Ping Zhang · Qingquan Zhao · Xiaoqian Ma ·Ling Ma

Abstract The hazelnut weevil (Curculio dieckmanni Faust.) is a major pest of Asian hazel (Corylus heterophylla Fisch.) in China. Dead hazelnut weevil larvae were examined and the associated pathogenic bacterium was identified as Serratia marcescens Bizio. This significantly shortened the lifespan of hazelnut weevil. Larval weight was reduced as a function of S. marcescens concentration and exposure time. The structure of infected midgut cells was altered,with necrosis of the wall tissues and many cells becoming dislodged, creating cavities. The S. marcencens strain inhibited digestive enzyme activity and protective enzymes in the midgut of adult hazelnut weevil. Inhibition on S. marcencens strain increased with treatment time. S. marcescens directly destroyed the midgut cells and interfered with digestive and protective enzymes. This decreased the food intake and increased mortality of hazelnut weevil. S. marcescens appears to be an effective bacterium for the control of hazelnut weevil but requires further study, including biological formulation development and field application.

Keywords Biocontrol · Curculio dieckmanni · Corylus heterophylla · Pathogenic mechanism · Serratia marcescens

Introduction

Asian hazel (Corylus heterophylla) produces edible nuts containing unsaturated fatty acids that are benef icial to human health (Liu et al. 2018). The species is widely cultivated in northeastern China (Cheng et al. 2015a, 2019). As planting areas have expanded, pest problems have increased.The hazelnut weevil (Curculio dieckmanni) is the primary pest causing nut damage in China. Adults feed on leaves and shoots (Ak?a and Tuncer 2005; Cheng et al. 2015b; Cheng et al. 2016). After mating, eggs are laid on the young nuts;the larvae feed on the kernels of the developing nuts and are difficult to control (Ak?a and Tuncer 2005; Batallacarrera et al. 2016).

Several insecticides have been evaluated for the potential control of hazelnut weevils. The plant-derived pesticide azadirachtin was not acutely toxic to Curculio nucum Col.after 72 h. Beta-cypermethrin and carbosulfan were the most toxic, with lethal concentrations (LC50) of 3.1 mg AI/L and 12.4 mg AI/L, respectively, followed by furathiocarb, methiocarb, and benfurocarb insecticides, The lowest activity was observed for carbaryl (Tuncer et al. 2007). The greatest yield losses of Asian hazel, compared to other pests, are caused by hazelnut weevil. Insecticides are frequently applied for control and can lead to excessive chemical residues on the nuts (Burger et al. 2019).

Hazelnut weevil larvae at different altitudes vary in their susceptibility to the fungus Beauveria bassiana, while Metarhizium anisopliae pathogenicity was unaffected by altitude (Batallacarrera et al. 2013b). Entomopathogenic nematodes can effectively control hazelnut weevil larvae.Within nine weeks of treatment, pathogenic nematodes could be detected at a soil depth of 40 cm, and long-term control was achieved (Batallacarrera et al. 2013a, 2014;Christine 2015). Natural biological agents and their preferred environmental conditions should be considered in control efforts. For example, entomopathogenic nematodes are highly dependent on soil type, moisture, temperature,and chemistry (Stuart et al. 2015; Singh et al. 2016). The use of novel biocontrol agents in specific areas rather than commercially available agents may enhance control. During pest outbreaks, biocontrol agents may not achieve the desired levels of control and may require use in a combination with pesticides (Burger et al. 2019). There is def initely a need to identify new strains of biological control agents and new modes of application. During the process of collecting damaged fruit, red-colored larvae that had died in the fruit were found but the cause of death was unclear. The objective of this study was to isolate pathogenic bacteria from dead larvae in hazelnuts and to evaluate the toxicity and pathogenic mechanisms of the bacterial isolates against adult hazelnut weevil.

Materials and methods

Insect source and feeding conditions

Hazelnut weevils were captured in a hazel orchard in Diaobingshan city by shaking the trees and placed in culture flasks, five per bottle. Three pieces of fresh hazel leaves were placed in each bottle and the flasks stored at 4 °C. The leaves were replaced every seven days. The longest-lived weevil expired after 60 days.

