JOURNAL OF BACTERIOLOGY, Dec. 1980, p. 1152-1158 Vol. 144, No.

3
0021-9193/80/12-1152/07$02.00/0

Myceloid Growth of Arthrobacter globiformis and Other

Arthrobacter Speciest
JAMES J. GERMIDA AND L. E. CASIDA, JR.*
Department of Microbiology, The Pennsylvania State University, University Park, Pennsylvania 16802

Transitory myceloid growth occurs in certain complex media with Arthrobacter

globiformis strain ATCC 8010. This type of growth, however, was not observed
in a medium which contained an array of metal ions but did not contain agents
able to complex metal ions. Addition of metal-complexing agents to this medium
caused an interruption in the life cycle of strain 8010 so that growth occurred
only as the myceloid form. It appeared that manganese was the critical metal
that was removed by the metal-complexing agents. During growth, the myceloid
cells started to fragment, but wall septation was incomplete. A. globiformis strain
ATCC 4336 and several other Arthrobacter species and soil isolates, but not
Arthrobacter crystallopoietes, responded to metal-complexing agents as did strain
8010. Biotin and vitamin B12 were not involved in this myceloid growth.
Arthrobacter species are found in soil (11-13, plated on soil extract agar. The other isolate came
17) and other natural habitats (23). They exhibit from soil that had been incubated 1 week with the
a pleomorphic life cycle which varies or can even nicotine salts medium of Casida and Rosenfield (4).
be controlled, depending on nutritional condi- Streaking was on nicotine agar, and blue colonies were
tions and growth rate (8, 10, 14, 15, 18, 19, 26). selected.
Media. GYEN agar contained 0.1% glucose, 0.1%
InArthrobacterglobiformis the sequence of this yeast extract, 0.8% nutrient broth, and 1.5% agar. The
life cycle is coccoid cells, myceloid cells, rods, chemically defined glucose-salts medium, which was
and finally, coccoid cells or cystites. Of these, designated as the cation-complete medium, was de-
the myceloid phase may be fleeting or may not scribed by Chan (6). The concentrations (based on
occur at all, and relatively little is known about amounts added) of component salts were: 49.3 mM
it. For example, transitory myceloid growth oc- KNO3, 2.9 mM K2HPO4, 3.7 mM KH2PO4, 0.82 mM
curs during coccoid cell outgrowth in complex MgSO4. 7H20, 0.17 mM NaCl, 36 mM FeSO4. 7H20, 58
media (3, 14, 24, 26, 29). Longer-lasting, ,tM MnSO4.H20.
branched myceloid forms, however, are pro- Sugars added to media at 1% (wt/vol) were prepared
as 30% stock solutions in double-distilled water and
duced by vitamin B12 deficiency (9) or biotin sterilized separately by autoclaving for 15 min at
deficiency (6). Production of the myceloid forms 1210C. A biotin stock solution was prepared in double-
with biotin deficiency is not related to growth distilled water and sterilized by passage through a 0.3-
rate (31). ,um-pore size membrane filter. It was added at 0.4 ng/
The present study considers the effect of ml to cation-complete medium for some of the studies
metal ion deficiency and metal ion-complexing involving A. globifor7nis strain ATCC 4336.
agents on the growth and morphogenesis of A. A medium containing low levels of several metal
globiformis strain ATCC 8010 and some other ions was designated as the cation-deficient medium. It
Arthrobacter species. Of particular interest was differed from the complete medium in that the con-
the effect of Mn2+ on the myceloid phase of centrations of metal salts were: 0.082 mM MgSO4.
7H20, 0.36 ,iM FeSO4.7H20, 0.1 uM MnSO4.H20, 2
growth. ,uM ZnSO4.7H20. This medium was prepared by a
MATERIALS AND METHODS modification of the 8-hydroxyquinoline extraction
