summarizes the variations in protocol and outcomes of five hybridization experiments. Both Fig. 3-2 and Fig 3-4 were taken from Experiment 2. Despite the somewhat disappointing results, we can conclude from Fig. 3-4 that the Azospirillum nif structural genes are carried on pAZl, the molecule which probably represents the chromosome of these bacteria. Since the control hybridization (Fig. 3-2) gave such a strong signal it is doubtful that our difficulty in detecting Azospiril lum nif sequences reflects a serious flaw in experimental procedure. Rather, the problem seems to be one of sensitivity; the autoradiogram illustrated in Fig. 3-4 required 8 days exposure and shows only faint bands of hybridization. There would appear to be two explanations for this low level of hybridization. First, assuming that

PAGE 57

52 the Azospirillum nif genes are carried on pAZl, less than 0.5% of the DNA in the band representing pAZl is capable of hybridizing to the probe. Second, since there is some degree of divergence in the DNA sequences of nif structural genes in various bacteria (59), the probe DNA is not entirely homologous to that portion of pAZl which is capable of hybridization. DNA preparations from different strains of Rhizobium hybridize with differing intensities to Klebsiella nif DNA (59). We do not think that the low level of hybridization exhibited by Azospirillum DNA reflects poor transfer from gel to nitrocellulose. Gels were always restained and examined for DNA after transfer; none was ever found.

PAGE 58

53 +j c 1 p E E m nj O H --H P -P >i N P a) O W c o P X -P -P Q C O •h 1 p •H > H f0 4-> a> (C o p O Dj rH 14-t IH rH O o a, CO 4-) a. 0)00

PAGE 59

r 5 4 A B Fig. 3 _ 1Agarose gel electrophoresis of pWK27 digested with Hindi I I (D-F) or EcoRI (G-l). Some undigested OC and CCC plasmid DNA is visible. Lanes (A-C) and (j) show lambda Hindlll fragments used as size standards. Volumes of DNA solutions added to wells were as follows: (A) 1 ul; (B,J) 2 ul; (C) 4 ul; (D,G) 3 ul; (E,H) 6 ul; (F,l) 9 ul. Electrophoresis was at 60 V for I) 6 ul; 4 h.

PAGE 60

55 Fig. 3-2. Autoradiogram of gel shown in Fig. 3-1 • Gel was blotted and hybridized to labeled pWK27 as described in the text. Exposure was for 2 h. (A-C,J) Hindlll-digested lambda; (D-F) Hindlll-digested pWK27; (G-I) EcoRI-digested pWK27 .

PAGE 61

56 Fig. 3 3> Agarose gel electrophoresis showing DNA molecules recovered from Azospirillum strains of diverse geographic origin. Electrophoresis was for 30 h at 80 V. A. brasilense strains: (A) Sp7; (B) Spl3t; (C) Sp84; (D) Cd; (E) JM82A,; (F) JMl25Ap, A. lipoferum strains: (g) SpUSA5b; (h) SpRGDxx.

PAGE 62

57 D E Fig. 3~^' Autoradiogram of gel shown in Fig. 3 3« The gel was trimmed of material above pAZl, blotted onto nitrocellulose, and hybridized to labeled pWK27 as described in the text. Exposure was for 8 days. A) Sp7; (B) Spl3t; (C) Sp84; (D) Cd; (E) JM82A.,; F) JM125A 2 j (G) SpUSA5bj (H) SpRG6xx.

PAGE 63

APPENDIX A NORMALIZATION OF GEL MOBILITY DATA In order to calculate a regression line relating electrophoretic mobility to molecular weight, it was necessary to pool mobility data from 20 gels run for various lengths of time under slightly differing conditions. Therefore, the mobility data from 19 of the gels had to be normalized to the data from one gel chosen as a standard. The gel illustrated in Fig. 1-6 was chosen as the standard because it includes the greatest number (15) of DNA bands. The standard gel was designated B and the mobilities of its DNA bands were B L , B 2 , , B n » The mobilities of the bands in gel A (which had in common with gel B n different plasmid bands numbered l,2,...,n) were designated A l' A 2' • • • ' A n* ™* le constant K a was calculated for which the expression Z I K a A i -B i l/(K a A i +B i ) where i=l, 2, . . . ,n was at a minimum. Then the mobilities of the plasmids in gel A were multiplied by K a different K = was calculated for each gel a. a. to be normalized. 58

