Serratia entomophila bet gene induction and the impact of glycine betaine accumulation on desiccation tolerance



Mark R.H. Hurst, AgResearch Lincoln, Private Bag 4749, Christchurch 8140, New Zealand. E-mail:



The genes involved in choline transport and oxidation to glycine betaine in the biopesticidal bacterium Serratia entomophila were characterized, and the potential of osmoprotectants, coupled with increased NaCl concentrations, to improve the desiccation tolerance of this species was investigated.

Methods and Results

Serratia entomophila carries sequences similar to the Escherichia coli betTIBA genes encoding a choline transporter and dehydrogenase, a betaine aldehyde dehydrogenase and a regulatory protein. Disruption of betA abolished the ability of Ser. entomophila to utilize choline as a carbon source. Quantitative reverse-transcriptase PCR analysis revealed that betA transcription was reduced compared to that of the upstream genes in the operon, and that NaCl and choline induced bet gene expression. Glycine betaine and choline increased the NaCl tolerance of Ser. entomophila, and osmotically preconditioned cultures survived better than control cultures following desiccation and immediately after application to agricultural soil.


Addition of glycine betaine and NaCl to growth medium can greatly enhance the desiccation survival of Ser. entomophila, and its initial survival in soil.

Significance and Impact of the Study

Serratia entomophila is sensitive to desiccation and does not persist under low soil moisture conditions. Techniques described here for enhancing the desiccation survival of Ser. entomophila can be used to improve formulations of this bacterium, and allow its application under a wider range of environmental conditions.


Serratia entomophila is a soilborne pathogen of the New Zealand grass grub, Costelytra zealandica (Coleoptera: Scarabaeidae). Once ingested by grass grub larvae, Ser. entomophila initiates a disease process known as amber disease characterized by gut clearance, cessation of feeding and ultimately death by starvation (Jackson et al. 1991, 1993; Jackson 2007). The pathogenic properties of Ser. entomophila have led to the development of this organism as a biopesticide for use in New Zealand pastures (Jackson et al. 2001). Similar to other members of the Enterobacteriaceae, Ser. entomophila is sensitive to ultraviolet light and osmotic and desiccation stress, so the bacteria must be applied by drilling to ensure they are placed beneath the surface of the soil. Therefore, use of the biopesticide, known as Bioshield, is currently limited to areas where subsurface application is feasible and there is adequate soil moisture to ensure bacterial survival, such as irrigated pasture (Jackson et al. 1991). Enhancing the environmental tolerance of Ser. entomophila to desiccation would greatly increase the efficacy and versatility of the biopesticide.

The trimethylammonium compound glycine betaine (betaine) is preferentially used as an osmoprotectant by bacteria, fungi and plants (Bremer and Kramer 2000; Rasanen et al. 2004). Betaine is a neutral solute that can be tolerated intracellularly at molar concentrations without interfering with essential biochemical functions of the cell (Brown 1976). Betaine, like other common osmoprotectants such as proline and ectoine, is accumulated by cells in response to changes in osmotic pressure caused by fluctuations in external osmolyte concentrations (Csonka and Hanson 1991; Bremer and Kramer 2000). Osmoprotectants have been used commercially to maintain the viability of bacterial cultures under hypersaline growth conditions (Prasad et al. 2003). Osmotically adapted cells also display a greater tolerance to subsequent periods of desiccation, thus allowing the production of viable desiccated cultures for biotechnological applications (Garcia de Castro et al. 2000; Tunnacliffe et al. 2001). This technique is referred to as anhydrobiotic engineering (Garcia de Castro et al. 2000; Manzanera et al. 2002) and is a rapidly expanding research area for the stabilization of biocontrol products (Garcia 2011).

In Gram-negative bacteria, betaine is either taken up from the environment by specific transporters or synthesized from its metabolic precursor, choline (Landfald and Strom 1986). In Escherichia coli, choline is imported into the cell through the high-affinity choline transporter, BetT, and then oxidized to form betaine aldehyde by the BetA choline dehydrogenase. The aldehyde is then oxidized by a betaine aldehyde dehydrogenase, BetB, to produce betaine. The second oxidation step can also be carried out by BetA (Bremer and Kramer 2000). The BetI protein regulates the expression of the bet genes through binding to the betT and betIBA promoter regions and inhibiting transcription. The presence of choline in the growth medium appears to alleviate BetI repression (Lamark et al. 1996; Malek et al. 2011). The betIBA genes of E. coli are transcribed as a single operon, with the betT gene being transcribed from a separate divergently orientated promoter (Lamark et al. 1996). Betaine can also be transported directly into the cell via the multisubstrate ProU and ProP transporters (Mellies et al. 1994, 1995), while some bacterial species also contain specific betaine transporters, such as the dual osmosensing/betaine transporter BetP from Corynebacterium glutamicum (Kramer and Morbach 2004).