Isolation and identification of insect pathogenic bacteria

Dead, red-colored hazelnut weevil larvae from hazelnuts were dissected according to Wang et al. (2013). The larvae were immersed in 3% sodium hypochlorite for 30 s, then 70% ethanol for 1 min, and rinsed five times with sterile water. The larvae were incised in the body cavity with sterile scissors and the reddish tissue harvested with a needle and inoculated into Luria Bertani medium. Sterile water was also applied to the medium as a control to assure that the insect surface was completely sterilized. All samples were incubated at 25 °C for three days. Colonies with a reddish color in the medium were subcultured for 10 consecutive generations, with three days per generation until the colonies were uniform and colony morphology was consistent. This isolated strain was named AZW1.

Determination of AZW1 toxicity to adult hazelnut weevil

AZW1 suspensions at concentrations of 0.9 × 108, 1.8 × 108,2.7 × 10 8 , 3.6 × 10 8 , and 4.5 × 10 8 cfu/mL were prepared based on their optical density (OD) 600 as measured, using an ultraviolet spectrophotometer. Fresh C. heterophylla leaf discs were obtained with a 1-cm diameter punch and immersed in the suspension liquid for 5 min; the controls were immersed in sterile water. The leaves were dried at room temperature on filter paper, and 10 discs placed in each box. Tests were carried out in a climate chamber at 26 °C, 60% RH, and a 14:10 (L:D) photoperiod. The surfaces of the insects were sterilized by spraying with 70%ethanol before transfer. Twenty insects were placed in each box, with three replicates per concentration. The number of dead insects was recorded at one, two, three, four and five days after treatment. Dead insects were removed and placed onto a microbalance (Shimadzu AUW220D, Japan),and weighted each day at 3:00 p.m. Weight gain was calculated as: weight gain = daily average weight ? initial average weight. The inhibition rate of insect body mass growth (%)was calculated by (weight gain in the control group ? weight gain in the treatment group) / (weight gain in the control group × 100) as described by Lou et al. (2014). Insects that did not move when probed with a pin were considered dead and were immediately dissected, their intestines removed,frozen in liquid nitrogen, and stored at ? 80 °C.

Determination of midgut enzyme activity in adult hazelnut weevil

Insect guts were transferred to mortars pre-cooled to -20 °C.After adding liquid nitrogen, samples were ground and transferred to sterilized 1.5 mL centrifuge tubes and a buffer added for a ratio of ω (insect gut tissue g): v (0.02 mol L ?1 phosphate buffer) = 1:9. Samples were vortexed for 1 min,centrifuged at 1600 × g and 4 °C for 10 min. A 50 μL portion of the supernatant was transferred into a 5 mL centrifuge tube. The following was carried out: (1) absorbance was measured at 253 nm with an ultraviolet spectrophotometer (Biochrom Ultrospec 2100, London, UK) based on the trypsin-catalyzed hydrolysis of the ester chain of the substrate arginine ethyl ester to test for cell damage caused by excessive digestion of insects; (2) the digestive capacity of the insects was determined because amylase with Ca2+acts as a stabilizing factor and activator to cause a sharp drop in the viscosity of the substrate solution. The absorbance was measured at 660 nm with an ultraviolet spectrophotometer;(3) the absorbance was also measured at 420 nm to assess the ability of insects to hydrolyze oils in the body; this was done because of the hydrolysis of the substrate under the action oflipase; (4) to protect insect body cells from degrees of poisoning, the reaction of decomposing H2O 2 based on catalase was quickly stopped by the addition of ammonium molybdate. The remaining H2O 2 reacted with ammonium molybdate to produce a pale-yellow complex which was then measured at 405 nm; (5) peroxidase enzyme activity was determined by measuring the change in absorbance at 420 nm to assess the degree of damage to insect midgut cells since peroxidase (POD) catalyzes the reaction of hydrogen peroxide; (6) superoxide dismutase (SOD) activity was assayed using the xanthine oxidase method based on the production of O 2?anions, and we assessed the ability of production O 2?anions in the intestines of insects. Enzyme activity was calculated according to the manual (Instruction number 001-1, A007-1, A084-1, 080-1, 016-1, 054-1).

Identification of the pathogenic bacteria

DNA was extracted with a bacterial DNA extraction kit(OMEGA D3370, Guangzhou, China) using DNA as a template. The primer was 27F: 1429R for the polymerase chain reaction (PCR) amplification of strain 16S rDNA. The amplification reaction system consisted of 94 °C for 5 min followed by 30 cycles of 94 °C for 45 s, 55 °C for 45 s,and 72 °C for 1 min, followed by a final round of 72 °C for 10 min. A PCR product reference gel recovery kit (Promega A9281, Madison, WI, USA) was used for 10 g/L agarose gel electrophoresis detection. The samples were sent to Shanghai Biotech for sequencing and then submitted to the National Center for Biotechnology Information (NCBI)database in Maryland, USA for blast homology comparison.A phylogenetic tree was constructed using MEGA 7.0 software (Tamura et al. 2013).

