Termite-killing components in Serratia marcescens (SM1)

The bacteria, Serratia marcescens (SM1) was previously obtained from the black-winged termite, Odontotermes formosanus Shiraki. SM1 was highly toxic to O. formosanus, however, the mechanism of toxicity is unclear. In this study, toxicity test results showed that the main components that affected O. formosanus were in a supernatant and that the insecticidal protease in the supernatant resulted in the death of O. formosanus. In addition, zinc sulphate recovery experiments indicated that the metalloproteinases in the supernatant were more harmful. These results provide a theoretical foundation for the future biological control of termites, the basis for the development of pest control technology and the discovery of new pesticides.


Introduction
including the cotton bollworm, Helicoverpa armigera Hüber (Chen et al. 2005), Myrmeleotettix palpalis (Jin et al. 2005), Phyllotreta striolata (Yang et al. 2014), the tobacco cutworm, Spodoptera litura Fabricius (Niu et al. 2015;Aggarwal et al. 2017), S. exigua (Niu et al. 2015), and the larvae and eggs of the Asian palm weevil, Rhynchophorus ferrugineus Oliver (Zhang et al. 2011). Secretory product of S. marcescens, prodigiosin, have also been found to be pathogenic to some organisms, such as Bursaphelenchus xylophilus (Hu et al. 2017), the yellow fever mosquito, Aedes aegypti L. and the Indian malaria mosquito, Anopheles stephensi Liston (Patil et al. 2011). Niu et al. (2018) used S. marcescens mixed with five insecticides to treat Laodelphax striatellus and found that the mortality rate of L. striatellus to the insecticides could be improved. It is clear that S. marcescens is pathogenic to many types of pests and has wide market prospects in biological control.
Our laboratory recently isolated a red pigment-producing bacterium from dead termites and identified it as Serratia marcescens (SM1) (Fu et al. 2019). SM1 was significantly toxic to O. formosanus (Fu et al. 2020), but the specific components and mechanisms of this toxicity need further examination. Research has shown that the pathogenesis of S. marcescens is due to chitinase enzymes (Regev et al. 1996;Zhang et al. 2000;Xu and Peng 2004;Yin et al. 2004;Jin et al. 2005), and Tao 2006 has reported that the insecticidal protein of S. marcescens is a metallic protein that exists in a supernatant of the bacteria. However, the active components of S. marcescens that are toxic to termites have not been identified. The purposes of this study are: (1) to clarify the termite-killing components of S. marcescens; and, (2) to provide a basis for the biological control of termites.

Insects
Six colonies of Odontotermes formosanus were collected from Jurong in Zhenjiang, Jiangsu Province, China. Termites were kept in sealed plastic containers in total darkness at 27 ± 1 °C and 75 ± 1% relative humidity. All colonies were maintained under laboratory conditions without soil and with moist filter paper for one day before the subsequent experiments.
Culture and concentration determination of the S. marcescens strain, SM1 SM1 was isolated from infected O. formosanus and deposited at Nanjing Forestry University, Nanjing. A solid bacterial medium (1 L) was prepared, consisting of 10 g peptone, 20 g beef extract, 2 g NaCl, 2 g K 2 HPO 4 , 18 g agar and 1 L H 2 O with pH 7.2-7.4. A seed medium (1 L) of 10 g peptone, 20 g yeast extract, 2 g NaCl, 2 g K 2 HPO 4 in 1 L of H 2 O was also prepared. A zymotic medium (1 L) was prepared of 10 g peptone, 30 g soybean oil, 2 g NaCl, 2 g K 2 HPO 4 in I L H 2 O. As the amount of SM1 increased, its optical density at 600 nm (OD 600 ) also increased. The OD 600 value was linearly related to the number of S. marcescens SM1 bacteria and used to determine the bacterial count of the liquid culture. To obtain the concentration of the S. marcescens SM1 suspension, the OD 600 value was measured on a spectrophotometer.

Bioassay of components of S. marcescens SM1 fermentation
The concentration of the fermentation medium of S. marcescens SM1 was 1.17 × 10 10 cells/ml, and the fermentation medium of S. marcescens SM1 was centrifuged at 5000 r for 20 min at 4 °C. After centrifugation, a surface oil liquid (pigment layer), a middle layer supernatant and a lower layer of bacterial precipitate were obtained. The lower layer of the bacteria precipitate was dissolved by medium. According to the transfer toxicity method (Tang et al. 2007), quality filter paper was first wetted, placed in a 9-cm diameter glass Petri dish, and five O. formosanus worker termites were dipped in each treatment solution. Five soaked termites marked in red on the abdomen and twenty untreated worker termites were placed in the culture dish. The five marked worker termites dipped in the medium and the 20 untreated worker termites were the controls. The termites were observed and recorded in darkness at 25 °C. The experiment group and the control groups were repeated three times in the same manner.