To date, very little is known about the osmoregulatory pathways of Ser. entomophila. We have found that it is possible to significantly improve the environmental persistence of Ser. entomophila using anhydrobiotic engineering, and have identified the genes responsible for betaine synthesis from choline. Understanding the conditions necessary to induce betaine synthesis may allow the development of a more robust biopesticide.

Materials and Methods

Bacterial strains, media and culture conditions

Bacterial strains and plasmids are described in Table 1. Serratia entomophila strains A1MO2 and 626 were isolated from the gut of amber disease–infected C. zealandica larvae and are held in the AgResearch Lincoln culture collection. Strain 626 is used for the production of Bioshield, while A1MO2 is the more commonly studied laboratory strain that is amenable to molecular techniques. Serratia entomophila strains and E. coli strain DH10B were grown on Luria-Bertani (LB) medium at growth temperatures of 30 and 37°C, respectively. M9 minimal medium (Miller 1972) broth used for NaCl tolerance and gene induction assays contained 0·4% glucose as the sole carbon source to produce a high concentration of cells for RNA extraction. Solid M9 minimal medium used for plate assays contained 0·4% casamino acids as the sole carbon source and molecular-grade agarose (12 g l−1) (AppliChem, Darmstadt, Germany) as a solidifying agent. Where stipulated, NaCl (5 mol l−1) was added to media prior to autoclaving to achieve a final concentration of 0·6–0·8 mol l−1, depending on the experimental conditions being examined. Choline, glycine betaine, ectoine, proline and trehalose (Sigma Aldrich, St Louis, MO, USA) were added to growth media where stipulated to a final concentration of 1 mmol l−1. Antibiotics for the selection of Ser. entomophila and E. coli were used at the following respective final concentrations (μg ml−1): ampicillin 400 or 100, chloramphenicol 90 or 30, kanamycin 100 or 50 and spectinomycin 100. Caprylate-thallous agar (CTA) (Starr et al. 1976), DNase agar (Difco/BD, Franklin Lakes, NJ, USA), adonitol agar (ADN) (O'Callaghan and Jackson 1993) and itaconate agar (ITA) (O'Callaghan and Jackson 1993) were used for the selective isolation and identification of Ser. entomophila in desiccation and soil survival experiments.

Table 1. Bacterial strains and plasmids used in this study
Strain or plasmidDescriptionReference or source
  1. a

    AgResearch Lincoln culture collection.

Escherichia coli
DH10BF- araD139 Δ(ara, leu)7697 ΔlacX74 galU galK rpsL deoR Φ80lacZΔM15 endA1 nupG recA1 mcrA Δ(mrr hsdRMS mrcBC)Grant et al. (1990)
Serratia entomophila
A1MO2Path+ (pADAP)+ DNase+Grimont et al. (1988)
626Path+ (pADAP)+ DNase+AgRCCa
ΔbetTbetT::SmR/SpcR, A1MO2 derivativeThis study
ΔbetAbetA::CmR, A1MO2 derivativeThis study
pACYC184CmR, TcRChang and Cohen (1978)
pAOP8CmR, 6·5-kb amplicon encompassing betT, betI, betB and betA cloned into the XbaI and EcoRV sites of pACYC184This study
pAOPSCmR, SpcR, SmR, Ω fragment from pHP45 inserted in the BamHI site of pAOP8This study
pGEM-T EasyApR, lacZ multicloning sitesPromega
pHP45ΩApR, SmR, SpcR, contains Ω fragmentPrentki and Krisch (1984)
pJACTcR, CmR, 3·5-kb betA-CmR amplicon cloned into the BamHI site of pJP5608This study
pJP5608TcR, mob(RP4), lacZ multicloning sitePenfold and Pemberton (1992)
pLAFR3TcR, pRK290 with λcos lacZ and multicloning site from pUC8Staskawicz et al. (1987)
pLTSTcR, SpcR, SmR, 3·0-kb betT-Ω(SpcR) fragment amplicon cloned into pLAFR3This study

Cloning and sequencing Serratia entomophila bet genes

Escherichia coli (NC_002655, NC_000913.2) and Serratia marcescens (NC_009832) bet genes were aligned using ClustalW2 ( and primers designed to conserved regions flanking the putative ORFs. Serratia entomophila strain A1MO2 bet genes were PCR-amplified and cloned using the pGEM-T Easy vector system (Promega, Madison, WI, USA). The cloned amplicons were sequenced using universal M13 forward and reverse primers, and the sequence was assembled using the ContigExpress function of VectorNTI Advance 9 (Invitrogen, Carlsbad, CA, USA).