PAGE 64

APPENDIX B REGRESSION LINE CALCULATIONS If the logarithms of the average normalized mobilities of the standard plasmids were designated as X^, X2,...X n and the logarithms of their molecular weights (in megadaltons) were designated Y lf Y 2 ,...Y n then the logarithm of the molecular weight (Y fc ) of the unknown molecule could be calculated from the logarithm of its mobility (X t ) according to the formula Y t = A+BX t where A=T-BX EX 2 I (EX.) 2 n X = EX./n Y = EY./n 59

PAGE 65

LITERATURE CITED 1. Aaij , C, and P. Borst. 1972. The gel electrophoresis of DNA. Biochim. Biophys. Acta 269 ;192-200. 2. Ahmad, M. H. 1978. Influence of nitrogen on growth, free amino acids and nitrogenase activity in Spirillum lipof erum . J. Gen. Appl. Microbiol. 24 : 271-278 . 3. Albrecht, S. L. , and Y. Okon. 1980. Cultures of Azospirillum , p. 746. In A. San Pietro (ed.), Methods in enzymology, vol. 69. Academic Press, New York. 4. Barker, G. R. 1978. Genetic expression and its control in naturally occurring bacterial plasmids, pp. 25-29. In P. W. Kent (ed.), New approaches to genetics: Developments in molecular genetics. Oriel Press, Boston. 5. Barker, G. R. , and N. Hardman. 1978. The effects of acridine orange on deoxyribonucleic acid in Escherichia coli . Biochern. J. 171 :567-573. 6. Brent, T. P. 1972. Repair enzyme suggested by mammalian endonuclease activity specific for ultravioletirradiated DNA. Nature (London), New Biol. 239 :172-173. 7. Bujard, H. 1970. Electron microscopy of single-stranded DNA. J. Mol. Biol. 49:125-137. 8. Cairns, J. 196 3. The chromosome of Escherichia coli . Cold Spring Harbor Symp. Quant. Biol. 28: 43-45. 9. Carr, T. C. 1978. Cryptic plasmids in Azospirillum . Master's Thesis, University of Florida, Gainesville. 10. Casse, F., C. Boucher, J. S. Julliot, M. Michel, and J. Denarie. 1979. Identification and characterization of large plasmids in Rhizobium meliloti using agarose gel electrophoresis. J. Gen. Microbiol. 113 : 229-242. 11. Chopra, I. 1975. Mechanism of plasmid-mediated resistance to cadmium in Staphylococcus aureus . Antimicrob. Agents Chemother. 7_:8-14. 12. Clark, D. , A. A. Weiss, and S. Silver. 1977. Mercury and organomercurial resistances determined by plasmids in Pseudomonas. J. Bacteriol. 132:186-196. 60

PAGE 66

61 13. Clewell, D. B. , and D. R. Helenski. 1969. Supercoiled circular DNA-protein complex in Escherichia coli : Purification induced conversion to an open circular DNA form. Proc. Natl. Acad. Sci. U.S.A. 62 ; 1159-1166. 14. Cooper, S., and C. E. Helmstetter. 1968. Chromosome replication and the division cycle of Escherichia coli. B/r. J. Mol. Biol. 3JL :519_54 °15. Crick, F. H. C, L. Barnett, S. Brenner, and R. J. Watts-Tobin. 1961. General nature of the genetic code for proteins. Nature (London) 192 ; 1227-1232. 16. Currier, T. C, and E. W. Nester. 1976. Evidence for diverse types of large plasmids in tumor-inducing strains of Agrobacterium . J. Bacteriol. 126 : 157-165. 17. Davis, J., and D. I. Smith. 1978. Plasmid-determined resistance to antimicrobial agents. Ann. Rev. Microbiol. 32:469-518. 18. Davis, R. W. , M. Simon, and N. Davidson. 1971. Electron microscope heteroduplex methods for mapping regions of base sequence homology in nucleic acids. In L. Grossman and K. Moldave (ed.), Methods in enzymology, vol. 21. Academic Press. New York. 19. Day, J. M. , and J. Dobereiner. 1976. Physiological aspects of N -fixation by a Spirillum from Digitaria roots. Soil Biol. Biochem. £: 45-50. 20. Denhardt, D. 1966. A membrane filter technique for the detection of complementary DNA. Biochem. Biophys. Res. Commun. 23: 641-646. 21. Dobereiner, J., I. E. Marriel, and M. Nery. 1976. Ecological distribution of Spirillum lipoferum Beijerinck. Can. J. Microbiol. 22: 1464-1473. 22. Eckhardt, T. 1978. A rapid method for the identification of plasmid desoxyribonucleic acid in bacteria. Plasmid JL:534-583. 23. Emerich, D. W. , and R. H. Burris. 1978. Complementary functioning of the component proteins of nitrogenase from several bacteria. J. Bacteriol. 134 :936-943. 24. Fangman, W. L. 1978. Separation of very large DNA molecules by gel electrophoresis. Nucleic Acids Res. _5:653-665. 25. Fisher, M. P., and C. W. Dingman. 1971. Role of molelecular conformation in determining the electrophoretic properties of polynucleotides in agarose-acrylamide composite gels. Biochemistry 10 : 1895-1899.

