Identification of Genes Required by Bacillus thuringiensisfor Survival in Soil by Transposon-Directed Insertion SiteSequencing

Alistair H. Bishop • Phillip A. Rachwal •

Alka Vaid

Received: 6 October 2013 / Accepted: 14 October 2013 / Published online: 6 December 2013

� Her Majesty the Queen in Rights of the United Kingdom 2013

Abstract Transposon-directed insertion site sequencing

was used to identify genes required by Bacillus thuringi-

ensis to survive in non-axenic plant/soil microcosms. A

total of 516 genetic loci fulfilled the criteria as conferring

survival characteristics. Of these, 127 (24.6 %) were

associated with uptake and transport systems; 227 loci

(44.0 %) coded for enzymatic properties; 49 (9.5 %) were

gene regulation or sensory loci; 40 (7.8 %) were structural

proteins found in the cell envelope or had enzymatic

activities related to it and 24 (4.7 %) were involved in the

production of antibiotics or resistance to them. Eighty-three

(16.1 %) encoded hypothetical proteins or those of

unknown function. The ability to form spores was a key

survival characteristic in the microcosms: bacteria, inocu-

lated in either spore or vegetative form, were able to

multiply and colonise the soil, whereas a sporulation-

deficient mutant was not. The presence of grass seedlings

was critical to colonisation. Bacteria labelled with green

fluorescent protein were observed to adhere to plant roots.

The sporulation-specific promoter of spo0A, the key regu-

lator of sporulation, was strongly activated in the rhizo-

sphere. In contrast, the vegetative-specific promoters of

spo0A and PlcR, a pleiotropic regulator of genes with

diverse activities, were only very weakly activated.

Introduction

Bacillus thuringiensis [41] is a biotechnologically exploi-

ted biopesticide that has far greater usage world-wide than

any other environmental micro-organism [2]. In spite of

this, and in common with other endospore-forming bacte-

ria, relatively little is known about the mechanisms by

which it survives in the environment. The proliferation

during pathogenesis within susceptible invertebrates is well

studied [41] and it has been shown that spores in the soil

are able to colonise the phylloplane from germinating seeds

[4]. There is evidence that B. thuringiensis can exist in the

guts of earthworms [20] which has also been reported to be

a facet of the ecology of Bacillus anthracis in the soil [42].

These two species share such a great genetic similarity that

they should be considered to be variants of the same spe-

cies [15]. As a consequence, a further usage of B. thurin-

giensis is as a surrogate for the threat agent, B. anthracis

[38].

In spite of the very high degree of genetic homology

between these two species [22], one notable difference is

that B. thuringiensis is ubiquitous in the soil, even in

extreme latitudes [35], while B. anthracis only occurs

naturally in highly defined locations [18].There is no

obvious explanation for this, where B. anthracis does

multiply in the soil the presence of plant roots seems to be

important [25, 39]. One proposal has been that B. anthracis

has a non-functional version of the pleiotropic genetic

regulator PlcR [13, 27]. In B. thuringiensis this regulator is

important in invertebrate pathogenesis and controls a wide

range of functions including degradative enzymes, bacte-

riocins, transporters and environmental interactions [13,

19].

Signature-tagged mutagenesis (STM) [16] has fre-

quently been used as a negative screening technique,

Electronic supplementary material The online version of thisarticle (doi:10.1007/s00284-013-0502-7) contains supplementarymaterial, which is available to authorized users.

A. H. Bishop (&) � P. A. Rachwal � A. Vaid

Detection Department, Defence Science and Technology

Laboratory, Porton Down, Salisbury, Wiltshire SP4 0JQ, UK

e-mail: AHBishop@dstl.gov.uk

123

Curr Microbiol (2014) 68:477–485

DOI 10.1007/s00284-013-0502-7

particularly for genes important in pathogenesis [29]. In

contrast, very few similar investigations of microbial

ecology have been carried out [8, 14, 28, 33]. By exploiting

the power of high throughput pyrosequencing it has

become possible to sequence very large libraries of DNA

fragments. Using this approach, a technique termed trans-

poson-directed insertion site sequencing (TraDIS) has been

developed [12, 24, 45]. A saturation library of non-tagged

transposon mutants is made and constitutes the input pool.

Those bacteria remaining after a survival challenge are

termed the ouptut pool. Multiple sequencing reads are

obtained of the loci adjacent to the transposon in both

pools. Those loci present in the input pool but absent in the

output are assumed, if they pass certain statistical criteria,

to have been important in the survival challenge: the

inactivation of that gene by transposon insertion has

resulted in the loss or decrease in number of bacteria car-

rying the insertion at that site. Moreover, a numerical

reflection of the importance of that gene can be derived by

dividing its frequency in the input pool by that in the output

pool. This figure can be termed as ‘Survival Factor’. We

used TraDIS to identify the genes that are required by B.

thuringiensis to survive in a plant/soil microcosm. This

information could be useful in increasing the exploitation

of this organism as a biopesticide. Furthermore, in view of

their close similarity, information on the soil survival

strategies of B. thuringiensis could provide important

insights into the persistence mechanisms of B. anthracis

and might lead to novel decontamination methods.

Materials and Methods

Bacterial Strains, Vectors and Culture Conditions

Bacillus thuringiensis strain 158-S-2 was selected from a

number of isolates recovered from the phylloplane of clo-

ver [3, 5] because of its ability to colonise the rhizosphere

and its amenability to electroporation and DNA extraction.