Reverse-transcriptase PCR analysis of Serratia entomophila bet operon structure

Total cellular RNA was isolated using an RNeasy Mini kit (Qiagen, Valencia, CA, USA). Serratia entomophila cultures were grown to mid-log-phase and aliquots transferred into tubes containing RNAprotect bacteria reagent (Qiagen). Pellets were processed immediately following the RNeasy Mini kit protocol for enzymatic lysis and proteinase K digestion of bacteria (Qiagen) and then subjected to two rounds of DNase digestion. First-strand cDNA synthesis reactions were carried out using random hexamer primers (Promega) and SuperScript III Reverse Transcriptase (Invitrogen) according to manufacturer's instructions. Resultant cDNA was used as template for PCR analysis of the bet gene mRNA transcript. Primers were designed to amplify the regions spanning the junctions of the bet genes (Fig. 1, Table S1).

Figure 1.

Schematic representation of the Serratia entomophila bet gene arrangement and RT-PCR confirmation of operon structure. (a) Organization of the bet genes and locations of primers used for RT-PCR analysis. The location of the hairpin secondary structure depicted in (b) is indicated by a star. (c) Agarose gel photographs showing the amplification of intergenic PCR products from RT-PCR analysis of the bet genes. M indicates DNA molecular weight marker, C denotes cDNA template, R denotes RNA template, - denotes H2O negative control, and D denotes DNA template. (d) G+C content of the Ser. entomophila bet gene region (window size 36, global GC content 60%). ORFs are indicated by block arrows with corresponding G+C% of the region superimposed. G+C content analysis carried out using PlasmaDNA (Angers-Loustau et al. 2007).

Construction of betT and betA insertion mutants

Primer pairs BetTF/BetTR and BetAF/BetAR (Table S1) were used to amplify the betT and betA genes, respectively, from the Ser. entomophila strain A1MO2. The resulting amplicons were ligated into pGEM-T Easy. The spectinomycin resistance gene from the Ω fragment of vector pHP45Ω was amplified using EcoRV-flanked primers SPRVF and SPRVR (Table S1) and ligated into the MscI site of the betT amplicon. The entire betT-SpR fragment was excised from pGEM-T Easy by digestion with BamHI and ligated into the analogous site of vector pLAFR3, forming pLTS. The chloramphenicol resistance cassette from pACYC184 was amplified using primers ChlF and ChlR (Table S1) and ligated into a unique AgeI site located within the betA gene. The betA-CmR fragment was excised from pGEM-T easy and ligated into suicide vector pJP5603, generating pJAC.

pLTS and pJAC were transformed separately into Ser. entomophila A1MO2, and mutants in which the antibiotic resistance gene had integrated into the genomic bet gene were identified as described previously (Hurst et al. 2000). The recombinants were validated by PCR and DNA sequence analysis and designated betT::aadA and betA::cat.

Complementation of betT and betA mutant strains

Long-range PCR was carried out using the Expand Long Template PCR System (Roche, Madison, WI) to amplify the entire bet gene region from strain A1MO2. Primers OPF2 and OPR2 (Table S1) amplified a 6·5-kb product which was cloned into pGEM-T Easy and validated by DNA sequence analysis. The bet region was excised from pGEM-T Easy using SpeI and XmnI and ligated into the XbaI and EcoRV sites of vector pACYC184, resulting in pAOP8. pAOP8 was transformed into Ser. entomophila strain betT::aadA. To enable selection in the CmR background of strain betA::cat, a spectinomycin resistance cassette (derived from pHP45Ω) was amplified using the primer pair SPHIF/SPHIR (Table S1) and ligated into the BamHI site of pAOP8, generating pAOPS. pAOPS was then transformed into the Ser. entomophila A1MO2 derivative betA::cat.

Choline utilization assays

A1MO2 wild-type, the betT and betA insertional mutants, and the complemented mutant strains were grown in broth cultures of M9 minimal medium and M9 containing 0·6 mol l−1 NaCl with or without choline or betaine. Three replicate flasks for each treatment were set up. Samples were taken every 2 h over a 24-h period for optical density reading and enumeration of viable cells.