PAGE 67

62 26. Franche, C, and C. Elmerich. 1981. Physiological properties and plasmid content of several strains of Azospirillum brasilense and A. lipoferum . Ann. Microbiol. (Inst. Pasteur) 132 A: 3-18. 27. Gauthier, D. , and C. Elemrich. 1977. Relationship between glutamine synthetase and nitrogenase in Spirillum lipoferum . FEMS Microbiol. Letters 2_:101-104. 28. Gillis, M., J. De Ley, and M. De Cleene. 1970. The determination of molecular weight of bacterial genome DNA from renaturation rates. Eur. J. Biochem. 12:143-153, 29. Hansen, J. B. , and R. H. Olsen. 1978. Isolation of large bacterial plasmids and characterization of the P2 incompatibility group plasmids pMGl and pMG5. J. Bacteriol. 135 :227-238. 30. Hirota, Y. 1960. The effect of acridine dyes on mating type factors in Escherichia coli . Genetics 46 : 57-64. 31. Hohn, B., and D. Korn. 1969. Cosegregation of a sex factor with the Escherichia coli chromosome during curing by acridine orange. J. Mol. Biol. 45 : 385-395. 32. Jacob, F., S. Brenner, and F. Curzin. 1963. On the regulation of DNA replication in bacteria. Cold Spring Harbor Symp. Quant. Biol. 28:329-348. 33. Kato, A. C., and M. J. Fraser. 1973. Action of a singlestrand specific Neurospora crassa endonuclease on ultraviolet light-irradiated native DNA. Biochem. Biophys. Acta 312 :645-655. 34. Kline, B. C., and J. R. Miller. 1975. Detection of nonintegrated plasmid deoxyribonucleic acid in the folded chromosome of Escherichia coli : Physicochemical approach to studying the unit of segregation. J. Bacteriol. 121 : 165-172". 35. Kline, B. C., J. R. Miller, D. E. Cress, M. Wlodarczyk, J. J. Manis, and M. R. Otten. 1976. Nonintegrated plasmid-chromosome complexes in Escherichia coli . J. Bacteriol. 127 :881-889. " "" 36. Klipp, W. 1980. Personal communication. 37. Kuehn, G. D. , and B. A. McFadden. 1968. Factors affecting the synthesis and degradation of ribulose1, 5-diphosphate carboxylase in Hydrogenomonas facilis and Hydrogenomonas eutropha . J. Bacteriol. 95:937-946.