To evaluate the proportion of bacteria in a population in the

sporulated form, a heat shock of 20 min at 68 �C was

employed to kill vegetative cells. The difference in colony

counts between the number appearing on Tryptone Soy

Agar (TSA) (Oxoid), with and without heat shock, repre-

sented the proportion of heat-sensitive, vegetative cells.

The minimal medium used was BM [32] with 0.1 % glu-

cose (w/v) as the carbon source. Routine liquid culturing

used Tryptone Soya Broth (TSB) (Oxoid). Escherichia coli

was grown using Luria–Bertani (LB) [40] agar and broth.

The plasposon pAW068, developed for use with B.

anthracis, was a gift from (Prof. Marta Perego, Scripps

Institute, La Jolla, CA, USA). It was electroporated into B.

thuringiensis 158-S-2 and transformants selected for

growth at 28 �C in the presence of chloramphenicol

(7.5 lg ml-1) and spectinomycin (100 lg ml-1). Trans-

formed cells were cultured overnight at the non-permissive

temperature for plasmid replication (35 �C) [48]. Cells

were selected with a phenotype of chloramphenicol sensi-

tivity but spectinomycin resistance, indicating integration

of the transposon with loss of the plasmid. Inverse PCR

[48] and sequencing of eight clones indicated that random

integration had occurred (data not shown). This population

of bacteria were expected, collectively, to represent a sat-

uration library of transposon insertion throughout the

genome and was termed the ‘input pool’. Prior to inocu-

lation into the microcosms, the input pool was grown

statically at room temperature in a mixture of BM:TSB

media (75:25) overnight. The early vegetative phase bac-

teria were quantified spectrophotometrically, harvested by

centrifugation at 6,0009g and washed in PBS.

The GFP/lux reporter plasmid (pDestpUNK) [34] was a

gift from Prof. Phil Hill (University of Nottingham, UK).

TOPO-TA kit for cloning PCR products and TOP10 E. coli

competent cells were purchased from Promega. The

chloramphenicol resistance-encoding plasmid pBBRMCS1

[43] and the suicide plasmid pRN5101 [46] were provided

by Dr Colin Berry (University of Cardiff, UK).

Construction of spo- strain of B. thuringiensis

PCR primers (Supplementary Information, Table 1) were

based upon the published sequence of the spo0A gene in B.

thuringiensis [26]. Amplification (94 �C, 2 min; 35 cycles

of: 94 �C for 30 s, 52 �C for 30 s, 72 �C for 30 s; 72 �C for

7 min) resulted in a 1.7 Kb product from strain 158-S-2

which was cloned into the TOPO-TA vector (Invitrogen).

The chloramphenicol-resistance gene from pBBRMCS1

was amplified using the thermal cycling profile, above, and

the primers listed in Supplementary Information, Table 1,

which incorporate AccI restriction enzyme sites at the 50

ends. After digestion with AccI the chloramphenicol gene

was ligated into the spo0A gene in TOPO-TA plasmid. The

resulting constructs were transformed into competent cells

of E. coli (TOP 10, Invitrogen) and selected on LB agar

containing chloramphenicol (10 lg ml-1). The resulting

clones were screened on the basis of producing amplifi-

cation with the spo0A primers commensurate with the

insertion of the chloramphenicol gene. The disrupted gene

was amplified with M13 primers (Supplementary Infor-

mation, Table 1), blunt ended and ligated into the EcoRV

cloning site of the temperature-sensitive suicide plasmid

pRN5101. This was electroporated into B. thuringiensis

strain 158-S-2 and transformed colonies selected on LB

agar containing chloramphenicol (10 lg ml-1). Electro-

transformants were grown overnight in TSB containing

chloramphenicol (10 lg ml-1) at the non-permissive

478 A. H. Bishop et al.: Soil Survival Genes of Bacillus thuringiensis

123

replication temperature for pRN5101 (37 �C). Surviving

cells could only maintain the required chloramphenicol

resistance gene by homologous recombination between the

functional, chromosomal copy of the spo0A gene and the

disrupted, plasmid-borne fragments. Colonies were

screened for the correct phenotype and PCR amplification

products. These were then verified by DNA sequencing.

Construction of Promoter Reporter Plasmids

The sporulation-specific (‘spospo’) and vegetative-specific

(‘spoveg’) promoters of the B. thuringiensis spo0A gene

[26] were amplified from B. thuringiensis 158-S-2 using

the primers listed in Supplementary Information, Table 1

and the amplification conditions above. Synthetic com-

plementary oligonucleotides (Supplementary Information,

Table 1) of a consensus sequence from 28 PlcR activation

boxes [13] were synthesised (Sigma Genosys). After

annealing, the double-stranded DNA was amplified using

the Gateway cloning plcR primers [26] (Supplementary

Information, Table 1). All amplifications used Platinum

Pfx polymerase (Invitrogen) and the amplification profile

above. All three promoters were inserted into the pDest-

PUNK reporter plasmid using the Gateway cloning kit

(Invitrogen).

Electroporation

Electroporation was carried out as previously described in

[6], except that 4-mm gap cuvettes (MicroBiosystems)

were used and that bacteria were incubated after electro-

poration in brain heart infusion (BHI) broth (Oxoid) con-

taining 0.5 % glycerol (Sigma) for 90-min at 30 �C, with

shaking, before plating onto BHI agar containing the

appropriate selective antibiotic.