Quantitative reverse-transcriptase PCR (RT-qPCR) analysis

Cultures were grown to mid-log-phase in M9 minimal medium before induction with an inhibitory but sublethal concentration of NaCl (0·6 mol l−1), choline, NaCl and choline, or NaCl and betaine. An uninduced control culture was included to examine natural fluctuations in bet gene transcription throughout the experiment. RNA and cDNA were prepared as described above. Primers were designed for each of the bet genes as well as the reference genes rpoD, 16S ribosomal RNA and the Ser. entomophila virulence–associated plasmid pADAP replication gene, repA (Hurst et al. 2011) (Table S1). Being a plasmid, the expression of repA is intrinsically linked to the host cell replication. Primer concentrations used in final reactions were determined empirically. The amount of transcript in each culture at regular time points was compared to amounts present before cultures were induced. Expression of each of the bet genes was normalized against that of the reference genes, and significant changes in gene expression levels were determined using the REST-MCS program ver. 1.9.6 ( Each reaction was carried out three times in duplicate using independent samples.

NaCl-challenge Serratia entomophila growth assays

M9 minimal medium (50 ml) containing 0·6 mol l−1 NaCl was inoculated with overnight starter culture of wild-type A1MO2. Osmolarity challenge assays were conducted as follows: choline, betaine or trehalose was added to each flask and cultures were then incubated at 30°C with aeration. Positive (M9-only, no NaCl) and negative (NaCl, not supplemented with osmoprotectant) control cultures were included for comparison. Aliquots of each culture were removed at 30-min intervals for optical density readings and viable cell counts. Measurements were taken every 2 h for 20 h, or until cultures reached stationary phase. All growth curves were performed in triplicate.

Desiccation experiments

Serratia entomophila was grown to late log-phase in M9 minimal medium alone or with the addition of 0·6 mol l−1 NaCl and 1 mmol l−1 betaine. Cells were pelleted by centrifugation at 4°C and washed twice with sterile cold water. Cell pellets were resuspended in a final volume of 1·5 ml water or 1 mmol l−1 betaine solution. Aliquots of cell suspension (50 μl) were transferred into microcentrifuge tubes and placed with open lids into a Concentrator 5301 (Eppendorf, Hamburg, Germany) and desiccated for 2 h at 30°C. Samples were weighed before and after desiccation to determine the per cent moisture loss. Following desiccation, samples were placed in closed cardboard storage boxes and maintained at 4°C or 20°C. At each sampling point (prior to desiccation, immediately postdesiccation, and 1, 7, 14, 30 and 60 days postdesiccation), 50 μl sterile LB broth was added to triplicate sample tubes before incubation at 20°C for 15 min to rehydrate the desiccated cells. Total viable counts were determined for each sample.

Persistence of Serratia entomophila in soil

Serratia entomophila cultures were grown to late log-phase (OD600~1·5) in 50 ml volumes of M9 minimal medium or medium containing 0·6 mol l−1 NaCl + 1 mmol l−1 betaine, then incubated on ice for 30 min. Cultures were pelleted and the supernatant was discarded. Cell pellets were resuspended in 25 ml sterile cold water and recentrifuged. Cell pellets were then resuspended in cold water to achieve a final volume of 2 ml.

Pastoral soil that had not previously been exposed to chemical or biological pesticides was collected and sieved to remove large sediment and plant residue. Soil was weighed into sterile Petri dishes to give 10 g dry soil weight per dish. Bacterial cell suspensions were diluted in water and then added to soil to give a final cell count of ~1 × 107 colony forming units (CFU) g−1 dry soil and soil moisture contents of 15% or 25%. A third treatment condition, referred to as the ‘drying’ sample, was prepared with an initial soil moisture of 25% and allowed to dry out throughout the experimental period. A sterile spatula was used to mix the moistened soil and ensure the cell suspension was evenly distributed throughout each soil. Each Petri dish was weighed so that any moisture loss throughout the experiment could be determined. The decrease in soil moisture throughout the experimental period was comparable between the three treatment conditions (data not shown).

Samples were transferred to Scienceware desiccator jars (Bel-Art, Pequannock, NJ, USA) containing either 300 ml water (for constant soil moisture treatment) or 10 g dry desiccator beads (for drying soil treatment). Desiccator beads were replaced every 14 days to ensure steady drying of soil. Desiccator jars were stored at a constant temperature of 19°C for the duration of the experiment (90 days). Triplicate samples of each treatment were sampled at each of the seven time points: immediately postinoculation and days 1, 7, 14, 30, 60 and 90. The initial number of CFU added to each soil was determined by serially diluting the inoculum culture and plating on CTA agar plates. The experiment was carried out twice, using freshly collected field soil.

Enumeration of bacteria from soil

At each time point, soils were destructively sampled to determine the numbers of surviving bacteria. Each Petri dish sample was weighed to determine any weight loss corresponding to moisture loss throughout the storage time. Samples were then transferred to sterile flasks and soil weights recorded. Approximately 90 g sterile peptone water was added to each flask to produce a 1 : 10 dilution of the samples. Flasks were transferred to a sonicating water bath for 3 min, and soil suspensions were then serially diluted in phosphate buffer. Samples were plated onto Serratia selective agar (CTA) and incubated at 30°C for 6 days. After colonies had been enumerated, 10% of colonies with a colony phenotype typical of Serratia spp. were identified as described previously (O'Callaghan and Jackson 1993).