Data analysis

One-way analysis of variance (ANOVA) was conducted using the Origin 2018 (Learning version) statistical software. The significance of each treatment was tested by Duncan’s new complex range method (P< 0.05). A P value of 0.05 indicated a significant difference between the groups.

Results

AZW1 effects to the cumulative mortality of C.dieckmanni

The cumulative mortality rates of hazelnut weevil treated with different concentrations of AZW1 fermentation broth differed significantly (Table 1). At 2rd day, the 0.9, 1.8, 2.7,3.6, and 4.5 cfu/mL (colony-forming unit/mL) mortalities were 7%, 11%, 15%, 15%, and 20%, respectively. By day 5, mortality values were 31%, 35%, 58%, 70%, and 73%,respectively. All differed significantly from the control.These results indicate that different concentrations of AZW1 are toxic to adult hazelnut weevil. After feeding on leaf juice containing AZW1 for one to five days, weight gain of the beetle was reduced compared to the water controls (Fig. 1).The weight gains of adults in groups treated with 1.8 × 10 8 ,2.7 × 10 8 , and 3.6 × 10 8 cfu/mL were significantly lower than those of the controls. As the suspension concentration of AZW1 and treatment time increased, the reduction of beetle body weight was significant. Under the lowest concentration of 0.9 × 10 8 cfu /mL, weight gain rates, compared to the controls, on days 1, 2, 3, 4, and 5 were 87.0%, 81.7%,77.3%, 75.3%, and 65.3%, respectively. At the highest concentration of 4.5 × 10 8 cfu/mL, weight gain rates on days 1,2, 3, 4, and 5 were 69.3%, 67.3%, 46.3%, 38.0%, and 24.0%,respectively.

Table 1 Cumulative mortality of hazelnut weevil inoculated with different concentrations of AZW1 strains

Fig. 1 Weight changes in hazelnut weevils treated with AZW1 at different concentrations. Values are the means of four independent replicates (n = 4) ± standard error (SE). Lowercase letters indicate a sig-nificance difference between treatments. Data are from representative experiments performed at least four times with similar results

Fig. 2 Midgut tissue sections of adult hazelnut weevil infected with AZW1 strain (40 ×). A midgut tissue in the control group; B- F Midgut tissue 1, 2,3, 4, and 5 days after infection with AZW1. PM peritrophic membrane, CC columnar cell,EH empty hole, D midgut cell denuded, CP cytoplasmic projections

AZW1 effect on the midgut cell structure of hazelnut weevil

Figure 2 shows normal midgut columnar cells and the midgut tract visible under a light microscope at 40 × magnification. The serosa surface was smooth and the circular muscle layer was visible. At day 1, the midgut tract in the treatment group showed abnormalities in the columnar cells; cytoplasmic projections appeared at the apex. Some of cells had begun to dissolve, resulting in cavities of different sizes.By day 2, the cytoplasmic cavities had gradually increased;goblet cells were deformed, the peritrophic membrane and ring muscles had disappeared. Between days 3 and 4, the saccular process had increased significantly with tissue inf iltration into the cavity. In addition, the columnar cells were severely deformed, and there were numerous protrusions into the disappearing ring muscle. At day 5, all tissues had a frayed texture. There was a considerable necrosis and a large number of cells had detached into the midgut lumen. After dissecting, the body cavity tissues and organs could not be separated, and there was a pungent hydrogen sulf ide odor.

Enzyme assay of midgut digestive enzymes of adult C.dieckmanni

There was an insignificant decrease in proteinase enzyme activity from day 1 to day 3, being 0.103 ± 0.3 (U/g mass)and 0.076 ± 0.3 (U/g mass), respectively. From day 4 to 5,the enzyme activity dropped rapidly by 0.017 ± 0.3 (U/g mass) and 0.329 ± 0.3 (U/g mass), respectively. (Fig. 3 A). 1 d after treating with an AZW1 concentration of 1.8 × 108 cfu/mL, Amylase enzyme activity increased to 1.067 ± 0.3 (U/g mass), and then it gradually decreased from 2 to 5 d. Enzyme activity in the control group gradually increased from day 1 to day 4 by 0.956 ± 0.3 (U/g mass) and began to decline sharply after day 4 to 0.674 ± 0.3 (U/g mass) (Fig. 3 B). After the 1st day of treatment, lipase enzyme activity decreased by 0.0700 ± 0.3 (U/g mass) and was significantly lower than in the control group. From days 2 to 5, the increase in control group enzyme activity was insignificant (Fig. 3 C).