Microcosms

The soil was obtained from the site where the bacteria were

isolated and was a chalky loam whose composition was

previously reported [3]. The following plants (Emorsgate

Seeds) used were: Avenula pratense (meadow oat grass),

Dactylis glomerata (cocksfoot), Festuca pratensis (mea-

dow fescue), Festuca rubra (red fescue) Lolium perenne

(ryegrass), Schedonorus arundinacea (tall fescue), Phleum

pratense (Timothy) and Trifolium pratensis (red clover).

Seeds were germinated and grown in non-sterile tap water

to a root length of about 2 cm. Ten seedlings, two of each

species, were transplanted into each of 24 wells (3.5-cm

diameter) in multiwell plates (Nunc) and soil (*8 g)

added. The soil was then inoculated with a bacterial sus-

pension. Distilled water was added to moisten the soil and

the plates were then incubated at 20 �C, 12:12, light:dark

photoperiod and 80 % humidity in an environmental cab-

inet (Weiss-Gallenkamp).

Recovery of Bacteria From the Microcosms

After 2 weeks, the bacteria were recovered from the

microcosms by shaking the contents of each well in soil

extraction buffer (0.5 % (v/v) Tween 20 and 0.1 % (w/v)

sodium pyrophosphate) to a total volume of 45 ml for

20 min at setting 8 on a random motion oscillator (Weiss-

Gallenkamp). To enumerate the cell density serial dilutions

were made in sterile PBS and plated in duplicate onto TSA

containing spectinomycin (100 lg ml-1) with penicillin

(80 lg ml-1) and polymyxin (20 lg ml-1). Exploiting the

inherent resistance of B. thuringiensis to the latter two

antibiotics helped suppress the contaminants that were

otherwise able to grow through the spectinomycin selection.

To recover the bacteria, slurry (8 ml) obtained from

shaking the soil with extraction buffer was loaded onto

5 ml of Histodenz (Sigma), made to a density of 1.3 g l-1,

in 15 ml tubes (Falcon). These were centrifuged at

5,0009g for 1.5 h. The layer containing the bacteria was

aspirated with a sterile Pasteur pipette. The population of

B. thuringiensis recovered was termed the ‘output pool’.

Preparation of Sequencing Libraries

Genomic DNA was initially extracted from the input and

output pools of both B. thuringiensis 158-S-2 using the

Pure Gene kit (Qiagen). Each DNA pool (5 lg) was frag-

mented by nebulisation (30 psi nitrogen for 6 min using an

Invitrogen nebuliser) and the resulting fragments size-

assessed on an Agilent Bioanalyser High Sensitivity chip.

Fragmented DNA was resolved on a 2 % (w/v) agarose gel

and DNA in the size range 250–400 bp excised and puri-

fied using a QiaExII gel extraction kit (Qiagen). The

fragment libraries were end-polished and A-tailed using a

NEBNext DNA library preparation kit for the Illumina

sequencing (New England Biolabs), ready for adapter

ligation. Double stranded adapters Ind_Ad_T and

Ind_Ad_B (Supplementary Information, Table 1) were

annealed and ligated to the fragment libraries. These were

quantified by qPCR using the primers Ad_T_qPCR1 and

Ad_B_qPCR2 (Supplementary Information, Table 1) using

an Illumina library quantification kit (KAPA Biosystems),

according to the manufacturer’s instructions.

Adapter-ligated fragments were then enriched for those

specifically containing a transposon insertion site using

PCR primers with homology to the 30 inverted repeat

sequence of pAW068. The PCR thermal cycle was as

follows: 90 �C for 30-s, followed by 24 cycles of 98 �C for

10-s; 65 �C for 30-s and 72 �C for 30-s. The first four

cycles contained only the transposon-specific primer to

A. H. Bishop et al.: Soil Survival Genes of Bacillus thuringiensis 479

123

maximise enrichment of the transposon-associated frag-

ments. The transposon specific primers also contained the

Illumina P5 end for attachment to the Illumina flow cell:

pAW068-p5 (Supplementary Information, Table 1). The

reverse primer, RInV3.3, (Supplementary Information,

Table 1) which contained the Illumina P7 end, was added

at the start of the fifth cycle. To limit PCR bias, multiple

PCR reactions were run in parallel and pooled after size

selection to generate sufficient concentrations of library

material. PCR products were resolved on a 2 % (w/v)

agarose gel. Successfully amplified products were purified

and size-fractionated by gel electrophoresis and the DNA

recovered using a QiaExII gel extraction kit (Qiagen).

DNA was eluted in 20 ll elution buffer and quantified by

quantitative PCR (KAPA Biosystems). Sequencing was

performed by the University of Exeter Sequencing Service

on an Illumina HiSeq 2500 instrument.

Analysis of Survival Genes

The sequence data were analysed by the Integrative

Genomics Viewer software [37]. Statistical analysis [10] has

shown that a minimum of 32 sequence reads is required in

the input pool for a given gene locus to be confident that any

decrease in frequency in the output pool is significant.

Furthermore, the same study demonstrated that a twofold

decrease in the frequency of a gene locus in the ouput pool

compared to the input is the threshold level that a transposon

insertion has resulted in a loss of fitness. These criteria were

adopted for the data presented here. A survival factor was

obtained for each gene locus meeting these requirements by

dividing the number of reads for the output pool for that of

the input pool. The European Molecular Biology Laboratory

(EMBL) database was used to compare the DNA sequences

to the reference sequence of B. thuringiensis BMB171 [17].