Statistical analyses

Student's t-tests were carried out using GraphPad software where appropriate. REST-MCS ver. 2.0 (Pfaffl 2001) was used for the analysis of qPCR data. For the soil experiments, the observed bacterial counts of two cultures at time = 0 were compared to their theoretical values (values expected from the applied concentrations) using the two-sample Wilcoxon test. Because the theoretical values were different between the cultures, this comparison was made on normalized difference, that is, relative difference = (theoretical value – observed value)/theoretical value. The changes in bacterial count over the time of observation period were compared between the six treatment groups, which are defined by combinations of the two cultures and three soil moisture levels. These comparisons were made using a nonlinear regression approach by fitting treatment-group-specific exponential decay curves to log10-transformed observed counts. Each of these curves consists of two parameters: α and β, α corresponding to the estimated mean bacterial count at time = 0 and β the estimated average amount of change per day from the mean count. Significance was accepted at P < 0·05 (95% confidence interval) for all analyses.

Nucleotide sequence accession numbers

The nucleotide sequence of the Ser. entomophila bet gene region was deposited in GenBank under accession number EU399653.


DNA sequence analysis of the Serratia entomophila strain A1MO2 bet genes

A 6426-bp contig corresponding to the Ser. entomophila bet gene region was assembled. Analysis of the DNA sequence and translated protein sequence revealed four predicted open reading frames (ORFs). The Ser. entomophila bet region displayed a global GC content of 60%, with a marked GC decrease in the areas corresponding to the predicted intergenic regions (Fig. 1d). Upon comparison with the corresponding gene region in E. coli, the Ser. entomophila bet gene cluster was found to have a relatively large (226-bp) region separating the betB and betA genes that was not present in E. coli (Fig. 1a).

Analysis of the DNA secondary structure in the Ser. entomophila betBbetA intergenic region using mFold revealed a double hairpin structure 23 bp downstream of betB. The 66-bp region contained two sets of inverted repeats forming the stems of the hairpins, with three nucleotides separating the two hairpin structures (Fig. 1b). A ΔG of −16·6 kcal mol−1 was calculated for the double hairpin, suggesting that the structure is stable at physiological growth temperatures (30–37°C). Analysis of related Gram-negative species revealed that Ser. proteamaculans 568, Pseudomonas aeruginosa PAO1, Acinetobacter ADP1, Rhizobium etli CFN42 and Sinorhizobium meliloti 1021 also had a slightly larger intergenic region separating the betB and betA genes than that described in E. coli. Although there was no significant similarity between the sequences, all strains with >20 bp separating betB and betA contained multiple hairpin structures in the noncoding region (data not shown).

betI, betB and betA are transcribed as a polycistronic operon

The presence of the large noncoding region between betB and betA in the Serentomophila sequence suggested that these genes may not be transcribed as part of a single betIBA operon. To ascertain whether this was the case, cDNA was prepared and assayed by RT-PCR using intergenic primer sets. Products were amplified from cDNA template in the areas corresponding to the predicted betB–betA and betI–betB intergenic regions (Fig. 1c), confirming the presence of these three genes on a single mRNA transcript. A product spanning the betTbetI gene region could be PCR-amplified from genomic DNA, but not when cDNA was used as template (Fig. 1c). These results indicate that the Ser. entomophila genes betI, betB and betA are transcribed as an operon, and confirm that betT is divergently transcribed.

betA is essential for utilization of choline as an osmoprotectant

Reduced ability to utilize choline as an osmoprotectant under conditions of osmotic stress has been reported previously for bet gene mutants (Rasanen et al. 2004). To determine whether the bet genes are essential for the conversion of choline to betaine in Ser. entomophila, betT and betA mutants and their respective complemented strains were constructed. These four strains and the A1MO2 wild type were grown in M9 broth medium with NaCl and choline or betaine. Supplementation of M9 broth with 0·6 mol l−1 NaCl (Fig. 2b) resulted in a decrease in CFU of 1–2 orders of magnitude for all strains compared to the M9 control medium (Fig. 2a). In M9 broth supplemented with both NaCl and choline, the betA mutant was the only strain not able to utilize choline as an osmoprotectant, demonstrated by a viable cell count 10-fold lower than that of the other strains (Fig. 2c). When grown with NaCl plus betaine, the growth of the betA::cat mutant did not significantly differ from wild type (Fig. 2d), confirming that disruption of betA abolished the ability of Ser. entomophila to convert choline to betaine for use as an osmoprotectant. Complementation of the betA mutant restored the ability of the strain to utilize choline (Fig. 2c). Disruption of betT in Ser. entomophila by insertional mutagenesis did not inhibit the growth under conditions of increased osmolarity when the medium was supplemented with choline, indicating that betT may not be the sole choline transporter for this species.