Organisms. The principal strain used in this study technique of Waring and Werkman (30). The basal
was A. globiformis ATCC 8010. For some experiments solution (phosphate buffer and KNO3), the glucose
A. globiformis strain ATCC 4336, Arthrobacter oxy- stock solution, and the double-distilled water used for
dans strain ATCC 14358 (yellow and white biotypes), washing cells or preparation of cation salt stock solu-
Arthrobacter crystallopoietes strain ATCC 15481, and tions and solutions of organic growth factors were
five soil isolates were used. The last were tentatively extracted by this technique. For this technique, 200
identified as Arthrobacter species based on morphol- mg of 8-hydroxyquinoline dissolved in 3 ml of CHC13
ogy (28). Four of the isolates were isolated from soil was added to 400 ml of the particular solution or water
that had been suspended in sterile tap water and to be extracted. This was shaken by hand and then
allowed to stand for 30 to 60 min. The metal chelates
t Paper no. 6030 in the journal series of the Pennsylvania that formed were removed by three sequential extrac-
Agriculture Experiment Station. tions with 10-ml portions of CHC13. This entire process
1152
VOL. 144, 1980 MYCELOID GROWTH IN ARTHROBACTER SPECIES 1153
was repeated five times, except that the last addition 0.3-,um-pore size membrane filter. The tripotassium
of 8-hydroxyquinoline was allowed to stand overnight citrate solution was autoclaved, however. When added
before four sequential extractions were performed with to media for growth studies, 8-hydroxyquinoline was
10-ml portions of CHC13. The extracted basal solution dissolved in absolute ethanol (22) and sterilized by
or water was then stored in acid-washed polypropylene filtration. Stock solutions of biotin (Sigma Chemical
flasks (Kimble Products, Toledo, Ohio) at 5°C for no Co., Saint Louis, Mo.), vitamin B12 (Merck & Co., Inc.,
longer than 1 week. The metal cation salts to be added Rahway, N.J.), and yeast extract were dissolved in
to extracted basal solution were made up as individual extracted water and sterilized by autoclaving.
stock solutions with extracted water. These stock so-
lutions were sterilized separately by autoclaving. All RESULTS
glassware used for the above had been soaked in 5 N
HC1 for 24 h, then rinsed 8 to 10 times with double- Growth of A. globiformis strain ATCC 8010
distilled water, and allowed to drain dry. in cation-complete medium is shown in Fig. 1.
Growth studies. For most studies, inoculum was The generation time was 4 h, and more than a
prepared by adding one loopful of cells from a stock 100-fold increase in colony-forming units per
culture slant to 30 ml of cation-complete medium. milliliter occurred during growth. Coccoid cells,
This was shaken for 72 h on a reciprocal shaker at added as the inoculum, proceeded to elongate in
27°C, then 0.3 ml was transferred to 30 ml of fresh the classic manner to form rods without produc-
cation-complete medium and incubated as above. A ing myceloid cells. The rods then divided by
different inoculum medium was used for growth stud- binary fission throughout the exponential phase
ies involving cation-deficient medium. In this case,
cation-complete medium, but without added Mn2", of growth. Their cell lengths were approximately
was prepared with components and water that had two to three times those of the coccoid cells.
been extracted with 8-hydroxyquinoline (as for the Finally, the rods underwent reductive fragmen-
cation-deficient medium). This medium contained ap- tation to yield coccoid cells at the onset of the
proximately 0.1 ,uM Mn2", which was due to trace stationary phase of growth.