Effect of protease inhibitors on the bioassay of the SM1 fermentation
Bioassay of components of S. marcescens SM1 fermentation showed that the culture supernatant of the bacterial fermentation had a strong toxicity to O. formosanus. Therefore, the supernatant was treated separately with proteinase K at 50 μg/ml, ethylenediaminetetraacetic acid (EDTA) at 10 mmol/L, 1, 10-phenanthroline at 10 mmol/L, and Corrected mortality rate treatment (%) = mortality rate treatment − mortality rate ck ∕ 100 − mortality rate ck × 100 Termite-killing components in Serratia marcescens (SM1) phenylmethanesulfonyl fluoride (PMSF) at 1 m mol/L. There was also a treatment without reagents added and was only heated at 60 °C for 15 min. The toxicity assay was then performed using the transfer toxicity method. The five marked worker termites dipped in the medium and the 20 untreated worker termites were the controls. The experimental groups and the control groups were assayed three times.

Recovery experiment of ZnSO 4
After the bioassays using the supernatant treated with different protease inhibitors, the supernatants treated with the two types of protease inhibitors, EDTA and 1,10-phenanthroline, were subjected to zinc sulphate (ZnSO 4 ) recovery treatments. The two supernatants were incubated with ZnSO 4 at a concentration of 0.5 mmol/L for 1 h at 22 °C, and the same toxicity assay performed again. The medium incubated at 22 °C for 1 h with a concentration of 0.5 m mol/L ZnSO 4 was the control. The concentration of the SM1 fermentation was 1.69 × 10 10 cells/ml and the toxicity assay was then performed using the transfer toxicity method. The experimental group and the control group were independently assayed three times.

Statistical analysis
The results were subjected to an analysis of variance (ANOVA) using InStat software (GraphPad, San Diego, CA, USA) with a level of significance at P < 0.05. Tukey's test was used for multiple comparisons.

Bioassay of components of the SM1 fermentation
The transfer toxicity bioassay against O. formosanus was carried out using the components of the SM1 fermentation. The results show that the toxicity of the components was not different from that of the supernatant (P > 0.05). After 36 h, the corrected mortality rate of the supernatant was approximately 50%, and at 54 h, the rate was close to 80%. After 54 h, toxicity of the lower bacterial layer was relatively poor at only 28.5%. Among the three components, the uppermost oil pigment layer had the poorest toxicity (Table 1).

Effect of protease inhibitors on the bioassay of SM1 fermentation
After the middle layer supernatant was identified as the main active component, it was treated with protease inhibitors. It was first heated at 60 °C for 15 min to denature the protease. The results show that the toxicity effect of the supernatant decreased significantly (P < 0.05) after heating; the toxicity became stronger after 48 h and the corrected mortality reached 60% (P > 0.05). Four different protease inhibitors, PMSF, EDTA, proteinase K and 1, 10-phenanthroline, were added separately to the supernatant. The results show that the toxicity decreased when protease inhibitors were used. When proteinase K was used, the toxicity of the supernatant decreased most, and the corrected mortality rate was 10.4% at 54 h; when EDTA was used, the toxicity of the supernatant was greatly reduced, and the corrected mortality rate was 22.4% at 54 h; when PMSF or 1, 10-phenanthroline was used, the toxicity of the supernatant was reduced, and the corrected mortality rate failed to reach 50% by 54 h (Table 2).