Midgut protective enzyme assay results of hazelnut weevil

POD enzyme activity in the treatment groups was higher than the controls on day 1, with an increase of 0.954 ± 0.3(U/g mass), the increase in control group enzyme activity was insignificant (Fig. 4 A). One day after treatment, catalase activity in the midgut rose sharply (0.983 ± 0.3 (U/g mass) compared to the controls. Enzyme activity increased significantly in the treatment group within one to two days,increasing by 1.91 ± 0.3 (U/g mass) before reaching a maximum. After day 3, enzyme activity began to decrease, dropping by 3.72 ± 0.3 (U/g mass). The increase in the controls was moderate, although not significant (Fig. 4 B). Superoxide dismutase activity in the controls tended to be stable throughout the five-day period, and the differences were insignificant. SOD enzyme activity in the treatment group showed an upward trend during the first three days, increasing by 2.3 ± 0.3 (U/g mass). There was a decline from day 3 to day 5 by 3.2 ± 0.3 (U/g mass). (Fig. 4 C).

Identification of insect pathogenic bacteria

Fig. 3 Changes in the midgut digestive activity of hazelnut weevil at different times after infection. A proteinase activity; B amylase activity; C lipase activity. Values are the means of four independent replicates (n = 4) ± standard error (SE). Lowercase letters indicate a significance level difference between treatments. Data are from representativ experiments performed at least four times with similar results

Fig. 4 Changes in midgut protective enzyme activity of hazelnut weevil at different times after infection. A peroxidase activity; B catalase activity; C superoxide activity. Values are the means of four independent replicates (n = 4) ± standard error (SE). Lowercase letters indicate a significance difference between treatments. Data are from representative experiments that were performed at least four times with similar results

Table 2 Physiological and biochemical characteristics of strain AZW1

The bacterial strains were identified by morphological, physiological, and biochemical characteristics and by 16 s rRNA gene sequencing. Colonies of AZW1 were rose, round,smooth, and viscous on the surface and edges (Fig. S1).The color gradually deepened as the culture time increased;growth was better on potato dextrose agar than on the Luria Bertani medium. Microscopic examination at 100 × revealed bacteria that were near spherical, short rod-like motile cells of about 0.5 × (0.5-1.0) microns. Cells were Gram-negative,facultatively aerobic, methyl red negative, and odorless. The bacteria can utilize D-mannitol, sorbitol, inositol, sucrose,and citrate as carbon sources, but not D-fructose, cottonseed sugar, D-galactose, D-xylose, arabinose, or cellobiose. Cells were arginine double hydrolysis negative and pyruvate dehydrogenase positive (Table 2). The 16S rDNA sequence of AZW1 was similar to the standard strain of Serratia marcescens DSM 30,121 AJ233431.1, with 98% homology (Fig. 5).

Discussion

The hazelnut weevil, C. dieckmanni, is common in northeastern China and often causes major economic losses. The larvae feed on hazel nuts, and since larval control is diff icult, control efforts focus mainly on adults. In this study,the bacterial strain AZW1 was isolated from the body cavities of field-collected larvae. The bacteria were identified as S. marcescens, and we studied its pathogenicity to adult hazelnut weevil. S. marcescens has been used to control a variety of pest species. Patil et al. (2011) used a S. marcescens mannitol extract to control the yellow fever mosquito,Aedes aegypti L. and a major malaria mosquito, Anopheles stephensi Liston larvae. Aggarwal et al. (2015) investigated the insecticidal activity of S. serrata strain SEN against several instars of the tobacco cutworm, Spodoptera litura Fab.larvae. Nehme et al. (2007) found that injection of S. marcescens into the common fruit fly, Drosophila melanogaster Meigen, caused rapid death. In a number of similar reports,the same bacteria was reported to be pathogenic to a variety of insects. For example, Erwinia spp. is pathogenic to a bark beetle when infected by Beauveria bassiana (Xu et al.2019a, b). Erwinia has also been reported to be a pathogen to aphids. In the present study, the highest mortality rate was 73% by day 5 after feeding on leaves coated with AZW1, and this was significantly greater than the controls. These results indicate that AZW1 has insecticidal activity and conf irms Serratia marcescens as the causative agent of hazelnut weevil larval mortality in the field.