Attachment of Fluorescent Bacteria to Roots

Microcosms were inoculated as above but using B. thur-

ingiensis containing the pDestpUNK expression vector

under the control of either the plcR, spoveg or spospo

promoter. After 1 week, the seedlings were carefully

pulled from the soil and washed in PBS. The roots were

excised and mounted on a slide for examination by con-

focal microscopy (Zeiss Observer Z1).

Results and Discussion

Preparation of B. thuringiensis Inoculum

The initial intention was to use a sporulation-deficient

(spo0A-) mutant to ensure that the bacteria had to use

mechanisms to survive that were available in the vegetative

state rather than immediately sporulating. The survival of

this mutant in the microcosms was very poor compared to

the wild-type (Fig. 1) which proliferated. The wild-type

strain was used for all subsequent experiments. Prior to

inoculation of the saturation transposon library into the

microcosms it was cultured through several cycles of

sporulation and germination to remove any mutants that

were deficient in sporulation, as this was now known to be

an important survival factor. Furthermore, the saturation

library was cultured in minimal medium to remove muta-

tions in central metabolic pathways which might not be

lethal in a rich medium, but which would not be directly

related to soil survival. Bacteria grown on minimal med-

ium directly before inoculation survived poorly in the

microcosms (data not shown). For this reason the input

pool was cultured in a mixture (75:25) of minimal:rich

media (see ‘Materials and Methods’ section). The mini-

mum inoculum density that would result in colonisation of

the microcosms was 4 9 105 CFU g-1 soil-1 (data not

shown). Higher inocula did not produce an appreciably

greater final cell density than that shown in Fig. 1. The

minimum inoculum level possible was chosen to avoid so-

called stochastic errors or selection bottlenecks [10, 29]:

mutants that might have been capable of persisting through

the challenge are lost simply because too many bacteria

were initially inoculated. Equally, attenuated bacteria sur-

vive because they do not experience a sufficiently stringent

survival challenge. This results in random loss or survival,

respectively, in the output population. Sufficient bacteria

were harvested in the output pool to provide greater than a

100-fold increase beyond the input pool to expect, at the

95 % confidence interval, that a particular mutant has been

lost due to failure to survive as opposed to chance [9].

Fig. 1 Colonisation of soil/plant microcosm by B. thuringiensis

158-S-2, wild-type (light grey lines) and spo- mutant (dark grey

lines). Plate counts were obtained from six replicate wells from three

separate experiments. The error bars represent standard deviation

480 A. H. Bishop et al.: Soil Survival Genes of Bacillus thuringiensis

123

Summary of Genes Required for Survival

in the Microcosms

Bacillus thuringiensis strain BMB171 [17] has a genome of

5.64 Mb and contains 5,088 open reading frames. It was

used to annotate the TraDIS sequences from the environ-

mental strain of B. thuringiensis used. A total of 516

genetic loci fulfilled the criteria for conferring survival

characteristics. Of these, 127 (24.6 %) were associated

with uptake and transport systems (Supplementary Infor-

mation, Table 2A); 227 loci (44.0 %) coded for enzymatic

properties; 49 (9.5 %) were gene regulation or sensory loci

(Supplementary Information, Table 2B); 40 (7.8 %) were

structural proteins found in the cell envelope or had

enzymatic activities related to it (Supplementary Informa-

tion, Table 2C); and 24 (4.7 %) were involved in the

production of antibiotics or resistance proteins (Supple-

mentary Information, Table 2D). 83 (16.1 %) encoded

hypothetical proteins or those of unknown function. The

full list of genetic loci identified as being involved in soil

survival are listed in Table 2, Supplementary Information.

Taken together, these findings demonstrate that the

microcosm challenge was a very stringent test of envi-

ronmental survival. This can be seen from the dependence:

(i) on nutrient uptake and transport systems (ii) regulatory

mechanisms to optimize gene expression (iii) the produc-

tion of and protection against antibiotics and other harmful

substances and (iv) the production of enzymes to hydolyze

nutrient sources and metabolise those carbon sources found

in the soil and rhizosphere.

Cell Envelope

Of the 40 cell envelope proteins required for soil survival

over half were involved in peptidoglycan synthesis (Sup-

plementary Information, Table 2C). Given that the bacteria

had been grown in minimal medium prior to inoculation

into the microcosms it might seem unexpected that any

mutants in peptidoglycan synthesis remained. It is well

known, however, that bacteria can change their cell enve-

lope in response to the external environment and their

physiological state. In a classic chemostat study [11] it was

shown that Bacillus megaterium actively replaced acces-

sory polymers to optimize nutrient availability. The

thickness and degree of cross-linking of peptidoglycan is

also variable and it may be that a stronger wall is required

in the soil than in laboratory culture media. Autolysins may

also play a part in biofilm production and cell–cell inter-

actions [21]. Proteins active in the cell envelope during cell

division were also highlighted, indicating that this process

may be more exacting in the environment than in vitro.

Whatever the reasons, many examples exist of transposon

mutagenesis uncovering peptidoglycan synthesis genes

being required during pathogenesis and environmental

persistence [14, 28, 30, 31]. An inability to sporulate is also

disadvantageous, as illustrated in Fig. 1. In corroboration

of this, the survival of mutants in the sporulation inhibitor

gene, soj (BMB171_C5080) [36], was also decreased

(Supplementary information, Table 3).

