PERSISTENCE OF Bacillus thuringiensis AND Bacillus pumilus POTENTIAL BIOLOGICAL CONTROL AGENTS OF THE COFFEE BERRY BORER UNDER FIELD CONDITIONS OF PUERTO RICO

Field trials were conducted in two years to evaluate persistence of B. thuringiensis and Bacillus pumilus under field conditions. Strains of B. thuringiensis (3971) and B. pumilus (4185) were isolated from coffee agro-ecosystem in Puerto Rico. Field trials consisted in plots of nine coffee trees arranged in a randomized complete block design, sprayed with Bt-3971 and Bp-4185 at 10 8 , 10 9 , or 10 10 colony forming units CFU ml -1 or non-inoculated. In year 1, B. thuringiensis and B. pumilus survived up to six months after inoculation.


INTRODUCTION
Coffee is a crop of higher social, economic, and ecological importance of the mountains of Puerto Rico. The coffee zone is located in the central west, including 21 municipalities, with Adjuntas being the leader in coffee production (Rullan, 2006). Coffee production for the 2013-2014 cropping season was 10.31 metric Tons for a crop value of $ 26.3 million (Monroig-Ingles, 2015;Robles and Ferré-Sadurní, 2017). In September 2017, after Hurricane Maria tore through Puerto Rico, 90% of coffee plants were destroyed (Marino-Cardenas et al., 2018), being not only an economic collapse for coffee industry, but also a much wider and more serious social collapse in the mountainous region of the island.
The coffee berry borer (CBB), Hypothenemus hampei (Ferrari) (Coleoptera: Curculionidae), is a major threat to the coffee industry. CBB originated from Africa (Damon, 2000), and in America, it is found from Mexico to Brazil, including some of the countries of the Caribbean Bassin such as Cuba, Jamaica, the Dominican Republic, and Puerto Rico (Barrera and Lopez-Arroyo, 2007). The pest feeds and reproduces inside the coffee fruit, reducing its weight and quality, and in turn causing losses of up to 50% in the harvest if control measures are not carried out (De la Rosa et al., 2005). CBB is an elusive pest to control due to its cryptic nature. The whole life cycle of the insect occurs inside the fruit where the insect feeds and reproduces (Vega et al., 2009).
Bacillus sp. belongs to the Bacillaceae family. It is a rod-shaped, aerobic, gram-positive bacterium with the presence of endospores. The endospores have an ellipsoidal shape and are in the central part of the cell. These endospores are resistance structures that can remain active for long periods (Stahly et al., 1992). In addition to the endospores, the bacterium produces a para-sporal body or crystal proteins. B. thuringiensis crystals have different forms during sporulation and are protein-based (De la Rosa et al., 2005). B. thuringiensis can synthesize some toxic proteins such as δendotoxin called Cry and Cyt. These proteins are toxic when ingested by susceptible insects, which include important agricultural pests (Bravo et al., 2007). The toxins produce a spore-crystal complex that possesses active insecticidal properties for the control of Lepidoptera, Diptera, and Coleoptera . When the crystal is ingested, the Cry protein dissolves in the mesenteron of the insect creating pores in the intestinal membrane, which leads to the end of the intake and eventually dies of septicemia.
Bacillus para-sporal crystals, formed in endospores, are highly commercialized as biocontrol of pests. Bacillus thuringiensis corresponds to 90% of the commerce of biocontrol agents (Kergunteuil et al., 2016). In the other hand, Bacillus pumilus is a Gram +, rod-shaped, spore-forming bacteria, found mainly in soils. B. pumilus has shown effectivity against certain insects (Molina et al., 2010).
The use of Bacillus for biological control is considered as environmentally safe . Bacillus δ-endotoxins are part of the integrated pest management strategy and are an alternative to chemical insecticides for controlling pest. However habitat related factors in the field such as physic-chemical parameters, rainfall, relative humidity or UV-light can influence their ability to exert biological control. B. thuringiensis  and Bacillus pumilus (Bp-4185) were recovered from coffee agro-ecosystem and are well adapted to the environmental conditions of Puerto Rico. In a previous study we determined the mortality of Hypothenemus hampei surpassed 80% with 1.80x10 10 spores mL -1 of Bt-3971, and reached 100% after 96 hours post inoculation with 3.10 x 10 12 spores mL -1 of Bp-4185. Also, large conjugative plasmids of 12 kbp (contain cry genes) were consistently detected in Bt-3971 and Bacillus pumilus Bp-4185, showing both strains to be lethal and antagonistic against CBB (Vazquez et al., 2020).
This research was conducted to establish the survival of two Bacillus strains, B. thuringiensis (Bt-3971) and B. pumilus (Bp-4185) applied to coffee trees under field conditions during two years in Puerto Rico.

MATERIALS AND METHODS
Field experiments were carried out at the Adjuntas Experimental Station, University of Puerto Rico (N18 ° 10.310; W 066 ° 47.612), during the years year 1 and year 2.