contamination (calculation based on manufacturer's Pyrophosphate (7 mM) depressed total
analyses) from other metal cation salts. For both me- growth by approximately 22%, (Fig. 1A), but the
dia, the cells were harvested by centrifugation at 72 h generation time was not noticeably affected. The
(approximately 450 to 500 Klett units [no. 66 filter] or total cell mass (dry weight) obtained in the
109 colony-forming units). The cells were then washed
three times with sterile distilled water or extracted presence of Na4P207 was 55 to 65% of that pro-
water and suspended in the respective fluid at a final duced in its absence. Expression of growth as
concentration of approximately 109 colony-forming colony-forming units per milliliter was depressed
units per ml. This suspension was used to inoculate by 90% (Fig. 1B), and the apparent generation
media for the growth studies and provided approxi- time was 13 h. The difference, as observed by
mately 107 colony-forming units per ml. For the growth microscopy, was due to the formation of myce-
studies, 500-ml nephelometric flasks (Bellco Glass, loid cells that did not fragment. The myceloid
Inc., Vineland, N.J.) contained 30 ml of medium. Cot- cells from the stationary phase of growth were
ton plugs were used in these flasks for studies with
cation-complete medium. Foam rubber stoppers and washed and then incubated in cation-complete
acid-washed glassware were employed for studies with medium lacking glucose and pyrophosphate.
cation-deficient medium. All flasks were incubated at Under these conditions, the myceloid cells pro-
270C on a reciprocal shaker. ceeded to fragment, and the colony-forming
Morphology of the cells was followed by phase- units per milliliter increased to a level approxi-
contrast microscopy of wet mounts and bright-field mately 50% of that for total growth in cation-
microscopy of crystal violet stains of smears. Both complete medium.
types of preparations were photographed, and mea- The coccoid cells, added as the inoculum to
surements of cell size were made on the photomicro- cation-complete medium containing Na4P207,
graphs. Thin sections of bacterial cells were prepared swelled initially so that they resembled cystites
and observed by transmission electron microscopy as
described by Balkwill and Casida (1). (14, 27). During the first 24 h these cystite-like
The numbers of colony-forming units per milliliter cells underwent germination at 2 or 3 points on
were determined by plate counts on GYEN agar. the cell surface (Fig. 2, 17 h), and this was
Chemicals. The components of the complex media followed by elongation and rudimentary branch-
used in these studies were products of Difco Labora- ing (the myceloid form) during the rest of the
tories, Detroit, Mich. All chemicals, including EDTA, growth period (Fig. 2). The total linear length
used for preparing the cation-complete or cation-defi- (exclusive of branches) of individual cells was 10
cient media, and as metal-complexing agents, were to 20 times that of the coccoid or rod-shaped
analytical reagent grade (Fisher Scientific Co., Pitts- cells produced in the absence of Na4P207. The
burgh, Pa.). [Ethylenebis-(oxyethylenenitrilo)]tetra-
acetic acid (EGTA), however, was from J. T. Baker myceloid forms had incomplete septation as re-
Chemical Co. (Phillipsburg, N.J.). The metal-complex- gards their walls (Fig. 3), and the walls along the
ing agents were prepared as stock solutions in double- sides of the cells appeared thin and stretched.
distilled water and sterilized by filtration through a Sequestering agents for metal ions were tested
1154 GERMIDA AND CASIDA J. BACTERIOL.