Transport and Competition Mechanisms

Further evidence of the dynamic nature of the existence of B.

thuringiensis in the microcosms is given by the fact that

nearly 25 % of the persistence genes identified were

involved in transport mechanisms (Supplementary Infor-

mation, Table 2A). The necessity for the bacteria to import

metal ions and small organic compounds, for example, was

evidently crucial for their ability to colonise and persist. In

addition to the genes involved in antibiotic production or

resistance (Supplementary Information, Table 2D) at least

13 of the transport mechanisms identified in Supplementary

Information, Table 2A, could function as antibiotic efflux

pumps. Together, this represents 37 genes, 7.2 % of the total,

which participated in antibiotic production or resistance. The

concept of Bacillus species residing in the soil predomi-

nantly in spore form would appear to be unjustified: spores

do not participate in transport mechanisms, nor do they

produce or become susceptible to antibiotics. Clearly, B.

thuringiensis was actively involved in extensive competition

in order to establish niches in these microcosms. In addition

to antibiotics, both B. thuringiensis and B. anthracis engage

in a more subtle means of undermining some competitors.

Acyl homoserine lactones (AHLs) are quorum-sensing

pheromones that some Gram-negative bacteria produce to

communicate with each other [44]. They are used during

pathogenesis and in biofilm formation, for example, and also

to initiate antibiotic synthesis. Both Bacillus species are able

to produce lactonases which degrade these messengers [44].

It was shown here that disruption of the lactonase gene

(BMB171_C2975; Supplementary Information, Table 3)

decreased microcosm fitness.

Gene Regulation

Nearly 10 % of the loci required for persistence in the

microcosms were gene regulators or sensory mechanisms

(Supplementary Information, Table 2B). Their appearance

is further evidence that the inoculated bacteria were meta-

bolically active and that a wide range of functions was nec-

essary. The regulators with the four highest survival factors

are: a two-component, intracellular signal transduction reg-

ulator (BMB171_C2783); the main house-keeping RNA

polymerase signal factor (BMB171_C3957); the regulator of

an enzyme with a gluconeogenic function (BMB171_

C3961) and a regulator of autolysis (BMB171_C5036). This

A. H. Bishop et al.: Soil Survival Genes of Bacillus thuringiensis 481

123

demonstrates the diversity of functions found necessary for

persistence in the microcosms. It is also of interest that

mutants in five methyl-accepting chemotaxis genes

(BMB171_C0363, BMB171_C1792, BMB171_C3193,

BMB171_C4620, BMB171_C0497) exhibited decreased

survival in the microcosms (Table 2B). Strain 158-S-2

which was used in these experiments is amotile, at least

in vitro. The chemotaxis genes must, however, be required

for some, as yet, unknown function. B. anthracis is also

amotile but contains chemotaxis genes.

Energy Generation and Carbon Metabolism

Key indicators that nutrient supply was not plentiful are the

attenuation caused by mechanisms orchestrating the

response to limited carbon levels such as carbon starvation

protein A (BMB171_C5010) and the regulator of carbon

metabolism, CcpA (BMB171_C4314) (Supplementary

Information, Table 2B). Isocitrate lyase (BMB171_C0994)

and malate synthase (BMB171_C0993) which are enzymes

of the glyoxylate, anaplerotic pathway were also survival

factors (Supplementary Information, Table 3). In addition,

storage compounds were important for persistence, e.g.

glycogen synthase (BMB171_C4503), 3–hydroxybutyrate

dehydrogenase (BMB171_C3694) and 3–hydroxybutyryl

CoA dehydrogenase (BMB171_C4939) (Supplementary

Information, Table 3).

The archetypal member of its genus, Bacillus subtilis, and

its closest relatives, are specialist soil-dwelling organisms

and have a large part of their metabolism devoted to carbo-

hydrate utilization. In contrast, B. thuringiensis and B. an-

thracis have a limited number of sugars and polysaccharides

that they can metabolize [19]. Instead, they both have a

greater number of peptide and amino acid transporters and

amino acid degradation pathways than B. subtilis. This has

been taken as evidence that a pathogenic life-style is much

more common for both B. thuringiensis and B. anthracis

[19]. As evidence in support of this the results here show few

pathways involved in the degradation of extracellular poly-

saccharides (e.g. glycoside hydrolase, BMB171_C2152 and

pullulanase, BMB171_C2448) but a large number of prote-

ases (Supplementary Information, Table 3) and transport

systems for amino acids (Supplementary Information,

Table 2A) are required for environmental survival. Collagen

adhesion protein is conserved between members of Bacillus

cereus sensu lato. Attachment to collagen during pathogen-

esis has been shown to be a feature of some strains of B.

cereus sensu stricto [23]. This may be advantageous during

interactions with invertebrates by both B. thuringiensis and

B. anthracis [20]. In the context of environmental survival

this wall-anchored adhesin might be used to attach to

invertebrates or to another macromolecule altogether. Two

genes encoding collagen adhesion proteins were identified as

survival factors (Supplementary Information, Table 3:

BMB171_C2270 and BMB171_C0765). This might seem co-

incidental on its own, but two genes with high survival factors

(BMB171_C3206 and BMB171_C3432) had collagenase

activity. Conceivably, the degradation of collagen or related

proteins is an important survival mechanism in these bacteria

which rely heavily on amino acid metabolism. Inactivation of

immune inhibitor A, a metalloprotease capable of cleaving

insect antibacterial peptides (BMB171_C0585), also resulted

in attenuated persistence (Supplementary Information,

Table 3). An ortholog of this gene occurs in B. anthracis and

its value in environmental survival, like collagenase, may be

as a non-target specific protease.