Figure 2.

Growth of Serratia entomophila A1MO2 wild-type, betT::aadA, betA::cat and complemented mutant strains in M9 minimal medium alone (a) or medium treated with NaCl (b), NaCl and choline (c) or NaCl and betaine (d). NaCl concentration was 0·6 mol l−1, and osmoprotectant concentrations were 1 mmol l−1 each. Graphs represent the average colony numbers determined from three independent experimental replicates after 24 h of incubation at 30°C. The observed growth inhibition of strain betA::cat in the NaCl and choline treatment compared to the M9-only control is statistically significant (P = 0·002), as is the resumption of growth of this strain upon supplementation with betaine (P < 0·001).

Serratia entomophila bet gene transcription is induced by NaCl and choline

Quantitative RT-PCR was used to examine the levels of Ser. entomophila A1MO2 bet gene transcript present in the 60-min interval following amendment of the medium with NaCl and/or choline or betaine. A statistically significant (P < 0·05) increase in gene expression was observed under several conditions (Table 2). The large increase in gene expression observed following the addition of NaCl plus choline confirmed that while bet gene expression was osmotically induced in Ser. entomophila, the levels were enhanced by the addition of choline. The addition of betaine in conjunction with NaCl prevented the increase in betT and betI expression that was observed upon the addition of NaCl alone, suggesting that betaine acts as a corepressor of the bet operon and thereby an inhibitor of choline conversion to betaine. The results showed that at 5 min postinduction, the only significant increase in gene expression in the presence of choline was that of betI (P = 0·003). This rapid induction suggested that the addition of choline induced betI transcription. The decrease in betI induction levels after 60 min may reflect steady-state transcription in the absence of hyperosmotic stress.

Table 2. Effect of NaCl, choline and betaine on Serratia entomophila bet gene expression
ConditionaTime (min)betTbetIbetBbetA
  1. a

    Substrate added to M9 growth medium to induce changes in gene expression.

  2. Change refers to the fold change compared to the uninduced control. Significant fold changes and their respective P-values are indicated in boldface.

NaCl + choline5−1·540·7722·720·3371·200·964−1·410·857
NaCl + betaine5−2·710·316−1·770·619−2·960·299−1·950·550

Osmoadaption in Serratia entomophila

The ability of small organic solutes to aid in the NaCl tolerance of Ser. entomophila has not previously been examined. We first determined that NaCl at a concentration of 0·6 mol l−1 inhibited the growth of Ser. entomophila in M9 minimal medium, as shown by no increase in culture density or viable cell numbers over a 20-h incubation period (Fig. 3a,b). The addition of choline or betaine to the culture medium allowed the cells to overcome NaCl inhibition, with the cultures having similar growth kinetics to the positive control (Fig. 3a,b). Trehalose, a well-described osmolyte, did not alleviate NaCl-induced growth inhibition (Fig. 3a,b) in this model. An environmental stress, such as exposure to extreme temperature or osmolarity, has been shown to cross-protect bacteria against subsequent and divergent stresses (Garcia de Castro et al. 2000; Mitchell et al. 2009). We investigated whether exposure of Ser. entomophila to NaCl during growth could assist in desiccation tolerance and storage survival.

Figure 3.

Growth of Serratia entomophila A1MO2 in M9 minimal medium with and without 0·6 mol l−1 NaCl and 1 mmol l−1 osmoprotectant over 20 h at 30°C. Measured by optical density (a) and CFUs (b). Positive control is M9 minimal medium alone (no NaCl), and negative control is M9 minimal medium with 0·6 mol l−1 NaCl only (no osmoprotectant). Osmoprotectants and NaCl were added to growth medium prior to inoculation with bacterial culture. Error bars represent the standard error of the means of three individual experimental replicates. (image_n/jam12052-gra-0001.png) M9, (image_n/jam12052-gra-0002.png) Nacl, (image_n/jam12052-gra-0003.png) Choline, (image_n/jam12052-gra-0004.png) Betaine, (image_n/jam12052-gra-0005.png) Trehalose.