1010

Ut)
z

Ui.
lI 0
y

6 12 18 24 30 36 42
HOURS HOURS
FIG. 1. Growth plotted as turbidity (A) or colony-forming units (B) for A. globiformis strain 8010. Cation-
complete medium (@); cation-complete medium with 7 mM Na4P207 (0); cation-deficient medium with 0.1
Mn (A); cation-deficient medium without added manganese (A).

for their effects on growth and morphology of concentration had a similar effect when it was
strain 8010. EGTA (0.2 to 7.8 mM) and citrate added to cation-complete medium containing
(15 to 60 mM) added to cation-complete medium either Na4P207, citrate, or EGTA.
caused the myceloid morphology along with par- Strain 8010 was grown in cation-deficient me-
tial growth inhibition. When the concentration dium, which contained low levels of several
of these agents was lower or higher than the metal ions, including Mn2+ (Fig. 1). Under these
range listed, normal growth and morphology or conditions the lag phase, based on turbidity, was
complete growth inhibition, respectively, was 6 to 9 h longer than it was in cation-complete
observed. The use of citrate as the sole carbon medium. The maximum turbidity obtained was
source also caused myceloid cell development. less than 50% of that observed in the complete
Based on all methods of measurement and ob- medium, although the generation time was ap-
servation, the growth patterns for cells treated proximately the same. Colony-forming units in
with EGTA or citrate were similar to those for the deficient medium (Fig. 1B) paralleled the
Na4P207. However, at the lowest active citrate turbidimetric measurements (Fig. 1A), but only
level (15 mM), some fragmentation of myceloid a 10-fold increase in colony-forming units per
cells was observed at the onset of the stationary milliliter occurred. The morphological cycle was
phase of growth. EDTA and 8-hydroxyquinoline as for growth in cation-complete medium, except
could not be used to produce myceloid cells. 8- that overall cell size was smaller. The limiting
Hydroxyquinoline completely inhibited growth cation in this deficient medium was iron. Thus,
at all concentrations tested (1.0 to 10 ,ug/ml). results similar to the above were obtained when
EDTA had no effect at concentrations below 0.8 we used cation-complete medium in which- the
mM but completely inhibited growth at greater iron level had been reduced to match that in
concentrations. cation-deficient medium. Morris (21) and Mid-
Stevenson (26) reported myceloid growth in dleton and Gunner (20) reported pigment pro-
the complex medium of Henderson and Snell. duction by A. globifonnis under conditions of
This medium happens to contain citrate. We iron limitation. We did not observe this pigmen-
repeated Stevenson's experiment and also ob- tation.
served myceloid growth. In addition, however, Growth of strain 8010 in cation-deficient me-
we found that the addition of 3 mM CaCl2 to dium without added manganese is shown in Fig.
this medium to tie up the citrate prevented 1. The effective level of residual manganese was
myceloid cell formation while still allowing the calculated to be approxiimately 1 nM, based on
normal morphological life cycle. CaCl2 at this the reported levels of contaminant manganese
VOL. 144, 1980 MYCELOID GROWTH IN ARTHROBACTER SPECIES 1155
prevented, and fragmentation could not be in-
duced if Ca2", Srt, Ni2", Fe3", or Ca2" plus Fe3+,
instead of Mn2", were added at 0.1 AM. Biotin or
vitamin B12 (both at 0.4 or 4.0 ng/ml) or yeast
extract (1 or 10 gg/ml) did not prevent myceloid
