Survival of two ecologically distinct bacteria (Flavobacterium and Arthrobacter) in unplanted and rhizosphere soil: Laboratory studies

0038-07174a 53.00 + 0.00

Soil Biol. Biochfm. Vol. 2, No. 8, pp. 10?9-1037. 1990 Printed in Great Britain. AlI rights reserved

Copyright C 1990 Pergamon Press pk

SURVIVAL OF TWO ECOLOGICALLY DISTINCT BACTERIA (FLA VOBACTERIUM AND ARTHROBACTER) IN UNPLANTED AND RHIZOSPHERE SOIL: LABORATORY STUDIES I . P . THOMPSON,‘*

K. A. COOK,*

G. LETHBRI~XX* and

R. G. BurtNstt

‘Biological Laboratory, University of Kent, Canterbury, Kent CT2 7NJ. England rShell Research Limited, Sittingboume Research Centre, Sittingboume, Kent ME9 SAG. England (Accepred 30 June 1990)

Summary-The survival of two ecologically distinct bacteria was studied for up to 100 days in laboratory soil microcosms. A triple antibiotic resistant Huvobocterium(P25; km’, sm’, rif’) was chosen as an example of “zymogenous” bacteria, defined as those which grow rapidly when simple nutrient sources are readily available. For comparison, a single antibiotic and methyl red resistant Arrhrobacrer strain (AlO9; sm’) was selected as an example of “autochthonous” bacteria, which are defined as those members of the soil microflora which tend to be present in constant numbers and show low but consistent levels of activity. Al09 survived longer and in greater numbers than did P25 in both non-sterile unplanted (Al09 190 = 41.2 days, 199 = > IOU days; P25 r90 = 4.8 days, 199 = 6.3 days, 199.9 = 8.2 days. BQL = I2 days) and rhizosphere soil (Al09 190 = 8.0 days, r99 = 78.2 days. r99.9 = > 100 days; P25 r90 = 2.8 days. 199 = 5.3 days. 199.9 = 9.0 days, BQL = 60.0 days); where 190, 199, 199.9 represents the real or extrapolated time for the population to decrease to 90%. 99% and 99.9% of the original inoculum density (4.0 x IO6 g-’ soil) and BQL is below quantifiable limits (< I.0 x IO'g-’ soil). Numbers of P25 fell BQL in non-sterile soil I2 days after introduction. Although P25 survived longer in the rhizosphere and rhizoplane than it did in unplanted soil, its numbers still fell to BQL within 60.0 and 41.3 days, rcspcctivcly. Both organisms survived better in heat-sterilized than non-stcrilc soil, suggesting that inoculum death in non-sterile soil was due to competition from and predation by the indigenous community plus a lack of soluble nutrients. The presence of an established inoculum reduced the survival of a second introduced bacterium, most noticeably when P25 was introduced 21 days after A109. Starving the culture or growing it under nutrient limiting conditions prior to introduction to soil at&ted survival of P25 but not Al09.

INTRODUCTION Microorganisms have keen intentionally into soil and rhizosphere environments

introduced

in attempts to enhance certain agriculturally beneficial activities such as nitrogen fixation (Okon. 1983, solubilization of phosphate (Lee and Bagyaraj, 1986), improvement of aggregate stability (Lynch, 1981), suppression of plant pathogens (Maplestone and Campbell, 1989) and promotion of plant growth (Lambert and Joos, 1989). However, one drawback of this is the frequent failure of inoculants to perform in the field, an observation that is difficult to explain because of a poor understanding of the factors which influence survival, proliferation and dispersal of soil inoculants. When laboratory-grown cultures are introduced to soil many bacterial species decline rapidly in numbers (Liang er al., 1982). This may be due to predation, competition or nutrient limitation, and varies according to the species of bacteria introduced, the soil type, and numerous other factors. To date most introduced bacteria have been of the zymogenous or “r” strategist *Present address: Institute of Virology and Environmental Microbiology, Mansfield Road, Oxford OXI 3SR. England. tTo whom all correspondence should be addressed.

type, broadly defined as those organisms that rapidly increase in numbers and activity in response to flushes of readily-available growth and energy substrates (Paul and Clark, 1989). When nutrients become limiting zymogenous bacteria decline in numbers. In contrast, autochthonous or “K” strategist bacteria, typically have low but consistent growth rates and numbers which are relatively constant. The object of the present investigation was to compare the survival of representatives of these ecologically-distinct groups. A Ffavobacterium sp. and an Arthrobacfer sp. were chosen as examples of zymogenous and autochthonous bacteria, respectively, and these were introduced to unplanted soil or to soil planted with wheat. A simple soil microcosm was used and techniques of microbial inoculation, soil and rhizosphere sampling, and bacterial enumeration chosen to facilitate comparative laboratory and field studies. MATERIALS AND ,METHODS

Microorganisms, antibiotic resistance and physiological studies A triple-antibiotic resistant Flovobacrerium strain [p25; kanamycin (km’), rifampicin (rifl), streptomycin (sml)] which most closely resembles F. bafustinum, methyl red resistant and a single-antibiotic,

1029

1030

I.