Bacterial strains
Inoculants were prepared using two species of Bacillus: B. thuringiensis (Bt-3971) and B.
pumilus . Bt-3971 and Bp-4185 were isolated from coffee agro-ecosystem, and conserved in silica-gel vials in the culture collection of the Department of Crop Protectionlaboratory of bacteriology at the University of Puerto Rico -Mayaguez.
Nephelo flasks were incubated at 28 ºC in an orbital shaker for 24 h, 48 h, and 96 h in Bt-3971 to reach 10 8 , 10 9 and 10 10 CFU ml -1 , respectively, and 9 h, 18 h, and 36 h in Bp-4185 to reach 10 8 , 10 9 and 10 10 CFU ml -1 , respectively. The experimental design consisted in a randomized complete block design (RCBD) with seven treatments and six replicates. The experimental plot had nine trees, and each block was separated by a row of border trees. Measurements of viable counts of total and Bacillus-type bacteria were conducted in three selected trees at each plot. The treatments were Bt-3971 incubated for 96 h (10 10 CFU ml -1 ), 48 h (10 9 CFU ml -1 ), and 24 h (10 8 CFU ml -1 ) and; and Bp-4185 incubated for 36 h (10 10 CFU ml -1 ), 18 h (10 9 CFU ml -1 ), and 9h (10 8 CFU ml -1 ), and non-inoculated trees (check). Bacterial inoculants were mixed with distilled water in polypropylene bottles (Nalgene®) to reach 6 liters of diluted nutrient broth. Inoculations one and two were made with a total of 14 liters of bacterial solution applied by each treatment. The volume of application was 260 ml / tree applied for 14 seconds covering the entire surface of the tree. In the third inoculation of year 2, the amount of bacterial inoculum was increased from to 400 ml / tree for each bacterial treatment (from 14 liters to 16 liters), since the height of the trees increased.
A CO2 backpack sprayer was used to apply the bacterial inocula. The sprayer was set at a constant pressure of 30 PSI. Bacterial inocula were placed in approx.18 liter (5-gallon) stainless steel tanks. Boom sprayer had a lance with a bronze Tee-Jet nozzle of the solid conical type (TG-2). Nine trees per replicate were selected at each experimental plot.

Monitoring and survival of aerobic endospore-forming bacteria
To determine the survival of bacteria in coffee trees under field conditions, collections of bark and fruit tissues were carried out (12 random samples, taken four from each tree) at 30, 60, 120 and 210 days after the first inoculation. Bacteria were isolated from bark tissue of coffee trees the day before the first inoculation from a random experimental unit from each block. A 2.5 cm cut was made on the tree branch one meter above the ground using a sterile scalpel. In June and September, sampling was done taking six green fruits instead of bark tissue. Bark tissues and fruits were placed in plastic bags and transported to the laboratory on ice.
The tissues were cut into pieces of approximately 0.4 cm in diameter and placed in 9 ml tubes with nutrient broth to make dilutions. Serial dilutions were made until 10 -10 . Then, 0.1 ml (100 μl) were transferred to plates with nutrient agar (beef extract, 1 g; yeast extract, 2 g; peptone, 5 g; NaCl 5 g; agar, 15 g) added with 2% cyclohexamide. Counts of bacterial colonies present in the plate after 72 hours were used to calculate the number of colony forming units per ml (CFU ml -1 ).
After the first inoculation, a monitoring of bark tissues was carried out at 14 days to determine the survival of the inoculated bacteria. Two replicates of each bacterial species (B. thuringiensis and B. pumilus) were used. A total of eight samples (one sample per tree) of approximately 2.5 cm was taken at each plot. The method described above was used for dilutions and counting. For each sample, three repetitions were performed, including a laboratory control and an internal control for Bt and Bp. CFU ml -1 were transformed to Log10 to perform the ANOVA and LSD test (α = 0.05). Non-inoculated trees were not used for ANOVA. T-test was used to determine differences between noninoculated and bacteria-inoculated trees/plots.
Once the bacterial colonies emerged, they were purified and classified by size, morphology, configuration, margin, elevation and color. Also, biochemical tests of KOH and differential staining were performed on all white and creamy colonies with the potential to be Bacillus. Bacteria were separated as aerobic endospore-forming bacteria and were contrasted against the total number of bacterial isolates detected using χ 2 chisquared test. Different bacterial strains were isolated and placed in vials with 2.5 ml nutrient agar, and in vials with 1 ml of silica gel to preserve the strains at -20 o C.