Colonisation of the Rhizosphere

A close association with plant roots has been implicated in

the presence of B. anthracis in a vegetative state in soil

[39]. This was further investigated using B. thuringiensis

expressing green fluorescent protein (GFP) from the

pDestpUNK expression vector. The sporulation- and veg-

etative-specific, ‘spospo’ and ‘spoveg’ promoters, respec-

tively, of spo0A and a concensus PlcR recognition box

were used as separate promoters for GFP (Supplementary

Information, Table 1). The bacteria were inoculated into

microcosms and incubated for 1 week. The seedlings were

Fig. 2 Confocal micrograph of B. thuringiensis cells (green) con-

taining the pDestpUNK expression vector incorporating the spo0A

sporulation-specific (‘spospo’) promoter. The bacteria were allowed

to colonise soil microcosms containing grass seedlings for 1 week

prior to extraction and washing. They were observed to be firmly

attached to the roots (red arrow) and root hairs (white arrow) (Color

figure online)

482 A. H. Bishop et al.: Soil Survival Genes of Bacillus thuringiensis

123

pulled out and the roots gently washed in phosphate-buf-

fered saline (PBS) to remove any residual soil. They were

then examined under confocal microscopy (Fig. 2). Fluo-

rescent bacteria, containing the spospo promoter were

clearly visible and were firmly attached to the roots and

root hairs. Being non-motile, like B. anthracis, attachment

to the root might be a vital persistence mechanism for B.

thuringiensis because the root is a source of organic acids,

sugars, polysaccharides, proteins and amino acids [1].

Nevertheless, the bacteria appeared to be in the process of

sporulating, having activated the sporulation-specific pro-

moter. This implies that a flux between spore and vegeta-

tive forms may be a feature of soil persistence: the

proportion of the transposon-tagged bacteria in the micro-

cosms present in spore form ranged between 30 and 70 %

over a 2-week period, varying unpredictably between the

times of sampling and within samples taken at the same

time. Even for strains of the B. cereus sensu lato group that

possess flagella, translocation through the soil seems to be

effected by a multicellular mode of growth involving cell

elongation rather than flagellar activity [47]. All three

promoter constructs resulted in strongly fluorescent bacte-

ria in laboratory media (although the signal from the spo-

spo promoter was noticeably stronger). Even under

prolonged exposure times the cells carrying the spoveg and

PlcR box promoters were only faintly visible in the

microcosms, implying that they were not strongly activated

under these conditions.

One difference between B. thuringiensis and B. anthracis

is that the arginine deiminase operon (Fig. 3) is deleted in

the latter [19]. Only one gene of the operon emerged as

being required for soil persistence for B. thuringiensis, the

arginine/ornithine transporter (BMB171_C0350). The other

members of the operon had insufficient reads to render a

statistically significant difference between input and output

pools. Incomplete identification of all members of an

operon has been previously reported [10]. It is interesting

that this pathway allows Streptococcus pyogenes to survive

acidic conditions in culture media through the release of

ammonia [7]. It has been postulated that this operon has

been deleted in B. anthracis because ammonia interferes

with its mechanism of pathogenesis [19]. While there is, as

yet, no proof that the lack of this pathway confines the

vegetative multiplication of B. anthracis to alkaline soils it

may be a contributing factor.

A very large amount of data have been generated which,

hopefully, will allow aspects of the soil survival of related

bacteria to be examined. To the best of our knowledge this

is the first report of the application of the TraDIS technique

to an environmental study. The use of non-axenic micro-

cosms was crucial for replicating the intense competition

that bacteria are imagined to experience in the environ-

ment. The requirements to scavenge nutrients and to

engage in antimicrobial interactions are examples of this.

Cell wall metabolism and sporulation appear to be as much

dynamic responses to the environment as is gene regula-

tion. It is evident, therefore, that even spore-forming

organisms have active mechanisms for establishing them-

selves in the soil and colonising the rhizosphere.

Acknowledgments The authors are very grateful to The Defense

Threat Reduction Agency (DTRA) for funding. AHB was also sup-

ported by funding from the Ministry of Defence, UK. We also

acknowledge the help from Gill Hartley (Dstl) for confocal micros-

copy and Bry Lingard (Dstl) for assistance with sequence analysis.

We thank Dr Sari Paikoff (DTRA) and Prof. Petra Oyston for support

and advice.