PAGE 68

63 38. Magalhaes, L. M. S., C. A. Neyra, and J. Dobereiner. 1978. Nitrate and nitrite reductase negative mutants of N„-fixing Azospirillum spp. Arch. Microbiol. 117:247-252. 39. Mazur, B. J., D. Rice, and R. Haselkorn. 1980. Interspecies homology of nitrogenase genes. Proc. Natl. Acad. Sci. USA 1J_: 191-195. 40. Meyers, J. A., D. Sanchez, L. P. Elwell, and S. Falkow. 1976. Simple agarose gel electrophoretic method for the identification and characterization of plasmid deoxyribonucleic acid. J. Bacteriol. 127 :1529-1537. (Erratum, J. Bacteriol. 129 :1171, 1977.) 41. Mishra, A. K. , P. Roy, and S. Bhattacharya. 1980. Deoxyribonucleic acid-mediated transformation of Spirillum lipoferum . J. Bacteriol. 137 :1425-1427. 42. Nakahara, H. , T. Ishikawa, S. Sarai, I. Kondo, and S. Mitshuhashi. 1977. Frequency of heavy metal resistance in bacteria from inpatients in Japan, Nature 266 : 165-167, 43. Nelson, L. M. , and R. Knowles. 1978. Effect of oxygen and nitrate on nitrogen fixation and denitrif ication by Azospirillum brasilense grown in continuous culture Can. J. Microbiol. 24:1395-1403. 44. Neyra, C. A., J. Dobereiner, R. Lalande, and R. Knowles. 1977. Denitrif ication by N 2 ~fixing Spirillum lipoferum . Can. J. Microbiol. 2^3:300-305. 45. Neyra, C. A., and P. Van Berkum. 1977. Nitrate reduction and nitrogenase activity in Spirillum lipoferum . Can. J. Micbriol. 2_3: 306-310. 46. Nur, I., Y. Okon, and Y. Henis. 1980. An increase in nitrogen content of Setaria Italica and Zea mays inoculated with Azospirillum . Can. J. Microbiol. _26:432-485. 47. Ohtsubo, E., R. C. Deonier, H. J. Lee, and N. Davidson. Electron microscope heteroduplex studies of sequence relations among plasmids of Escherichia coli . IV. The sequence in F14. J. Mol. Biol. 89: 565-584. 43. Okon, Y., S. L. Albrecht, and R. H. Burris. 1976. Factors affecting growth and nitrogen fixation of Spirillum lipoferum . J. Bacteriol. 127:1248-1254. 49. Okon, Y., S. L. Albrecht, and R. H. Burris. 1976. Carbon and ammonia metabolism of Spirillum lipoferum . J. Bacteriol. 128: 592-597.

PAGE 69

64 50. Okon, Y. , S. L. Albrecht, and R. H. Burris. 1977. Methods for growing Spirillum lipoferum and for counting it in pure culture and in association with plants. Appl. Env. Microbiol. 33.:85-88. 51. Okon, Y., J. P. Houchins, S. L. Albrecht, and R. H. Burris. 1977. Growth of Spirillum lipoferum at constant partial pressures of oxygen, and the properties of its nitrogenase in cell-free extracts. J. Gen. Microbiol. 9^:87-93. 52. Orgel, A., and S. Brenner. 1961. Mutagenesis of bacteriophage T4 by acridines. J. Mol. Biol. _3_ :762 -768. 53. Perry, R. D. , and S. Silver. 1981. Transport studies on cadmium resistance in whole cells and subcellular membranes of Staphylococcus aureus , p. 159. In Abstracts of the annual meeting of the American Society for Microbiology, 1981. ASM Publications, Washington, D.C. 54. Pettijohn, D. E. , and R. Hecht. 1973. RNA molecules bound to the folded bacterial genome stabilize DNA folds and segregate domains of supercoiling. Cold Spring Harbor Symp. Quant. Biol. 38: 31-41. 55. Polsinelli, M. , E. Baldanzi, M. Bazzicalupo, and E. Gallori. 1980. Transfer of plasmid pRDl from Escherichia coli to Azosp i rillum brasilense . Mol. Gen. Genet. 178:709-711. 56. Rainbow, A. J., and S. Mak. 1973. DNA damage and biological function of human adenovirus after U.V.irradiation. Int. J. Radiat. Biol. 24 : 59-72. 57. Rigby, P. W. J., M. Dieckmann, C. Rhodes, and P. Berg. 1977. Labeling deoxyribonucleic acid to high specific activity in vitro by nick translation with DNA polymerase. J. Mol. Biol. 113 :237-251. 58. Rothstein, A. 1959. Cell membrane as site of action of heavy metals. Fed. Proc. 18:1026-1035. 59. Ruvkun, G. B. , and F. M. Ausubel. 1980. Interspecies homology of nitrogenase genes. Proc. Natl. Acad. Sci. USA 72:191-195. 60. Sampaio, J. A. M. , E. M. R. daSilva, J. Dobereiner, M. G. Yates, and F. 0. Pedrosa. 1981. Autotrophy and methylo trophy in Derxia gummosa , Azospirillum brasilense , and A. lipoferum , p. 444. In A. H. Gibson and W. E. ~~ Newton (ed.), Current perspectives in nitrogen fixation. Australian Academy of Science, Canberra City.