Following desiccation and storage at 4°C or 20°C, Ser. entomophila cultures grown in the presence of NaCl and betaine were 1000-fold more tolerant of desiccation than those grown in M9 minimal medium alone (Fig. 4). Pre- and postdesiccation tube weights indicated that the cell pellets lost >98% of their original moisture content. As expected, cells stored at 4°C remained viable for longer than those stored at 20°C (Fig. 4). Resuspension of cell pellets in betaine solution prior to desiccation did not provide a survival advantage compared to cells resuspended in H2O alone (data not shown).

Figure 4.

Viability of Serratia entomophila A1MO2 following desiccation and storage at 4°C or 20°C. Cultures were grown in M9 minimal medium alone or with additional betaine (1 mmol l−1) or betaine and NaCl (0·6 mol l−1) prior to desiccation and storage. LSD bars indicate significance at P = 0·05 for each sampling point. Conditions are indicated as follows: M9 medium at 4°C, black line/solid circle; M9 medium at 20°C, dashed line/open triangle; betaine supplementation at 4°C, dashed line/filled square; betaine supplementation at 20°C, dashed line/open diamond; NaCl plus betaine supplementation at 4°C, dotted line/filled triangle; NaCl plus betaine supplementation at 20°C, dash plus dot line/open circle.

Enhancing the survival of Serratia entomophila in soil

The enhanced survival of desiccated cultures when preconditioned with NaCl and betaine led us to examine the use of this technique for enhancing survival of Ser. entomophila biopesticide strain 626 in soil. A small-scale experimental model was developed using various soil moisture conditions, and the experiment was carried out in duplicate. Similar trends were observed in the replicate experiments, but as rates of decline differed between them, data from each experiment were analysed separately. Overall, under all soil conditions tested, cells grown in M9 medium containing 0·6 mol l−1 NaCl and 1 mmol l−1 betaine (NaCl-bet medium) survived better than cells cultured in M9 medium (P < 0·05) (Fig. 5a–c). This difference was most apparent when cells were first applied to soil. An initial decrease in cell numbers of ~ 0·5-log10 CFU g−1 soil was observed in M9 cultures immediately after inoculation into soil (Fig. 5a–c). However, a similar decrease did not occur when the cells had been cultured in NaCl-bet medium, where a significantly greater number of cells survived the initial application to soil in both experiments (P < 0·05 and P < 0·001, for experiments 1 and 2, respectively). The greater numbers of the NaCl-bet-cultured cells were maintained under the three soil conditions at significantly higher (P < 0·05) cell numbers for approximately 7 days postinoculation (Fig. 5a–c). After this time, the difference between the two culture treatments did not significantly differ.

Figure 5.

Mean CFU (log10 ml−1) of Serratia entomophila 626 after inoculation into soil. Samples were inoculated into moist soil (25%SM) (a), dry soil (15%SM) (b) and drying soil (25% decreasing to 15%SM) (c). All samples were stored at 19°C throughout the time course. Error bars represent the standard deviation of the mean of triplicate samples. Data from experimental replicate 1 are presented; similar trends were observed in experimental replicate 2. (image_n/jam12052-gra-0001.png) M9, (image_n/jam12052-gra-0006.png) Nacl-Bet.


The use of Ser. entomophila as a biopesticide against New Zealand grass grub is hampered by the desiccation sensitivity of the bacterium, limiting its ability to be formulated in a dry state and its immediate survival postapplication to soil. We examined the ability of the osmoprotectant glycine betaine, coupled with pre-exposure to osmotic stress, to protect Ser. entomophila against subsequent desiccation damage upon long-term storage or application to desiccated soils. We first characterized the genes responsible for choline transport and subsequent oxidation into glycine betaine. The Ser. entomophila betTIBA genes were identified, and the gene arrangement was found to be similar to that of E. coli (Eshoo 1988). RT-PCR assays confirmed that the betIBA genes were transcribed as a single operon. Following construction of insertion mutants of the betT and betA genes, we confirmed that betA was essential for the utilization of choline to mitigate the growth inhibition observed under increased NaCl conditions. Interestingly, the betT mutation did not result in any growth or substrate utilization changes, suggesting that BetT is not the only choline transport protein used by Ser. entomophila. In E. coli, the periplasmic binding-protein-dependent ABC transporter, ProU, can bind choline with low affinity (Styrvold et al. 1986; Eshoo 1988; Choquet et al. 2005). It is possible that a Ser. entomophila ProU-equivalent transporter can compensate for the loss of BetT. The bacterium P. aeruginosa contains two BetT symporters, BetT1 and BetT3, which are alternately dominant under hypo- and hyperosmotic conditions, respectively (Malek et al. 2011). Nevertheless, the observed decrease in growth of the Ser. entomophila betA mutant in NaCl medium supplemented with choline, and the restoration of growth when betaine was added in place of choline, confirmed the functionality of the bet operon in Ser. entomophila.