growth when Mn2+ was not added. Also, the
IfI addition of a 1/100 dilution of sterile culture
filtrate from a naturally reverting myceloid cul-
1 o ture had no effect.
Cells were grown for varying periods of time
17 in cation-complete medium without added Mn2+
(approximately 0.1 AM Mn2+ present due to trace
contamination from other salts); then they were
';r -,... I washed and added to cation-deficient medium
|E . X I-~~~~olo without added manganese to test for myceloid
cell formation. Cells harvested from the com-
plete medium at 24 h (short rods) gave 3 to 5 h
of myceloid growth in the deficient medium,
followed by a rapid reversion to the normal life
p
4; t. "' x .. cycle. With 48- to 72-h cells (coccoid), synchro-
nous outgrowth and prolonged myceloid devel-
22 28 opment (lasting at least 30 to 48 h before frag-
menting) were observed in the deficient medium,
.~~~~~
.
. ~ ~ whereas 96- to 120-h coccoid cells demonstrated
SI' asynchronous outgrowth and prolonged myce-
loid development. Cystites of strain 8010, pro-
duced by growth in a minimal medium (14) or a
complex medium (3), responded to Mn2+ defi-
ciency with tripodal germination and prolonged
myceloid growth. All ofthe above cells of various
ages or types, however, produced myceloid cells
that did not fragment when added to cation-
complete medium containing EGTA, Na4P207,
t48 72 N\ or citrate.
A. globiformis strain ATCC 4336, A. oxydans
FIG. 2. Myceloid growth of A. globiformis strain strain ATCC 14358 (yellow and white biotypes),
8010 in cation-complete medium with added 7 mM and five Arthrobacter species soil isolates re-
Na4P207. Numbers on pictures are hour of incubation. sponded with myceloid growth (as for strain
Phase-contrast microscopy. 1,142-fold magnification.
8010) to Na4P207, EGTA, and citrate. In addi-
in the metal salts added to the medium. In this tion, strain 4336 (other species and strains not
case, the coccoid cells added as inoculum elon- tested) produced myceloid cells in cation-defi-
gated to give long filaments that demonstrated cient medium (manganese not added) containing
rudimentary branching (Fig. 4) and reached a 0.4 or 4.0 ng of biotin per ml. A. crystallopoietes
linear cell length 15 to 20 times that of normal strain ATCC 15481 did not grow well enough in
rod-shaped cells (cation-complete medium). the presence of citrate for testing. In the pres-
Thin sections appeared as for Fig. 3, but septa- ence of Na4P207 and EGTA, it grew only as
tion initiations were slightly less frequent. If somewhat. coccoid cells, although its growth was inhibited
Mn2" removal from this medium was complete,
the myceloid cells remained in this form
throughout growth without fragmenting to yield DISCUSSION
coccoid cells (as at 42 h, Fig. 4). Addition of 0.1 The polymorphic growth cycle of A. globifor-
,AM manganese to this medium after the myce- mis strain ATCC 8010 that occurred in cation-
loid cells had formed caused a slow fragmenta- complete medium without added metal-com-
tion to coccoid cells (over a period of 24 to 48 h) plexing agents was similar to that observed in
after an approximately 10- to 12-h delay. How- various media by other workers for this species
ever, addition of Mn2" at this level during the (6, 19, 26) and for other Arthrobacter species (2,
first 18 h of growth prevented formation of my- 18). Growth, however, occurred only as the my-
celoid cells. Formation of these cells was not celoid form without completing the cycle when
1156 GERMIDA AND CASIDA