P. THOMPSONet

Arthrobacter globiformis strain [A 109, streptomycin (sm?] were selected as representatives of zymogenous and autochthonous bacteria, respectively. Both bacteria were isolated from the soil with which all subsequent experiments were performed. P25 was isolated on nutrient agar plus cyclohexamide (50pgml-‘) and Al09 on arthrobacter agar (Hagedom and Holt, 1975). Numbers of each species in the original soil sample were estimated by dilution plate counts and were c 1 x lo’cfu g-’ soil (P25) and 5 x lticfu g-’ soil (A109). Mutants were selected by growing wild type strains in nutrient broth at 25’C. and adding a single antibiotic at the commencement of exponential growth (10 h) (Danso et al., 1973). Cultures were grown for a further 24 h. then transferred to fresh nutrient broth containing the same antibiotic. Growth was indicative of the presence of an antibiotic resistant mutant. The process was repeated for P25 on two further occasions until a spontaneous mutation, resistant to streptomycin (250 pg ml-‘), rifampicin (100 pg ml-‘) and kanamycin (50 /cg ml-‘), was selected. Although Al09 was resistant to only streptomycin (250 pg ml-‘), methyl red (I50 pg ml-‘) was added to A109 medium, since this is selective for arthrobacters (Hagedorn and Holt, 1975). In both P25 and A109. the mutation was chromosomally based (Wheatcroft and Williams, 1981) and stable for over 50 generations. Growth rates in nutrient broth (P25 0.40 h-‘; A109 0. I2 h ‘) nnd 23 biochemical properties (API-20E system API. France) of both mutant and wild type strains were compared and found to be identical. Cultures wcrc mnintaincd on nutrient agar slopes containing the appropriate antibiotics and subcultt’rsd cvcry I4 days. Soils trnd soil LwculaIion

A calcarcous grassland soil (42% sand, 32% silt, 26% clay; pH 6.9; organic C 2. IX) with no previous history of cultivation or agricultural treatment was used for all experiments. Soil was collected from a depth of 15 cm and any large pieces of organic debris and stones removed. The soil was immediately placed into plastic bin liners, returned to the laboratory and stored overnight at 4°C. Sub-samples were dried at 105’C to determine field water content (0.04-0.21 ml g-‘) and moisture holding capacity (MHC; 0.46 ml g-l). Prior to inoculation, the water content of the soil was adjusted with distilled water so that after the addition of the cell suspension the soil was at 70% moisture holding capacity (0.32 ml g-l). For studies with non-sterile soil, 5 g wet wt of inoculated soil (~3.78 g dry wt) was transferred to 5Ocm’ boiling tubes and plugged with nonabsorbent cotton wool. In order to determine inoculant survival in the absence of competition from the indigenous microflora and for inoculant respiration experiments, the soil was sterilized at 12l”C for 30 min on days 1.2.4 and 7 and kept at 20°C between autoclavings. This rigorous treatment was necessary to ensure sterility. Sterilized soil was brought to 70% MHC with sterile distilled water and inoculum solution at the start of the incubation period. P25 and Al09 were grown at 25°C in nutrient broth supplemented with the appropriate anti-

al.

biotic(s) for 36 h (late exponential-early stationary phase). Cultures were centrifuged (10.000 g, 15 min) and washed twice in 0.2 M sodium phosphate buffer @H 7.2) to remove nutrients. One ml of cell suspension (2.0 x IO’ cells ml-‘) was applied to the soil surface, giving an overall cell inoculum level of 4.0 x 106g-’ soil (dry wt), assuming even distribution. Initial densities in the surface layers would have been much higher, however. Soils were not homogenized following inoculation because application was intended to be analogous to the surface inoculation techniques used in subsequent field experiments. Soil microcosms were kept at 25’C in the dark. In competition studies, where both strains were introduced to the same soil microcosms. A109 and P25 were either inoculated at the surface simultaneously at day 0 or the first bacterial strain was added to soil at day 0 and the second strain at day 21. Rhirosphere soil and inoculation

In all rhizosphcre studies, soils in boiling tubes were planted with winter wheat (Trificum aesriuum var. Avalon). Seeds were surface sterilized by shaking in 70% (v/v) ethanol for 2 min. followed by I% (v/v) NaOCl for 2 min, and finally rinsed 3 times in sterile distilled water. Sterilized seeds were placed in oncthird strength nutrient agar and kept at room temperature until the radicals were 2-3 mm in length. A single seedling was placed into an 8 mm deep hole in the soil contained in the boiling tube. Unless otherwise stated. inoculum was injected onto the seed (I ml, 4.0 x lO”cfu) immediately after planting and before covering with soil. Planted soils wcrc placed in a growth room (25 C, 16 h light, 8 h dark) and were maintained at 70% MHC by regular addition of sterile water. Viable counts and biomass

Nutrient agar broth containing the appropriate selective antibiotics and an antifungal agent (cycloheximide 5Opg ml-‘) were used for isolating and cultivating P25 and Al09. Sampling was frequent early in the incubation period in order to detect any rapid increase or decrease in bacterial numbers. On each sampling occasion viable counts were made from triplicate tubes of soil. A IO ml sample of phosphate buffer was added to each tube and the contents agitated for 1 min using a vortex mixer. Serial dilutions of the resulting suspension were placed in Petri dishes to which molten nutrient agar (45°C) containing the appropriate mixture of filter sterilized selective antibiotics was added. Rhizosphere counts were determined by mixing the entire contents of the boiling tube (soil plus root) in phosphate buffer (pH, 7.2, 0.2 M). The washed root was removed and the soil suspension plated. Rhizoplane counts were made by macerating the washed root in a Waring blender for I min prior to dilution and plating as for soil. Rhizosphere numbers are expressed as g-’ soil (dry wt); rhizoplane numbers as g-’ root tissue (dry wt). Direct comparisons of total viable bacteria in planted soil with numbers in unplanted soil were made by adding rhizoplane and rhizosphere counts together and expressing them as bacteria g-’ soil per tube. For comparison purposes survival rates of