Discussion
Research on biopesticide proteins has been important in the development of biopesticides. Over the past few years, several studies of Bacillus thuringiensis insecticidal protein (Schnepf et al. 1998) have shown that protein crystals produced during sporulation are toxic to many organisms. The expression and structural functions of the protein genes have been studied by McGaughey et al. (1998) and Rajamohan et al. (1998). There are four gene clusters encoding the toxin protein complex in the luminescent bacillus, which makes these bacteria that are symbiotic with entomopathogenic nematodes a member of the microbial insecticide protein family (Bowen et al. 1998). Reports of nematophagous pathogenic bacilli also earlier suggested this symbiosis with entomopathogenic nematodes, and their insecticidal components have a lethal effect on a variety of pests (Forst et al. 1997). Studies have shown that active ingredients that play a major insecticidal role include small molecules (Volgyi et al. 1998), lipopolysaccharides (Smigielski and Akhurst 1995) and proteins. A 1.2 kb gene was obtained from the nematophagous pathogenic bacterium A24, which encodes a 30 kDa insecticidal protein. A few microliters of the fermentation containing the toxic protein can cause the death of insect larvae (Smigielski and Akhurst 1995). Moreover, Clostridium difficile (Barloy et al. 1998) and Bacillus sphaericus (Li et al. 2008) also produce insecticidal proteins which are toxic. Studies have shown that S. marcescens secretes several well-known extracellular proteins, including nuclease, phosphatase, hemolysin, iron-containing proteins, chitinase, protease, and lipase (Tao 2006). One or several of these components may become insecticide factors. Therefore, studying the specific components of S. marcescens SM1 can help to further clarify its mechanism of action on termites.
The uppermost layer of the pigment oil layer, the middle layer supernatant and the bacterial sediment layer were obtained by centrifugation of S. marcescens SM1 fermentation. The results show that the toxicity effect of the supernatant was the best and closest to that of the original fermentation. The toxicity of the bacterial sediment layer was poor and differed from that of the original fermentation. In addition, the pigment oil layer had the least effect, and the 3.3 ± 2.9b 6.8 ± 5.9a 6.8 ± 5.9b 8.8 ± 3.7b 7.2 ± 0.4c 11.0 ± 4.3c 8.1 ± 9.6c 10.4 ± 4.4d PMSF 0a 0b 0a 8.6 ± 4.1b 8.8 ± 3.7b 10.8 ± 0.6c 20.2 ± 5.3c 31.7 ± 7.0b 46.6 ± 0.0c EDTA 1.7 ± 2.9a 1.7 ± 2.9b 1.8 ± 3.0a 3.4 ± 3.0b 5.5 ± 5.3b 3.8 ± 3.3c 9.6 ± 8.9c 12.0 ± 0.8c 22.4 ± 3.3d 1,10-Phenanthroline 1.8 ± 0.0a 5.1 ± 3.3ab 6.8 ± 2.7a 13.6 ± 6.5b 12.3 ± 3.7b 12.6 ± 3.1c 20.0 ± 3.5c 23.6 ± 0.0bc 40.7 ± 6.5c 3.4 ± 3.0a 5.1 ± 3.3b 5.5 ± 5.3b 6.8 ± 5.9b 10.8 ± 0.6b 12.0 ± 0.8c 23.6 ± 3.0b 1,10-Phenanthroline 1.8 ± 0.0a 3.3 ± 2.9a 8.1 ± 7.6a 12.6 ± 3.1ab 12.0 ± 0.8ab 13.6 ± 6.5ab 20.2 ± 5.3ab 22.4 ± 3.3c 43.6 ± 3.3a EDTA; toxicity recovered after ZnSO 4 addition 3.4 ± 3.0a 3.4 ± 3.0a 8.6 ± 4.1a 8.8 ± 3.7b 20.2 ± 5.3ab 23.6 ± 0.0a 27.1 ± 1.8a 35.0 ± 5.7b 44.4 ± 5.3a 1,10-Phenanthroline; toxicity recovered after ZnSO 4 addition 0a 5.1 ± 3.3a 12.0 ± 3.5a 17.5 ± 2.2a 24.0 ± 11.2a 12.6 ± 3.1b 31.7 ± 7.0a 55.0 ± 6.2a 55.0 ± 6.2a Termite-killing components in Serratia marcescens (SM1) mortality rate of O. formosanus treated with the pigment oil layer was very close to that of the controls. Therefore, it was confirmed that the main active component of S. marcescens SM1 fermentation was in the supernatant, and that the pigment oil layer had no insecticidal properties. Since the bacterial precipitate was added to the medium during dissolution, the undestroyed cells might continue to ferment to produce S. marcescens SM1 fermentation and induce a poisoning effect in the subsequent toxicity test experiments. After identifying that the main active component was the supernatant, it was necessary to further clarify whether the active component was a protease by heating or adding different protease inhibitors. From the results, the toxicity effect of the heated supernatant was significantly reduced, and the corrected mortality rate of the treatment was only 30%. But as time increased, its toxicity gradually increased, which may be caused by the recovery of certain denatured proteases. Since protease K can inhibit most of the proteases, after protease inhibitor K was added, the toxicity effect of the supernatant was greatly reduced, and the mortality of treated O. formosanus was almost the same as that of the control. At the same time, the toxicity effects of the supernatants treated with the three types of protease inhibitors, PMSF, EDTA and 1,10-phenanthroline, were also reduced. At 48 h, the corrected mortality rate of each treatment group was less than 32%. These results indicate that there was more than one insecticidal protease in the supernatant and that these insecticidal proteins were one of the important causes of termite death.
Since both protease inhibitors EDTA and 1,10-phenanthroline inhibit metalloprotease, the supernatants treated with these inhibitors were incubated with 0.5 mmol/L ZnSO 4 for 1 h at 22 °C to confirm whether or not the metalloprotease was active. The results show that the insecticidal activity of the supernatant was restored after the ZnSO 4 recovery experiment, which indicated that the metalloprotease was present in the supernatant and the insecticidal protease had a good toxicity effect on O. formosanus. However, this experiment only preliminarily confirmed that the active components of S. marcescens SM1 fermentation were metalloproteases and other proteins. More precise active components are not known and further investigation is necessary.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creat iveco mmons .org/licen ses/by/4.0/.