References

1. Bais HP, Weir TF, Perry LG, Gilroy S, Vivanco JM (2006) The

role of root exudates in rhizosphere interactions with plants and

other organisms. Ann Rev Plant Biol 57:233–266

2. Bishop AH (2002) The insecticidal proteins of Bacillus thurin-

giensis. In: Hendrickx M (ed) Applications and systematics of the

genus bacillus and related organisms. Federation of European

Microbiological Societies. Blackwell, Oxford, pp 160–175

3. Bizzarri MF, Bishop AH (2007) Recovery of Bacillus thuringi-

ensis in vegetative form from the phylloplane of clover (Trifolium

hybridum) during a growing season. J Inv Pathol 94:38–47

4. Bizzarri MF, Bishop AH (2008) The ecology of Bacillus thur-

ingiensis on the phylloplane: colonisation from soil, conjugation

B. anthracis

Methyl accepting Exo-alpha- chemotaxis protein 1,4-glucosidase

B. thuringiensis

Arginine repressor Ornithine Carbamate ArgR carbamoyl kinase

transferase

Arginine Arginine/ ornithinedeiminase antiporter

Fig. 3 The arginine deaminase

operon of B. thuringiensis

which is deleted in B. anthracis

[19]. The shared genes flanking

the deletion (methyl accepting

chemotaxis protein and exo-

alpha-1,4-glucosidase) are

shown between the two

chromosomal segments

A. H. Bishop et al.: Soil Survival Genes of Bacillus thuringiensis 483

123

and interaction with larvae of Pieris brassicae. Microbiol Ecol

56:133–139

5. Bizzarri MF, Prabhakar A, Bishop AH (2008) Multiple-locus

sequence typing analysis of Bacillus thuringiensis recovered from

the phylloplane of clover (Trifolium hybridum) in vegetative

form. Microbiol Ecol 55:619–625

6. Bone EJ, Ellar DJ (1989) Transformation of Bacillus thuringi-

ensis by electroporation. FEMS Microbiol Lett 58:171–178

7. Degnan BA, Fontaine MC, Doebereiner AH, Lee JJ, Mastroeni P,

Dougan G, Goodacre G, Kehoe MA (2000) Characterization of

an isogenic mutant of Streptococcus pyogenes Manfredo lacking

the ability to make Streptococcal acid glycoprotein. Infect Immun

68:241–248

8. Deutschbauer A, Price MN, Wetmore KM, Shao W, Baumohl JK,

Xu Z, Nguyen M, Tamse R, Davis RW, Arkin AP (2011) Evi-

dence-based annotation of gene function in Shewanella oneid-

ensis MR-1 using genome-wide fitness profiling across 121

conditions. PLoS Genet 11:e1002385

9. Dziva F, van Diemen PM, Stevens MP, Smith AJ, Wallis TM

(2004) Identification of Escherichia coli O157: H7 genes influ-

encing colonization of the bovine gastrointestinal tract using

signature-tagged mutagenesis. Microbiology 150:3631–3645

10. Eckert SE, Dziva F, Chaudhuri RR, Langridge GC, Turner DJ,

Pickard DJ, Maskell DJ, Thompson NR, Stevens MP (2011)

Retrospective application of transposon-directed insertion site

sequencing to a library of signature-tagged mini-Tn5Km2

mutants of Escherichia coli O157:H7 screened in cattle. J Bac-

teriol 193:1771–1776

11. Ellwood DC, Tempest DA (1969) Control of teichoic acid and

teichuronic acid biosynthesis in chemostat cultures of Bacillus

subtilis var. niger. Biochem J 111:1–5

12. Gawronski JD, Wonga SMS, Giannoukos G, Ward DV, Akerley

BJ (2009) Tracking insertion mutants within libraries by deep

sequencing and a genome-wide screen for Haemophilus genes

required in the lung. Proc Natl Acad Sci USA 106:16422–16427

13. Gohar M, Faegri K, Perchat S, Ravnum S, Økstad OE, Gominet

M, Gilois N, Kolstø AB, Lereclus D (2008) The PlcR virulence

regulon of Bacillus cereus. PLoS ONE 3:e2793

14. Groh JL, Luo Q, Ballard JD, Krumholz LR (2007) Genes that

enhance the ecological fitness of Shewanella oneidensis MR-1 in

sediments reveal the value of antibiotic resistance. Appl Environ

Microbiol 73:492–498

15. Helgason E, Økstad OA, Caugant DA, Johansen HA, Fouet A,

Mock M, Hegna I, Kolstø A-B (2000) Bacillus anthracis,

Bacillus cereus, and Bacillus thuringiensis—one species on the

basis of genetic evidence. Appl Environ Microbiol 66:2627–2630

16. Hensel M, Shea JE, Gleeson C, Jones MD, Dalton E, Holden DW

(1995) Simultaneous identification of bacterial virulence genes by

negative selection. Science 269:400–403

17. He J, Shao X, Zheng H, Li M, Wang J, Zang Q, Li L, Liu Z, Sun

M, Wang S, Yu Z (2010) Complete genome sequence of Bacillus

thuringiensis mutant strain BMB171. J Bacteriol 192:4074–4075

18. Hugh-Jones M, Blackburn J (2009) The ecology of Bacillus an-

thracis. Mol Aspects Med 30:356–367

19. Ivanova N, Sorokin A, Anderson I, Galleron N, Candelon B,

Kapatral V (2003) Genome sequence of Bacillus cereus and

comparative analysis with Bacillus anthracis. Nature 423:87–91

20. Jensen GB, Hansen BM, Eilenberg J, Mahillon J (2003) The

hidden lifestyles of Bacillus cereus and relatives. Environ

Microbiol 5:631–640

21. Ju C, Gu H, Lu C (2012) Characterization and functional analysis

of Atl, a novel gene encoding an autolysin in Streptococcus suis.

J Bacteriol 194:1463–1473

22. Keim P, Gruendike JM, Klevytska AM, Schupp JM, Challacombe

J, Okinaka R (2009) The genome and variation of Bacillus cer-

eus. Mol Aspects Med 30:397–405

23. Kotiranta A, Haapasalo M, Kari K, Kerosuo E, Olsen I, Sorsa T,

Meurman JH, Lounatmaa K (1998) Surface structure, hydro-

phobicity, phagocytosis, and adherence to matrix proteins of

Bacillus cereus cells with and without the crystalline surface

protein layer. Infect Immun 66:4895–4902

24. Langridge GC, Phan M-D, Turner DJ (2010) Simultaneous assay

of every Salmonella typhi gene using one million transposon

mutants. Genom Res 19:2308–2316

25. Lee K, Costerton JW, Ravel J, Auerbach RK, Wagner DM, Keim

P, Leid JG (2007) Phenotypic and functional characterization of

Bacillus anthracis biofilms. Microbiology 153:1693–1701

26. Lereclus D, Agaisse H, Gominet M, Chaufaux J (1995) Over-

production of encapsulated insecticidal crystal proteins in a

Bacillus thuringiensis mutant. Nat Biotechnol 17:67–71

27. Lereclus D, Agaisse H, Gominet M, Salamitou S, Sanchis V

(1996) Identification of a Bacillus thuringiensis gene that posi-

tively regulates transcription of the phosphatidylinositol-specific

phospholipase C gene at the onset of the stationary phase.