Survival of bacteria in soil

inocula were expressed as time in days for a 90% (r90), 99% (199) and 99.9% (199.9) decline in numbers and were determined from the intercepts on the survival curves-thus I constants represent changes in order of magnitude from 4.0 x 106 to 4.0 x lo’, 4.0 x 104 or 4.0 x 10’. In addition, the time taken for the introduced bacteria to decline below quantifiable limits (BQL; < 1.0 x lo’ g-l) was extrapolated from the data. In the biomass studies, IO g (wet wt) of sterilized soil in 30 ml screw top universal vials was inoculated with 4.0 x lo6 g-’ bacteria. On each sampling occasion 5 g soil was diluted and plated to determine viable counts and the remaining 5 g was fumigated for IO days with ethanol-free chloroform, extracted with 2 M KCI and assayed for ninhydrin-reactive nitrogen. A K factor of 0.5 was used (Amato and Ladd, 1988). Respiration

1031

11 r

2

in inoculated soil

Respiration of introduced bacteria was determined using 25 g of sterilized soil contained in modified biometric flasks (Bartha and Pramer, 1965) in which the CO* evolved is absorbed by alkali and determined volumetrically. Biometric flasks were composed of two 250 ml Buchner flasks connected by butyl rubber tubing, 25 g of inoculated sterilized soil was placed in one flask and the other contained 5 ml of 0.25~ NaOH. The production of CO, was monitored over 40 days, initially every 48 h and then after IO days (as respiration rates dcclincd) cu every 72 h.

’ \

t 1

\ \

t

0

1

\ 40

20 Doys

60

60

100

tram introduction

Fig. I. Survival of Arhobacter Al09 (0. 0) or Navobacferium P25 (A.. A) in non-sterile (closed symbols) or sterilized (open symbols) soil. Each point is derived from mean counts of triplicate plates of three soil samples. Bars represent contidence limits of f I SD. Broken lines represent an extrapolation of the data below quantifiable limits. numbers of P25 survived in sterilized soil than in non-sterile soil. The survival of P25 was also greatly improved in sterilized soil. Following an initial increase (9.0 x IO’cfu g-l soil) that was maintained until day 22, there was a gradual decline in numbers although even at day 100 8.3 x IO’cfug-’ soil were detected. &cater

RESULTS

Survival of inoculu in unplanted soil In all unplanted soil experiments Arrhrobacter Al09 survived in greater numbers and for a longer period than did Fluvobacrerium P25. Following introduction to non-sterile soil (Fig. 1. Table I), Al09 numbers increased by >200-fold, reaching a maximum of 8.0 x IO’cfu g-l soil by day 2. Thereafter numbers declined to 2.5 x 1O’cfu g-’ soil by 48 days and remained at that level until the end of the experiment (100 days). In contrast P25 declined from an initial count of 4.0 x 106cfu g-’ soil to 5.2 x IO’ cfu g-’ soil by day 8 and to BQL by day 12. Both inoculants survived far longer and in greater numbers in heat-sterilized soil (Fig. I, Table 1). Numbers of Al09 immediately increased following introduction: 5.7 x IO”cfu g-’ soil were recorded at day 6 and counts remained significantly higher than the number introduced throughout the experiment.

Survival of inoculu in wheat rhizosphere

As was observed in unplanted soil, A109 survived in greater numbers than P25 in all wheat rhizospheres (Fig. 2, Table 1). Following inoculation of wheat seedlings in non-sterile soil, A109 rhizosphere numbers increased to 1.1 x IO’cfu g-’ soil by day I but then gradually declined to 1.3 x IO’cfu g-’ soil by day 100. This was a slightly poorer rate of survival than in unplanted soil, although numbers are not strictly comparable (see above). Even when rhizoplane counts were included (to give total bacteria g-’ soil per boiling tube) the comparable numbers

Table I. Survival of Ar~hroburrcr Al09 and F/m&acrcrium P2S in soil and wheat rhizosphcw Flovobocrrriw f constant

NS

S

r90 199 r99.9

4.8 6.3

61.0 94.6

a.2

Z-100

BQL

12.0

>I00

Arrhrobacrcr

P2S NR

2.8 5.3 9.0 60.0

SR

la.3 51.3 78.0 >I00

NS

S

41.2 >I00 >I00 9100

z-100 >I00 >lOO >I00

A IO9 NR 8.0

SR ,100

78.2 ~100 >I00 Z-100

>I00 >I00

NS. Non-sterile soil unplanted. S. Sterilized soil unplanted. NR. Non-sterile rhizosphcre soil. SR. Sterilized rhizosphcre soil. 190. r99.199.9 represent the time in days for a reduction of 90%, 99% and 99.9%. mpectively in the number of bacteria introduced. BQL is the directly measured or extrapolated time for bactenal numbers to decline below quantifiable limits (i.e. < I.0 x Id g-’ soil).

I. P. THOMPSON et al.

l/ \ 40

20

60

\ \ \ .

60

0 100

Days from Introduction

Fig. 2. Survival of Arfhrohucfrr Al09 (a, 0) or Flakbucterium P25 (A.. A) in non-sterile (closed symbols) or sterilized (open symbols) rhizosphcre soil when wheat seeds were inoculated immediately after planting. Each point is derived from mean counts of triplicate plates of three soil samples. Bars represent confidence limits of f I SD. Broken lines represent an extrapolation of the data Mow quantifiable limits.

of bacteria (1.2 x IO’cfu g-’ soil at day I and 3.5 x lO”cfug-’ soil at day 100) were still less than in unplanted soil. In contrast to Al09, P25 survived in greater numbers in planted soil than in unplanted soil. the decline to BQL being delayed until day 60. Both A109 and P25 survived longer in the wheat rhizosphere in sterilized soil than in non-sterile soil (Fig. 2). A109 numbers showed an initial rapid increase from 4.0 x IO6 to 3.7 x IO9cfu g-’ soil by day 2 followed by a further increase to 1.9 x 1O“‘cfu g-’ soil on day 19, a level that was maintained for the remainder of the experiment. P25 numbers remained constant (ca 1.0 x IO6cfu g-’ soil) until day 10, after which time a gradual decline occurred. However, the population was still just above quantifiable limits by day 100.