PAGE 70

6 5 61. Sharp, P. A., M Hsu, E. Ohtsubo, and N. Davidson. 197 2. Electron microscope heteroduplex studies of sequence relations among plasmids of Escherichia coli . I. Structure of F-prime factors. J. Mol. Biol. 71: 471-487. 62. Silver, S. , E. Levine, and P. M. Spielman. 1968. Acridine binding by Escherichia coli : pH dependency and strain differences. J. Bacteriol. 95: 333-339. 63. Smith, D. H. 1967. R factors mediate resistance to mercury, nickel, and cobalt. Science 156 : 1114-1116. 64. Smith, R. L. , J. H. Bouton, S. C. Schank, K. H. Quesenberry, M. E. Tyler, J. R. Milam, M. H. Gaskins, and R. C. Littell. 1976. Nitrogen fixation in grasses inoculated with Spirillum lipoferum . Science 193 :10031005. 65. Southern, E. M. 1975. Detection of specific sequences among DNA fragments separated by gel electrophoresis. J. Mol. Biol. 98:503-517. 66. Stonington, 0. G. , and D. E. Pettijohn. 1971. The folded genome of Escherichia coli isolated in a protein-DNA-RNA complex. Proc. Natl. Acad. Sci. USA 68:6-9. 67. Tarrand, J. J., N. R. Krieg, and J. Dobereiner. 1978. A taxonomic study of the Spirillum lipoferum group, with descriptions of a new genus, Azospirillum gen. nov. and two species, Azospirillum lipoferum (Beijerinck) comb. nov. and Azospirillum brasilense sp. nov. Can. J. Microbiol. _2_4: 967-980. 68. Thomashow, M. F., R. Nutter, A. L. Montoya, M. P. Gordon, and E. W. Nester. 1980. Integration and organization of Ti plasmid sequences in crown gall tumors. Cell 19 : 729-739. 69. Tynecka, Z., Z. Gos, and J. Zajac. 1981. Plasmiddetermined Cd2 resistance in Staphylococcus aureus : Accelerated efflux and reduced net uptake, p. 159. In Abstracts of the annual meeting of the American Society for Microbiology, 1981. ASM Publications, Washington, D.C. 70. Wahl, G. M., M. Stern, and G. R. Stark. 1979. Efficient transfer of large DNA fragments from agarose gels to diazobenzylozymethlyl-paper and rapid hybridization by using dextran sulfate. Proc. Natl. Acad. Sci. USA 76: 3633-3687. 71. Weiss, A. A., S. D. Murphy, and S. Silver. 1977. Mercury and organomercurial resistances determined by plasmids in Staphylococcus aureus . J. Bacteriol. 132 : 197-208.

PAGE 71

66 72. Willshaw, G. A., H. R. Smith, and E. S. Anderson. 1979, Application of agarose gel electrophoresis to the characterization of plasmid DNA in drug-resistant Enterobacteria. J. Gen. Microbiol. 114 : 15-25. 73. Worcel, A., and E. Burgi. 1972. On the structure of the folded chromosome of Escherichia coli . J. Mol. Biol. TL-.121-1M . 74. Worcel, A., E. Burgi, J. Robinton, and C. L. Carlson. 1973. Studies on the folded chromosome of Escherichia coli. Cold Spring Harbor Symp. Quant. Biol. 38:43-51.

PAGE 72

BIOGRAPHICAL SKETCH The author was born Alvin Gleave Wood on October 5, 1951, in St. Petersburg, Florida. He is the only son of Mary Gleave Harris and the late Rowland Emery Wood. Following graduation from Northeast High School in St. Petersburg, he attended the University of Chicago and received a B.A. in biology in June 1973. He received his M.S. degree from the Department of Microbiology and Cell Science, University of Florida, 1978. He is currently a candidate for the degree of Doctor of Philosophy, also in the Department of Microbiology and Cell Science. 67

PAGE 73

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in icope and quality, as a dissertation for the degree of Doctorof Philosophy. /, Dennis. E. Duggan, Chairman Associate Professor of Microbiology and Cell Science I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the deqree of Doctor of Philosophy.) — .<:", I -t . , ^" 7 ' . • V Francis C. Davis, Jr. / Associate Professor of Microbiology and Cell Science I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Lonnie 0. Ingram Associate Professor of Microbiology and Cell Science I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. ^ «S*' William W. Hauswirth Associate Professor of Immunology and Medical Microbiology

PAGE 74

This dissertation was submitted to the Graduate Faculty of the College of Agriculture and to the Graduate Council, and was accepted as partial fulfillment of %he requirements for the degree of Doctor of Philosophy. ( ) May 1982 XJ
PAGE 75

UNIVERSITY OF FLORIDA llllllllli 3 1262 08553 2124