J Bacteriol 178:2749–2756

28. Luo Q, Groh JL, Ballard JD, Krumholz LR (2007) Identification

of genes that confer sediment fitness to Deslfovibrio desulfuri-

cans G20. Appl Environ Microbiol 73:6305–6312

29. Mecsas J (2002) Use of signature tagged mutagenesis in patho-

genesis studies. Curr Opin Microbiol 5:33–37

30. Mei JM, Nourbakhsh F, Ford CW, Holden DW (1997) Identifi-

cation of Staphylococcus aureus virulence in a murine model of

bacteraemia using signature-tagged mutagenesis. Mol Microbiol

26:399–407

31. Menedez A, Fernandez L, Reimundo X, Guijarro JA (2007)

Genes required for Lactococcus garvieae survival in a fish host.

Microbiology 153:3286–3294

32. Nickerson KW, Bulla LA (1974) Physiology of spore forming

bacteria associated with insects: minimal nutritional requirements

for growth, sporulation, and parasporal crystal formation of

Bacillus thuringiensis. Appl Microbiol 28:124–128

33. O’ Sullivan LA, Weightman AJ, Hefin Jones T, Marchbank AM,

Tiedje JM, Mahenthiralingam E (2007) Identifying the genetic

basis of ecologically and biotechnologically useful functions of

the bacterium Burkholderia vietnamiensis. Environ Microbiol

9:1017–1034

34. Perehinec TM, Qazi SNA, Gaddipati SR, Salisbury V, Rees CED,

Hill PJ (2007) Construction and evaluation of multisite recom-

binatorial (Gateway) cloning vectors for Gram-positive bacteria.

BMC Mol Biol 8:80–91

35. Prabhakar A, Bishop AH (2012) Invertebrate pathogenicity and

toxin-producing potential of strains of Bacillus thuringiensis

endemic to Antarctica. J Inv Pathol 107:132–138

36. Quisel JD, Grossman AD (2000) Control of sporulation gene

expression in Bacillus subtilis by the chromosome partitioning

proteins Soj (ParA) and Spo0J (ParB). J Bacteriol 182:3446–3451

37. Robinson JT, Thorvaldsdottir H, Winckler W, Guttman M,

Lander ES, Getz G, Mesirov JP (2011) Integrative genomics

viewer. Nat Biotechnol 29:24–26

38. Ryan G, Sinclair RG, Rose JB, Hashsham SA, Gerba CP, Haas

CN (2012) Criteria for selection of surrogates used to study the

fate and control of pathogens in the environment. Appl Environ

Microbiol 78:1969–1977

39. Saile E, Koehler TM (2006) Bacillus anthracis multiplication,

persistence, and genetic exchange in the rhizosphere of grass

plants. Appl Environ Microbiol 72:3168–3174

40. Sambrook J, Fritsch EF, Maniatis T (1987) Molecular cloning: a

laboratory manual, 2nd edn. Cold Spring Harbor Laboratory,

Cold Spring Harbor

41. Schnepf E, Crickmore N, Van Rie J, Lereclus D, Baum J, Fei-

telson J et al (1998) Bacillus thuringiensis and its pesticidal

crystal proteins. Microbiol Mol Biol Rev 62:775–790

484 A. H. Bishop et al.: Soil Survival Genes of Bacillus thuringiensis

123

42. Schuch R, Pelzek AJ, Kan S, Fischetti VA (2010) Prevalence of

Bacillus anthracis-like organisms and bacteriophages in the

intestinal tract of the earthworm Eisenia fetida. Appl Environ

Microbiol 76:2286–2294

43. Seleem MN, Vemulappalli R, Boyle SM, Schurig GG, Sri-

ranganathan N (2004) Improved expression vector for Brucella

species. Biotechniques 37:740–744

44. Song Y, Xie C, Ong YM, Gan YH, Chua KL (2005) The BpsIR

quorum-sensing system of Burkholderia pseudomallei. J Bacte-

riol 187:785–790

45. van Opijnen T, Bodi KL, Camilli A (2009) Tn-seq: high-

throughput parallel sequencing for fitness and genetic interaction

studies in microorganisms. Nat Methods 6:767–772

46. Villafane R, Bechhofer DH, Narayanan CS, Dubnau D (1987)

Replication control genes of plasmid pE194. J Bacteriol

169:4822–4829

47. Vilain S, Luo Y, Hildreth MB, Brozel SV (2006) Analysis of the

life cycle of the soil saprophyte Bacillus cereus in liquid soil

extract and in the soil. Appl Environ Microbiol 72:4970–4977

48. Wilson A, Perego M, Hoch JA (2007) New transposon delivery

plasmids for insertional mutagenesis in Bacillus anthracis.

J Microbiol Methods 71:332–335

A. H. Bishop et al.: Soil Survival Genes of Bacillus thuringiensis 485

123