, 20

, ‘i,,,,

,

60

80

40

, 100

Days from introduction Fig. 3. Survival of Arthrobacrer A109 (0. 0) or F~o‘lovoboctrrium P25 (A.. A) on the rhizoplanc of non-sterile (closed symbols) or sterilized (open symbols) soil when wheat seeds were inoculated immediately after planting. Each point is derived from mean counts of triplicate plates of three soil samples. Bars represent confidence limits of k I SD. Broken lines represent an extrapolation of the data below quantifiable limits.

numbers. expressed

Furthermore,

as rhizoplane

counts

are

on a g-’ wt of root tissue basis, they are not comparable with rhizosphere counts. In non-sterile soil Al09 rhizoplane counts remained constant at 2.0 x IO’cfu g-’ root tissue from day I until day 19, rising to 6.3 x IO’cfu g-’ root tissue at day 66 and then declining to 1.8 x IO’cfug-’ root tissue at day 100. P25 rhizoplane counts declined from 4.7 x 106cfu g-’ root tissue at day 4 to BQL by day 41. In sterilized soil, in comparison with non-sterile soil, Al09 rhizoplane counts were initially much higher (1.5 x lo* cfu g-’ root tissue) at day 5 and remained in excess of 1.0 x IO’cfu g-’ for the entire experiment. P25 numbers were constant around 6.0 x IO6cfu g-’ root tissue until day 12 and thereafter declined gradually (190 = 22 days, r99 = 30.5 days, t 99.9 = 47 days). Respiration and biomass of inocula: survival us growth

Survival of inocula in wheat rhkoplanes

for rhizoplane bacteria were not determined but could be defined as the numbers of bacteria adhering to the seedling immediately after inoculation. If all introduced bacteria adhered to the seedlings, initial numbers would be in the order of 2.9 x IO9cfu g-’ root tissue (dry wt). However, this was unlikely as significant numbers of introduced bacteria would disperse rapidly into the surrounding soil. Therefore comparisons of changing populations within the rhizoplane (Fig. 3) were made from the first reliable count on either day 1 (non-sterile) or day 2 (sterilized) and not with estimated initial Initial

viable counts

A question of great significance when interpreting microbial survival data is whether the introduced bacteria merely survive to be counted at the various sampling times or that numbers are composed of introduced and new cells. If the second applies do inoculated cells proliferate so that the growth rate is in excess of, equal to, or less than death? When counts are significantly greater than inoculum levels it is reasonable to assume that proliferation has occurred. However, when numbers remain constant or decline it is not possible to state what combination of survival, death and proliferation has contributed to the measured density of cells.

Survival of bacteria in soil

It is anticipated that in sterilized soil the respiration rate of a population of actively dividing cells would exceed that of either a dormant surviving or a dying population. Following introduction of A109 to sterilized soil the respiration rate reached a peak of IOpg CO: g-’ soil h-’ within 48 h and then declined to below BQL (~0.5 CO2 g-’ soil h-‘) by day 15 (Fig. 4). P25 respiration rate in sterilized soil reached a peak (7.6 pg CO2 g-’ soil h-‘) within 48 h, after which it declined. reaching a constant level of 1.4 pg CO2 g-’ soil h-‘. If these rates are compared with the survival curves in sterilized soils (Fig. I), the respiration Rush for both inoculants coincides with the initial increase in numbers. Subsequently, whilst numbers of Al09 were maintained, the respiration rate decreased to BQL suggesting that the population was surviving rather than there being a steady-state between death and proliferation. The more gradual decline in P25 in sterile soil suggests that some proliferation occurred, but this was exceeded by cell death. Further evidence that the gradual decline in determined numbers of P25 were due to a combination of growth and death was obtained by recording survival in soils containing the protein synthesis inhibitor, chloramphcnicol (I40 pg g-l). Under thcsc conditions P25 dcclincd at a more rapid rate than in stcrilc soil without inhibitor (r9O = 18.2 days vs 61.0 days; t99 = 3 I .2 days vs 94.6 days; 199.9 = 44.5 days vs > 100 days; RQL = 65.0 days vs > 100 days). A previous cxpcriment had shown that chloramphenicol had no ctfcct on viable cells. The question of inoculum death and growth was further studied in a scparatc scrics of cxpcrimcnts by relating the survival curves to biomass N (Table 2). Although both spccics wcrc introduced at a higher density than in previous cxpcriments (P25 2.3 x lO”cfu g-’ soil; Al09 3.3 x IO’cfu g-’ soil) survival curves rcfccted similar trends to those shown in Fig. I. On introduction to sterile soil P25 viable numbers increased by 2.5-fold within 24 h, thereafter

Tabk

1033

2. Survival of Flmo&cterium

and Arlhrobacrcr A109

ES

inoculantsand microbial biomass N in sterilized unplanted soil P?5 (cfug-1 soil)

Sample

time (days) 0I 6 20 60 loo

2.3 5.8 x 4.6 x 2.3 x 8.8 x 1.0 x

IO’ 101 IO” lol IO’. IO’.