A
~ ~.
to~~ 1.J4'-
~~ s,
J. BACTERIOL.

4*.h.
I1 tB

FIG. 3. Thin sections of 72-h cells grown as for Fig. 2. (A) Section showing branching myceloid cell. Bar
marker, 1 pim. (B) Magnified portion of a myceloid cell. Incomplete septation in the presence of septum
membranes shown by arrow. Abbreviations: W, thin cell wall; B, possible bud formation. Bar marker, 0.25
um.
metal-complexing agents, such as Na4P207, edly to myceloid growth already occurring in
EGTA, or citrate were added to the cation-com- this medium. It would appear, therefore, that
plete medium. Growth in this form also occurred myceloid growth was regulated by Mn2" defi-
in a cation-deficient medium which contained ciency per se, or by metal-complexing agents
suboptimal levels of Mn2 . In this case polymor- which tied up Mn2".
phic growth resumed if Mn2+ was added belat- Several named Arthrobacter species and five
VOL. 144, 1980 MYCELOID GROWTH IN ARTHROBACTER SPECIES 1157

i 9). It would appear, however, that our myceloid

( forms are different from these, because myceloid
growth due to Mn2" deficiency was not affected
000'r "Molormm%fts by the addition of biotin, vitamin B12, or minute
concentrations of yeast extract. In addition, the
L. myceloid forms resulting from biotin deficiency
were "membrane-bound bodies within thick
walls" (7), whereas ours had a thin distinct wall
and incomplete septation.
Morris (21) reported myceloid growth of A.
globiformis when strain ATCC 4336 was grown
with citrate as the sole carbon source. He sug-
gested that cation sequestration might be in-
volved, but apparently neither he nor anyone
else experimentally approached this hypothesis.
Interestingly, Stevenson (26) observed myceloid
growth in the complex medium of Henderson
and Snell, which contains citrate. Our studies
showed that citrate caused the myceloid growth
in this medium, and that cation sequestration
was involved in the myceloid response. Thus,
the simultaneous or delayed addition of CaCl2
29 42 % prevented the myceloid growth induced by cit-
FIG. 4. Myceloid growth of A. globiformis strain rate or allowed resumption of normal morpho-
8010 in cation-deficient medium without added man- genesis, respectively. Lucas and Clark (18) ap-
ganese (residual manganese approximately 109 M). parently did not see myceloid growth when they
Numbers on pictures are hour of incubation. Bright- added citrate to the medium of Ensign and
field microscopy; crystal violet stain; 1,142-fold mag- Wolfe (15). Their low level of citrate, however,
nification. was borderline for possible activity.
Myceloid growth seemed to be associated with
Arthrobacter soil isolates exhibited myceloid incomplete wall septation. The electron micro-
growth in cation-complete medium which con- graphs showed this, and myceloid growth pro-
tained metal-complexing agents. Therefore, this ceeded to fragment (yielding coccoid cells)
type of growth was not unique to A. globiformis. within a reasonable period after manganese be-
The fact that A. crystallopoietes failed to dem- came available to the cells. There was, however,
onstrate myceloid growth under any of the con- a difference between the myceloid cells resulting
ditions tested was not totally unexpected. Al- with metal-complexing agents and those for di-
though this bacterium is considered to be a rect manganese deprivation. The myceloid
possible subjective synonym for A. globiformis growth due to metal-complexing agents demon-
(2), metabolic (18) and taxonomic (25) studies strated more branching and less elongation than
have indicated that this organism may not be myceloid forms due to Mn2' deficiency per se.
closely related to A. globifornis and other Ar- These differences, however, may be the conse-
throbacter species. quence of uncontrolled deficiencies of cations
The growth cycle of A. globiformis includes a (metals) other than Mn2 . It is known that the
transitory myceloid phase when this bacterium availability of different metal ions is affected by
is grown in various complex media (3, 14, 24, 26). the choice of metal-complexing agent used (5,
With manganese-deficient media, however, once 16). In addition, however, for the cation-deficient
myceloid growth is under way, it lasts through- medium lacking Mn2 , the nature of the extrac-
out the growth period. The transitory myceloid tion technique and the limited levels of metal
growth in complex media could be due to natu- ions other than Mn2+ that could be added to the
rally occurring chelators, such as the proteins or deficient medium without adding contaminant
amino acids in these media. This would mean manganese resulted in some reduction in con-
that once the complexing agent was metabolized, centrations of the other metals. Thus, the levels
or its complexing ability was affected by pH, the of the other metals seemed to be under better
availability of Mn2" would increase, and normal control and to have fewer deleterious effects on
morphogenesis would resume. Longer-lasting total growth when we used the metal-complex-
myceloid forms that are caused by a deficiency ing agents instead of compounding a medium
of organic growth factors have been reported (6, that was low in manganese. Therefore, we rec-
1158 GERMIDA AND CASIDA J. BACTERIOL.