Using reverse-transcriptase qPCR assays, we observed that the level of betA mRNA following induction with NaCl and choline was significantly less than that seen for betB, the gene located upstream of betA on the betIBA operon. This may be accounted for by the 3′ location of betA on the transcript, which is generally less stable than the 5′ end of the transcript. It is also known that the use of random hexamers to generate cDNA enriches for sequences at the 5′ end of a transcript. This may have reduced the ability to detect the betA transcript. However, it should be noted that the distance between the betB and betA qPCR products used in the study was 903 bp, a distance we considered too small to allow a disproportionately large 5′ bias due to the use of random hexamers. In addition, the noncoding region between the Ser. entomophila betB and betA genes contained a double hairpin structure 23 bp downstream of the betB gene (Fig. 1b) that may allow differential processing of the mRNA (Regnier and Arraiano 2000). For example, in the Bacillus subtilis pst operon, a stem-loop between the first gene of the operon pstS and the following pstA gene allowed differential processing of the operon mRNA. The 5′ portion of the mRNA encoding pstS had a half-life of 11 min, whereas the full-length transcript was rapidly degraded, with a half-life of less than 1 min (Allenby et al. 2004). Attempts to construct a betB–betA intergenic mutant strain in Ser. entomophila proved unsuccessful (data not shown), suggesting that the hairpin region may be essential. Because BetA converts choline to the toxic intermediate betaine aldehyde, it would be beneficial for Ser. entomophila to produce a lesser amount of BetA than BetB, to prevent the accumulation of the toxic molecule. Further work is required to determine whether the hairpin region has an effect on gene transcription or mRNA stability.

The finding that choline and NaCl induced betIBA operon expression in Ser. entomophila led us to investigate the use of anhydrobiotic engineering to prolong the shelf life of desiccated bacterial cultures. The decline in bacterial viability during storage is a major obstacle for the production of successful biopesticides. Pre-exposure to increased NaCl in the growth medium, along with the addition of betaine, produced a significant increase in the shelf life of desiccated Ser. entomophila. The storage survival experiments, depicted in Fig. 4, showed that there was no advantage conferred upon cells grown in medium supplemented with betaine alone, compared to the nonsupplemented medium (M9 only control). The only cells with a survival advantage were those grown with both NaCl and betaine. This finding confirmed that prestressing is required for the uptake of betaine, and therefore, both NaCl and betaine are necessary to achieve the final survival advantage. While this was an extreme example of the desiccation stresses imposed on the biopesticide during storage, the positive results suggest that Ser. entomophila formulations would benefit from NaCl and betaine supplementation during the production of inoculum to aid in their shelf life and possibly their survival following application to pastoral soil.

The soil trials revealed that the combination of stresses imposed upon bacterial cells after the addition to soil could be alleviated, at least initially, through the preconditioning of cells with NaCl and betaine. However, the lack of treatment effect beyond 7 days suggested that the protective effects of anhydrobiotic engineering were outweighed by the challenges faced by the inoculated bacteria in the soil environment, including competition with the natural soil microflora and predation. Cell numbers declined to below detectable levels after 60 days regardless of the growth medium; this decline was expected in the absence of the insect host, which is necessary to allow long-term survival of the pathogen. The increased survival in the first several days postapplication may nevertheless increase the efficacy of the biocontrol agent, as 7 days is sufficient time for the bacteria to be taken up by the insect larval host.

NaCl and betaine have been used to enhance the desiccation tolerance of Pantoea agglomerans, a biocontrol agent used to control fungal pathogens, when applied to the surface of apples stored for 15 days under various levels of relative humidity (RH). Osmoadapted cultures showed greatly enhanced survival and biocontrol properties on the surface of apples stored under low and fluctuating RH, but not under the conditions of high RH (Bonaterra et al. 2005). Anhydrobiotic engineering is also being investigated for the development of microbial strains for application to soil in a rhizoremediation capacity (Vilchez and Manzanera 2011). Formulations of Ser. entomophila that could be applied to the surface of pasture would be commercially desirable as it would extend the range of situations in which the biopesticide could be used. Further development of the anhydrobiotic engineering process in Ser. entomophila, possibly using a combination of osmoprotectants, may increase the advantage of the NaCl-bet-cultured cells that was observed for the first 7 days postinoculation into soil. The techniques described here present an opportunity for the development of bacterial inocula that are normally limited by adverse soil conditions.


This research was supported by a grant from the New Zealand Foundation for Research, Science and Technology (FRST), grant number C10X0301. The authors thank Dr Campbell Sheen for critical reading of this manuscript.