PZ5 ‘“gW;,;-’

@$

A109

?OY 34 ?25 135 x9 195

3.4 x 3.3 5.5 x 7.9 x 4.8 x 4.9 x

A109 “g-y,;-’

10’. 10. lo’ lo’ IO’ IV

1:. 189 189 194 181

Each value represents the mean of triplicate samples. *Significant difference from the previous value in the column at the 0.05% level.

numbers slowly declined to I.0 x IO’ cfu g-’ soil by day 100. However, throughout the 100 days, P25 biomass also increased from 202 pg N g-’ soil on day 1 to 295 pg N g-’ soil on day 100. Al09 viable counts also increased by IO-fold following introduction to sterile soil and remained at that level for the remainder of the study. Within the first 24 h, Al09 biomass increased ICfold and thereafter remained relatively constant. Eflec~ of plcrnt agr ON inoculu suroiltal in rhizosphere soil

In previous cxpcrimcnts bacteria were introduced directly onto whoat seeds immediately after planting. Application of bacteria to the soil surface after seedling establishment might bc more appropriate in some field situations and so survival was measured following surface inoculation of IO-day-old wheat plants growing in non-stcrilc soil (Fig. 5). A109 survived in lowor numbers when introduced in this way in comparison with inoculation of seed. and although P25 numbers incrcascd 7.5-fold by day 4, by

Br

0

20

40

60

80

100

Days from introduction

0

1

Fig. 5. Survival of Arrhrobacrer A109 (0, 0) or FlauobucA) in non-sterile rhizosphere soil when wheat seeds were inoculated immediately after planting (closed symbols) or bacteria were applied IO days after sowing (open symbols). Each point is derived from mean counts of triplicate plates of three soil samples. Bars represent confidence limits of f I SD. Broken lines represent an extrapol_. . . . .^ . . . ation 01 the data below quantlhable Iimlts. rerium (A..

10 Days from

20

30

40

introduction

Fig. 4. Respiration rates of Arrhrobac~er Al09 (0) or F/ucobacrerium P?S (A) in heat sterilized soil. The bar -_ represents the overall SE of all samples.

I.

1034

P. THOMPWN et

day 50 numbers were the same as in seed inoculation experiments.

al.

10r

Survival of A 109 and P2.5 inoculated simultaneously or sequentially Survival following simultaneous or sequential introduction of Al09 and P2S to sterilized soil was investigated (Fig. 6). When both species were inoculated simultaneously, the sharp increase in numbers recorded when Al09 was introduced on its own did not occur. Nevertheless, there was a gradual increase in A109 numbers and 6.4 x IO’cfu g-i soil were detected at day 100. Numbers of introduced P25 rapidly declined in comparison with survival of P25 as the sole inoculant (r90 = 22.0 days vs 61.0 days: t99 = 24.6 days vs 94.6 days; f99.9 = 27.0 days vs > 100 days; BQL = 3 I .O days vs > 100 days). When P2S was introduced 21 days before Al09, P25 survived for longer and in greater numbers, its decline was more gradual and 1.0 x IO’cfu g-i soil were detected by the end of the experiment (Fig. 7). This rate of decline was only slightly more rapid than when P25 was introduced to sterile soil on its own. When Al09 was introduced 21 days after P25. it did not survive as well as when used as the sole inoculant. showing only a slight increase in numbers by day 31 (i.e. IO days after introduction) and from day 66 Al09 dcclincd in numbers (final count 4.0 x IO’cfu g-l soil). In the rcvcrsc case, when Al09 was introduced 21 days before P25. AI09 numbers showed a similar rcsponsc to that when it was introduced on its own.

0

20

40

Days from

60

60

100

introduction

Fig. 7. Survival of Arfhrobacfer Al09 (0, 0) or F)uoobuclcrium P2S (A, A) in sterilized soil when inoculated together at day 0 (closed symbols) or sequently (open symbols): Al09 at day 0. P25 (A) at day 21 or P25 at day 0, A 109 (0) at day 21. The closed diamond indicates the time at which the second species was added. Each point is derived from mean counts of triplicate plates of three soil samples. Bars represent contidcnce limits of + I SD. Broken lines represent an extrapolation of the data below quantifiable limits.

P25 however declined rapidly to BQL by day 28 (i.e. 7 days after introduction).

0

IO Doys from

lntroductlon

Fig. 6. Survival of Arrhrobocfer Al09 (0, 0) or Fiarobuclerium P2S (A. A) in sterilized soil when inoculated separately (closed symbols) or together (open symbols). Each point is derived from mean counts of triplicate plates of three soil samples. Bars represent confidence limits of f I SD. Broken lines represent an extrapolation of the data below quantifiable limits.

Eflect of physiological state of inoculum on survivul In all routine experiments bacterial cultures harvested at late exponential phase were used and thus it was assumed that inoculants contained a high proportion of rapidly-growing cells. As nutrient levels in soil (or even at root surfaces) are likely to be well below those in broth culture, it was thought that exponentially-growing cells adapted to excess nutrients, might be at a disadvantage when introduced into soil. To investigate this hypothesis, harvested P25 and Al09 were starved in phosphate buffer for 100 days prior to introduction to sterile soil (Fig. 8). Counts of P25 cells declined from 4.0 x 106cfu ml-’ to 1.4 x IO’cfumlduring storage whilst Al09 numbers declined by just one order of magnitude (4.0 x 106cfu to 4.1 x IO’cfu ml-‘). Starved cells were concentrated prior to addition to sterilized soil at the usual rate (i.e. 4.0 x 106cfu g-l). Starved P25 cells initially proliferated to a much higher level (5.5 x IO’cfu g-1 soil by day 5) than did exponentially growing cells introduced immediately from nutrient broth (3.0 x IO’cfu g-t soil by day 5). The rate of subsequent decline was similar but final counts at day 100 were 2.0 x IO6cfu g-’ soil (starved cells) compared with 8.0 x IO’cfu g-’ (growing cells). In contrast to P25, starvation of A109 did not have any significant effect on its subsequent survival in soil.