ommend the use of metal-complexing agents for 15. Ensign, J. C., and R. S. Wolfe. 1964. Nutritional control
achieving myceloid growth in Arthrobacter spe- of morphogenesis in Arthrobacter crystallopoietes. J.
Bacteriol. 87:924-932.
cies. 16. Holloway, J. H., and C. N. Reilly. 1960. Metal chelate
ACKNOWLEDGMENTS stability constants of aminopolycarboxylate ligands.
Anal. Chem. 31:249-256.
This work was supported by grant DAAG29-79-G-0043 17. Lochhead, A. G., and F. E. Chase. 1943. Qualitative
from the U.S. Army Research Office and contract NGR 39- studies of soil microorganisms. V. Nutritional require-
009-180 with the National Aeronautics and Space Administra- ments of the predominant bacterial flora. Can. J. Mi-
tion. crobiol. 3:35-42.
LITERATURE CITED 18. Lucas, D. S., and J. B. Clark. 1975. Induction of mor-
phogenesis in the genusArthrobacter. J. Bacteriol. 124:
1. Balkwill, D. L., and L E. Casida, Jr. 1979. Attachment 1034-1036.
to autoclaved soil of bacterial cells from pure cultures 19. Luscombe, B. M., and T. R. G. Gray. 1974. Character-
of soil isolates. Appl. Environ. Microbiol. 37:1031-1037. istics of Arthrobacter grown in continuous culture. J.
2. Buchanan, R. E., and N. E. Gibbons (ed.). 1974. Ber- Gen. Microbiol. 82:213-222.
gey's manual of determinative bacteriology, 8th ed. 20. Middleton, K. S., and H. B. Gunner. 1968. Porphyrin
Williams & Wilkins Co., Baltimore. production by Arthrobacterglobiformis. Can. J. Micro-
3. Casida, L. E., Jr., and K. C. Liu. 1974. Arthrobacter biol. 14:1095-1103.
globiformis and its bacteriophage in soil. Appl. Micro- 21. Morris, J. G. 1960. Studies on the metabolism of Arthro-
biol. 28:951-959. bacterglobiformis. J. Gen. Microbiol. 22:564-582.
4. Casida, L. E., Jr., and R. Rosenfield. 1957. Bacterial 22. Morrison, N. E., A. D. Antoine, and E. E. Dewbrey.
oxidation of nicotine: formation of gama-aminobutyric 1965. Synthetic metal chelators which replace the nat-
acid. J. Bacteriol. 75:474-479. ural growth-factor requirements of Arthrobacter terre-
5. Chabereh, S., and A. E. Martell. 1959. Organic seques- gens. J. Bacteriol. 89:1630.
tering agents. John Wiley & Sons, Inc., New York. 23. Mulder, E. G., A. D. Adamse, J. Antheunisse, H.
6. Chan, E. C. S. 1964. Morphological aberrations of Ar- Deinema, J. W. Woldendorp, and L. P. T. M. Zev-
throbacter globiformis cells due to biotin deficiency. J. enhuizen. 1966. The relationship between Brevibacter-
Bacteriol. 87:641-651. ium linens and bacteria of the genus Arthrobacter. J.
7. Chan, E. C. S., B. Zyk, and M. Gomersall. 1973. Biotin Appl. Bacteriol. 29:44-71.
deficiency in Arthrobacter globifornis: comparative 24. Sguros, P. L. 1955. Microbial transformations of the
cell ultrastructure and nonreplacement of the vitamin tobacco alkaloids. I. Culture and morphological char-
by structurally unrelated compounds. J. Bacteriol. 113: acteristics of a nicotinophile. J. Bacteriol. 69:28-37.
394-402. 25. Stackebrandt, E., and F. Fiedler. 1979. DNA-DNA
8. Chaplin, C. E. 1957. Life cycles in Arthrobacter pascens homology studies among strains of Arthrobacter and
and Arthrobacter terregens. Can. J. Microbiol. 3:103- Brevibacterium. Arch. Microbiol. 120:289-295.
106. 26. Stevenson, I. L. 1961. Growth studies on Arthrobacter
9. Chaplin, C. E., and A. G. Lochhead. 1956. Abnormal globiformis. Can. J. Microbiol. 7:569-575.
morphology of a bacterium resulting from vitamin B12 27. Stevenson, L. L. 1963. Some observations of the so called
deficiency. Can. J. Microbiol. 2:340-342. "cystites" of the genus Arthrobacter. Can. J. Microbiol.
10. Clark, J. B. 1973. Morphogenesis in the genus Arthro- 9:467-472.
bacter. Crit. Rev. Microbiol. 1:521-544. 28. Stevenson, I. L. 1967. Utilization of aromatic hydrocar-
11. Conn, H. J. 1928. A type of bacteria abundant in produc- bons by Arthrobacter spp. Can. J. Microbiol. 13:205-
tive soils but apparently lacking in certain soils of low 211.
productivity. N. Y. State Agric. Exp. Stn. Tech. Bull. 29. Sundman, V. 1958. Morphological comparison of some
138:3-26. Arthrobacter species. Can. J. Microbiol. 4:221-224.
12. Conn, H. J. 1948. The most abundant group of bacteria 30. Waring, W. S., and C. H. Werkman. 1942. Growth of
in soil. Bacteriol. Rev. 12:257-273. bacteria in an iron free medium. Arch. Biochem. 1:303-
13. Conn, H. J., and I. Dimmick. 1948. Soil bacteria of the 310.
genus Arthrobacter. Soil Sci. 65:349-351. 31. Wills, A. P., and E. C. S. Chan. 1978. Morphogenetic
14. Duxbury, T., and T. R. G. Gray. 1977. A microcultural expression of Arthrobacter globiformis 425 in continu-
study of cystites, cocci and rods of Arthrobacter globi- ous culture with biotin and carbon limitations. Can. J.
fonnis. J. Gen. Microbiol. 103:101-106. Microbiol. 24:28-30.