Fig. 5 Neighbor joining (NJ) tree showing the phylogenetic positions of strain AZW1 and other related taxa based on 16S rRNA gene sequences. Bootstrap values (expressed as percentages of 1000 replicates) are shown at branch points. Bar: 0.005 nt substitution rate (Knuc)

Lipase, amylase, and protease are digestive enzymes in insect midguts that metabolize sugars, lipids, cellulose,and proteins; the enzymes are important in energy acquisition and in the metabolism of insect foods (Wigglesworth 1972). Significant reductions of proteases have been found in insects infected with S. marcescens (Hejazi et al. 1997; Bosa and Cotes 2004; Dillon et al. 2005). In this study, protease activity was significantly reduced in adult hazelnut weevil after one day of feeding on S. marcescens. As the treatment days increased, protease activity decreased to less than the controls, indicating that AZW1 had a significant inhibitory effect on midgut proteases. This study also showed that the levels oflipases and amylases fluctuated. Luntz and Blackwell (1993) and Rharrabe et al. (2008) found that azadirachtin exerted antifeedant effects and inhibited the growth and development of the Indian-meal moth, Plodia interpunctella (Hübner), larvae. Treated larvae had reduced body mass, developmental abnormalities, and significant mortality. The inhibitory effect of AZW1 on the weight gain of hazelnut weevil may be related to changes in the digestive enzymes. Protease, lipase, and amylase activities were inhibited, and this subsequently affected adult feeding, leading to reduced weight gain. Higher concentrations of AZW1 led to greater inhibition.

In the body of an insect stressed by exogenous toxic substances, many active oxygen radicals of superoxide anion O 2?and oxygen radical-OH are produced that help to kill pathogens and protect the body from damage. However,excessive oxygen can also cause damage, while the protective enzyme system removes excess free radicals (Jia et al. 2016). In this study, the initial positive responses of superoxide dismutase (SOD) and catalase (CAT) enzymes were inhibited over time. However, the initial response of peroxidase (POD) was signifi cantly inhibited, gradually weakened and then stabilized. In a similar study, Ali et al.(2017) reported that infection of the silverleaf whitefly,Bemisia tabaci (Gennadius), by the fungus Metarhizium muscarium led to decreases in activities of SOD, CAT and POD. In this study, AZW1 treatment of C. dieckmanni resulted in an induction of CAT and POD in the midgut. As the treatment time increased, enzyme activities gradually decreased, which may have been related to the amount of weevil feeding.

After one day of feeding, hazelnut weevil adults lost their peritrophic membranes, normally an effective barrier between the midgut wall and the midgut lumen. The membrane is mainly composed of chitin, sugar, and protein.Chitin is the main component of the peritrophic membrane skeleton (Moskalyk et al. 2010). The extracellular proteins secreted by S. marcescens include chitinase, hemolysin, and lipase (Aucken and Pitt 1998; Mohan et al. 2011; Nakamura et al. 2018), and chitinase can hydrolyze insect midgut peritrophic membranes. Mohan et al. (2011) reported that S. marcescens SRM strains can overcome insect host defenses and rapidly increase throughout the gut and haemocoel or primary body cavity. Tao et al. (2006) isolated a new insecticidal protein-zinc metalloproteinase from the culture solution of S. marcescens HR-3 and found that the enzyme possessed hydrolyzed protein activity. It may be inferred that after hazelnut weevil consumes AZW1, the bacteria multiply in the body, and a variety of enzymes are secreted. The enzymes destroy the protein in the peritrophic membrane, after which bacteria enter the midgut wall cells where the cell gap increases, and a large number of cells are shed. At this point, a large amount of toxin is produced in the insect’s body, and the enzymes scavenge excess free radicals; this explains the increase in enzyme activity. The damage to the midgut cells results in a decrease in secreted digestive enzymes and a compromised absorption function that causes the digestive enzyme activity of the midgut tract to continuously reduce. As a result, the nutrient utilization efficiency, insect feeding rate, and insect body weight are continuously lessened. This reduces growth and development, and often results in mortality. However, AZW1 acts directly on the intestinal tract of C. dieckmanni and interacts with intestinal microbes. This interaction will be the focus of further research.

In summary, a strain (AZW1) of S. marcescens bacteria from the body cavity of hazelnut weevil larvae was isolated and identified, and the pathogenicity and pathogenesis of this strain was determined. AZW1 is a candidate agent for the biocontrol of hazelnut weevil. However, the control effects of AZW1 will need to be evaluated in the field under a variety of environmental conditions.