Survival of bacteria in soil

1035

plating method were consistent and variability was welt within acceptable statistical limits, i.e. f 2 SD of the mean (95% confidence limits). Notwithstanding, the sensitivity of reliable enumeration was limited to densities > 1.0 x lo’ cfu g-’ soil. Below this level cells were sometimes detected but counts were not statistically valid. However, recent investigations (Mason and Bums, 1990) using monoclonal antibodies raised to P25 suggest that accurate quantification may be possible at c20 cells g-’ soil. Sumil~al of inocula in unplanted soil

2

0

I

20

40 Days

from

60

60

100

introduction

Fig. 8. Survival of Arrhohacrer A109 (0, 0) or Flucoburfcrium I’?5 (A, A) in non-sterile soil when inoculated immediately after exponential growth in cirro (closed symbols) or after IO0 days storage in phosphate buffer (open symbols). Each point is derived from mean counts of triplicate plates of lhra soil samplL3. Bars rcprcscnl confidence limits of f I SD. DlSClJSSlON

Use oJ’ spon!uneuus antibiotic dilution plute enumeration

resistunt mu~unts und

The use of bdCWial strains marked with spontancous antibiotic resistance for monitoring the fate of bacteria in soil, proved to be a useful technique allowing accurate quantification of introduced strains. Preliminary investigations showed that microbial growth rates and biochemical characteristics were not changed following mutation. Methods of agitation prior to dilution plating were discussed by Ramsay (1984). We found that using phosphate buffer and vortex mixing was an effective way of dispersing soil microbial cells such that subsequent dilutions were reproducible. Using this method, recovery rates of known numbers of cells extracted from unplanted and planted soil 3 h after introduction were 96% f 3%. However, there was no direct evidence that recovery rates remained high throughout the 100 day studies. Indeed it is likely that any growth of inoculants would be accompanied by surface attachment possibly resulting in a concomitant reduction in ease of dispersal (Burns, 1989). If this was the case, counts made after a period of incubation may have been an underestimate, since a disproportionate number of cells would have sedimented out with soil particles during dilution. However, as neither strain produced significant amounts of extracellular polysaccharide, adhesion of cells to soil surfaces may not be a factor in these experiments. Homogenization of roots prior to rhizoplane counts is likely to displace any adhered cells. Bacterial counts resulting from the extraction and

Survival of Flacobacterium P25 in non-sterile unplanted soil was poor in comparison to that in planted soil, numbers falling BQL within 12 days. In previous studies, on introduction to non-sterile unplanted soil, zymogenous bacteria such as Micrococcus luteus (Casida. 1980) and Xanthomonas campestris (Danso et al.. 1973) also rapidly decreased in numbers. No doubt unplanted soils are nutrient deficient in comparison with rhizosphere soils (Hirsch et al., 1979). although heat sterilization which solubilizes nutrients (Powlson and Jenkinson. 1976) will generally enhance survival, as was the case with P25 (BQL > 100 days). Arrhrobucrer Al09 survived well in non-sterile unplanted soil (190 = 41.2 days, 199 = > 100 days) and the ability of this genus to survive long periods of cnvironmcntal stress (such as starvation) is well known (Gray, 1976). As with P25, A109 survived for longer and in grcatcr numbers in heat-sterilized soil (190 = > 100 days), increasing 200-fold within 2 days of introduction. It is not known if Al09 responded to the same nutrients as P25 although solubilizcd humic components rather than simple C and N arc often substrates utilized by autochthonous bacteria (Stevenson. 1967). In the prcscnce of readily available nutrients rcleascd during heat-sterilization of soil, Al09 displayed rapid growth, but when the new nutrients were cxhaustcd it rcvertcd to its resting form. Using electron microscopy Labcda er al. (1976) observed that when nutrients became limiting, rod shaped Arthrobucter cells assumed a coccoid shape and in our study, direct observation of Al09 when starved in buffer confirmed that rod shaped inocula reverted to coccal forms. Survival o/inoculu

in wheat rhkosphere and rhizoplane

In comparison with unplanted soil, P2S survived in greater numbers in rhitosphere soil and on the rhizoplane (Figs 2 and 3). However, counts still fell BQL 60 and 53 days from introduction, respectively, suggesting that any beneficial rhizosphere effect was ephemeral. This may be because P25, unlike Al09, was not isolated specifically from the rhizosphere but rather from the bulk soil. Therefore P25 may be a poor competitor for root exudates (Wessendorf and Lingens, 1989). In contrast, Al09 survived well in the rhizosphere (BQL = > 100 days) and in unplanted soil. As with sterilized unplanted soil, both inocula survived for longer in sterilized rhizosphere soil and in the rhizoplane (Figs 2 and 3) than in non-sterile equivalents. This is not surprising since survival of inocula in non-sterile habitats is known to be adversely effected by predation and competition (Habte and Alexander, 1977; Casida. 1980).

1. P.

1036

THOMFWN CI al.

Survival of P25 and A 109 in phosphate buffer

It has been suggested that the persistence of introduced bacteria in nutrient limited habitats is closely related to their ability to resist starvation (Nelson and Parkinson, 1978; Sinclair and Alexander, 1984). The ability of P25 and A109 to survive nutrient starvation was tested by storing them in phosphate buffer. A109 survived much better in buffer (t90 =4.0 days; t99= > 100 days; t99.9= > 100 days; BQL = 100 days) than did P25 (t90 = 5.9 days; 199 = 25 days; ~99.9 = 38.1 days; BQL = 91.3 days). Similarly, Nelson and Parkinson (1978) reported greater starvation resistance of Arthrobacter than Pseudomonas and Bacillus in a C-free medium and Boylen and Ensign (1970) showed that A. crystallopoites was especially tolerant to long-term storage in phosphate buffer. Acea and Alexander (1988) observed a direct relationship between starvation resistance and the ability of bacteria to survive in soil and thus a bacterium’s capacity to tolerate starvation conditions in vitro may be a useful indicator of its survival in natural environments. Respiration and biomass of inocula: survival us growth One of the main strategies employed by autochthonous bacteria to survive competition in nutrient deficient habitats is reduction of metabolic rates (Boylen and Mulks. 1978). In our study inoculated Al09 probably rcspondcd to the solubilizcd nutrients rcleascd during heat-stcrihzation of soil and showed an initial increase in viable numbers (Fig. I). rcspiration (Fig. 4) and biomass N (Table 2). When new nutrients became dcplctcd however. the Al09 population assumed a quiescent stage. respiration dcclincd and viable numbers and total biomass became constant. The comparatively poor survival rate of P25 introduced to soil suggests that like other zymogenous bacteria (Klein and Casida, 1967) P25 was not only more susceptible to competition but also that it could not reduce metabolism and retain viability. Certainly lower numbers of P25 survived in phosphate buffer than A109. In sterile soil, viable P25 numbers continually declined over 100 days yet total extractable biomass C increased, indicating that some growth was taking place but that this was counteracted by a more rapid death

rate.

Effect of plant age on P2S and A 109 survival rhizosphere soil

in

Survival of A109 in the wheat rhizosphere was influenced by the age of the plant at inoculation. Al09 survived in greater numbers when introduced simultaneously with wheat seeds (Fig. 2) than when introduced to IO-day-old plants (Fig. 5). This may have been due to changes in the root exudates with time which can influence the composition of the microbial community (Barber and Lynch, 1977) or the inability of A109 to displace members of the established rhizosphere community. In contrast, plant age had little influence on the survival of P25. Survival of A 109 and P25 inoculated simultaneously sequentially

or

The influence of pre-emptive colonization (Compeau et a/., 1988) may explain the results observed in

competition studies between P25 and A109 in sterile soil (Figs 6 and 7). A109 survived in lower numbers in soil containing an established P2S population (introduced 21 days previously) than it did if introduced simultaneously with P25 at day 0 or as the sole inoculum. P25 survived for only 9 days when introduced to sterilized soil containing A109 introduced 21 days previously. This may be due to such features as primary colonizers filling all available physical and nutritional niches or to autogenic changes in the environment preventing the establishment of other species. Eflect of the physiological subsequent survival in soil

state

of the inocula on

Another feature found to have a major influence on inoculum survival was the physiological state of the cell prior to its introduction to soil. In our study P25 starved for 100 days in phosphate buffer survived in greater numbers when introduced into sterile soil than cells introduced immediately after exponential growth (Fig. 8). However, starvation of A109 had little influence upon its subsequent survival in sterile soil. Gauthier et al. (1989) have demonstrated that the survival of Escherichia coli in seawater depended on the age of the cells and on the physiological conditions during their growth. No doubt the survival of bacteria released to soil is dependent upon many factors including climatic parameters, genotypic and phenotypic properties of the strains introduced, their physiological state on release and the timing of inoculation. The results from this study show that simple laboratory microcosms can reveal major difference in the survival of different bacterial species added to soil and endorse recommendations that before any release to the environment the population dynamics of strains should be assessed on a case by case basis (Domsch et al., 1988; Wesscndorf and Lingens, 1989). Our data indicate that an autochonous bacterium such as Arthrobacter A109 competes well with the indigenous community and may establish itself as part of the soil microflora. In contrast, zymogenous strains such as Flavobacterium P25 are unlikely to survive in sufficient numbers to be of any benefit. The value of these microcosm experiments in predicting inoculant behaviour in the field will be discussed in a forthcoming paper. Acknowledgements-We thank Angela Self for technical assistance and Dr Joanna Harris for helpful discussions. REFERENCES

Acea M. J. and Alexander M. (1988) Growth and survival of bacteria introduced info carbon-amended soil. Soil Biology & Biochemtsrry 20. 703-71 I. Amato M. and Ladd J. N. (1988) Assay for microbial biomass based on ninhydrin-reactive nitrogen in extracts of fumigated soil. Soil Biology & Biochemirtry 20. 107-l 14.

Barber D. A. and Lynch J. M. (1977) Microbial growth in the rhizosphere. Soil Biology & Biochemistry 9, 305-308. Bartha R. and Pramer D. (1965) Features of a flask for measuring the persistence ‘and biological effects of pesticides. Soil Science 100, 68-70. Boylen C. W. and Ensign C. J. (1970) Long-term starvation survival of rod and spherical cells of Arthrobacfer crysfullopoites. Journal of Bacteriology

103, 569-577.

Survival of bacteria in soil Boylea C. W. and Mulks M. M. (1978) The survival of coryneform bacteria during periods of prolonged nutrient deprivation. Journal of General Microbiology 105, 323-334.

Burns R. G. (1989) Microbial and enzymic activities in soil biofilms. la Structure and Function of Biojilms (W. G. Characklis and P. A. Wilderer, Eds), pp. 333-349. Wiley, Chichester. Casida L. E. Jr (1980) Death of Micrococcus lure in soil. Applied and Ekironmenral

Microbiology 39, 1031-1034.

Compeau G., Al-Al&i B. J.. Platsouka E. and Levy S. B. (1988) Survival of rifampicin-resistant mutants of Pseudomonas fluorescens and Pseudomonas putida in soil systems. Applied and Environmental Microbiology 4, 2432-2438.

Danso S. K. A.. Habte M. and Alexander M. (1973) Estimating the density of individual bacterial populations introduced into natural ecosystems. Canadian Journal of Microbiology 19, 1450-1451.

Domsch K. H.. Driesel A. J., Goebel W.. Kraus P.. LindenMaier W., Lotz W.. Reber H. and Schmidt F. (1988) Considerations in the release of gene-technologically engineered microorganisms into the environment. FEMS Microbiology Ecology SJ, 261-272.

Gauthier M. J., Munro P. M. and Breittennayer V. A. (1989) Influence of prior growth conditions on low nutrient response of Escherichia coli. Canadian Journal of Microbiology 35. 379-383.

Gray T. R. G. (1976) Survival of vegetative microbes in soil. In The Sunical of Vegerarive Microbes (r. R. G. Gray and J. R. Postgate, Eds). pp. 325-364. Cambridge University Press, Cambridge. Habte M. and Alexander M. (1977) Further evidence for the regulation of bacterial population in soil by protozoa. Archives in Microbiology I 13, I8 I -I 83. Hagcdorn C. and Holt J. G. (1975) Ecology of arthrobacter in Clarion-Webster top sequences of Iowa. Applied .Uicrobiology 9, 2 I I-2 IN. Hirsch P.. Bernard M.. Cohen S. S., Ensign C. J., Jannasch H. W., Koch A. L., Marshall K. C.. Matin A.. Poindexter S. C.. Smith D. C. and Veldkamo H. (1979) Life under conditions of low nutrient concentration: group report. I. In Strategies of Microbial LI$ in Exrrem; E&ir&menrs (bf. Shilo. Ed.). DD. 357-372. Dahlem Konferenzen Life Sciences Research’Report 13..Verlag Chemie. Weinheim. West Germany. Klein D. A. and Casida L. E. Jr (1967) Escherichia coli die-out as related to nutrient availability and the indigenous microtlora. Canadian Journal of Microbiology 13. 1461-1470.

Labeda D. P., Kang-Chicn L. and Casida L. E. Jr (1976) Colonization of soil by Arrhrobacrer and Pseudomonas under varying conditions of water and nutrient availability as studied by plate counts and transmission microscopy. Applied and Environmenral Microbiology 31, 551-561.

1037

Lambert B. and Joos H. (1989) Fundamental rttizobacterial plant growth promotion

aspects of research.

TIBTECH 7, 215-219. Lee A. and Bagyaraj D. J. (1986) Effect of soil inoculation

with vesicular-arbuscular mycorrhizal fungi and either phosphate rock dissolving bacteria or thiobacilh or dry matter production and uptake of phosphorus by tomato plants. New Zealand Journal of Agricultural Research 29, 525-53

I.

Liang L. N., Sinclair J. L., Mallory L. M. and Alexander M. (1982) Fate in model ecosvstems of microbial snecies of potential use in genetic en&eering. Applied andz&vironmental Microbiology 44, 708-714.

Luscombe B. M. and Gray T. R. G. (1974) Characteristics and Arrhrobacrer grown in continuous culture. Journal of General Microbiology 82, 2 13-222.

Lynch J. M. (1981) Promotion and inhibition of soil aggregate stability by microorganisms. Journal of General Microbiology 126. 371-375.

Maplestone P. A. and Campbell R. (1989) Colonization of roots of wheat seedlings by bacilli proposed as biocontrol against take-all. Soil Biology t Biochemistry 21, 543-550. Mason J. and Bums R. G. (1990) Production of a monoclonal antibody specific for a Flavobacterium species isolated from soil. FEMS Microbiology Ecology 73, 299-308.

Nelson L. M. and Parkinson D. (1978) Effect of starvation on three bacterial isolates from an arctic soil. Canadian Journal of Microbiology 24, 1460-1467.

Okon Y. (1985) Azospirillum as a potential inoculant for agriculture Trend7 in Eiorechnology 3, 223-227. Paul E. A. and Clark F. E. (1989) Soil Microbiology and Biochemistry. p. 51. Academic Press, London. -_ Powlson D. S. and Jenkinson D. S. (1976) The effects of biocidal treatments on mctabolism‘in soil. II. Gamma irradiation, autoclaving, air drying and fumigation. Soil Biology & Biochemisrry 8. 179-188.

Ramsay A. J. (1984) Extraction

of bacteria from soil:

efficiency of shaking or ultrasonication as indicated by direct counts and autoradiography. Soil Biology & Eio-

chemistry 16, 475-48 I. Sinclair J. L. and Alexander M. (1984) Role of resistance to starvation in bacterial survival in sewage and lake water. Applied and Environmental Microbiology 24, 164-171.

Stevenson I. L. (1967) Utilization of aromatic hydrocarbons by Arrhrobacrer spp. Canadian Journal of Microbiology 13, 205-2 1 I.

Wessendorf J. and Lingens F. C. (1989) Effect of culture and soil conditions on survival of Pseudomonasfluorestens RI in soil. Applied Microbiology and Biorechnology 31, 97-102.

Wheatcroft R. and Williams R. (1981) Rapid method the study of both stable and unstable plasmids Pseudomonas. 433-437.

Journal

of General

Microbiology

for in

124.