RESULTS AND DISCUSSION
Total counts of bacterial isolates recovered from bark and fruits of coffee tissues were not significantly different (P > 0.05; χ 2 ). The time of incubation was a major difference between both isolates. Bt-3971 reached 10 8 , 10 9 and 10 10 CFU ml -1 after incubation for 24 h, 48 h, and 96 h, respectively. Meanwhile, under the same growth conditions, Bp-4185 reached 10 8 , 10 9 and 10 10 CFU ml -1 after 9 h, 18 h, and 36 h of incubation, respectively. This difference in growth rate was constant in the inoculations made in year 1 and year 2 in this study.
A total of 663 bacterial isolates were recovered from coffee trees inoculated with Bt-4185 and Bp-3971. These isolates were separated based on their colony color, Gram reaction, and presence of endospores. Bacillus-type (Gram+, KOH-, spore+) bacteria were 194 out of the 663 total isolates. At the beginning of the experiment, before the first inoculation, bacteria were isolated from bark tissue of coffee trees the day before the first inoculation.
A total of 13 strains were aerobic endospore-forming bacteria. Total counts of bacteria from bark tissue before the first inoculation are summarized in Table 1. 3.00 x10 9 9.48 In March year 1 and year 2, after the first inoculation, Bacillus-type bacteria were recovered from coffee trees; however, in September year 1 (six months after the last inoculation), as well as in September year 2 (after performing the third inoculation), Bacillus-type bacteria was higher and consistently present.
Values per experimental plot are the average of three trees Values represent total bacterial counts in plate by the inverse of the dilution The rates of Bp-4185 sprayed and total recovery were consistent in both years, the higher inoculum concentrations (10 10 spores ml -1 ), the higher bacterial recoveries. The total bacteria recovery for Bp-4185 inoculated with 10 10 spores ml -1 was 4.16x10 13 CFU ml -1 , meanwhile, with 10 8 spores ml -1 the total bacteria recovered was 8.04x10 12 CFU ml -1 . In Bp-3971, the highest recovery was detected in the 10 8 spores ml -1 treatment with 2.94x10 13 CFU ml -1 ; however, this value was in the same range than the treatment 10 10 spores ml -1 with 2.12x 10 13 CFU ml -1 .
The highest bacterial recovery in trees treated with Bt-3971 was in June for both years (year 1 and year 2). June year 1 was exceptionally dry and hot; therefore, bacterial spores which can resist desiccation help Bacillus-type bacteria to strive under these conditions.
Bacillus spores allow the bacterium to preserve and escape from harsh environmental conditions (Nicholson, 2002).
The first and second evaluations were performed in bark tissue, and the third and fourth evaluations were performed in fruit tissue. This differentiation in collected tissue was done because March and April are months of extensive vegetative growth of coffee trees.
Flowering is a phenotypic response to changes in water availability, air temperature and solar brightness, a coffee flower is open after 4 to 5 months of development. In countries such as Puerto Rico with a very marked rainy season, coffee trees tend to bloom and concentrate harvest in a specific time of the year; however, it is not uncommon to find trees with blooms and grains at different points of maturity.
Logarithmic means of the Bacillus-type bacteria of bark and fruit coffee tissues were used to compare the population levels at each sampling time.
Non-inoculated trees were used to establish if the inoculation increases the number of CFU of Bacillus-type bacteria. In June year 1, averages of 3.02x10 9 and 4.65x10 12 CFU ml -1 were detected in non-inoculated and inoculated trees, respectively (P = 0.0632), no significant differences were detected between non-inoculated and inoculated trees in
In year 1, Bt-3971 and Bp-4185 increased in total counts after the first inoculation (March).
However, only in Bp-4185 a significant increase was detected ( Figure 2). After the initial increase, Bacillus populations remained constant. Although B.
thuringiensis had consistently higher counts than B. pumilus, these differences were not statistically significant, except for the first sampling conducted in March year 1 (Figure 2).
In year 1, samplings conducted from March to September determined the survival of Bt-3971 and Bp-4185 with a maximum of 6 months under field conditions in Adjuntas, P.R.
Total counts of Bacillus-type bacteria recovered in year 2 are summarized in Figure 3. Bt-3971 had significantly more recovery rates than Bp-4185 after the first inoculation conducted in March, and at the last monitoring conducted in September. In year 2 Bacillus-type recoveries were consistently higher than in year 1. This difference is because in year 1 two inoculations were conducted, while in year 2 a third inoculation was made in June. It is uncertain the reason why the levels of Bp-4185 were significantly lower in September year 2 after the third inoculation ( Figure 3). This reduction could have been caused by rain registered after inoculation. In the other hand, the levels of Bt-3971 remained constant and were higher in September compared to June.
In year 1 and year 2, throughout the monitoring of Bacillus-type bacteria in coffee trees, we established the survival capacity of the strains of B. thuringiensis Bt-3971 and B.
pumilus Bp-4185. The laboratory tests used in this study allowed us to corroborate the presence of aerobic endospore forming bacteria inoculated in coffee trees. Survival of Bt-3971 and Bp-4185 was determined by a significant increase after the second (year 1) and third inoculation (year 2). Also, according to the means reflected in the monitoring carried out between the months of March to September year 2, it could be inferred that B.
thuringiensis and B. pumilus were present in the experimental plots after six months of inoculation. According to Sánchez and Peña (2000) there is little information about the ecology of Bacillus thuringiensis, in relation to vegetative cells and spores on plant tissues, specifically in the persistence of spores and the stability of protein crystals. This is because the viability of spores and crystals are affected by the germicidal action of visible and UV light. However, Smith and Couche (1991) reported in pine needles that B.
thuringiensis spores can persist for more than seven months in shaded environments.
Based on what was reported by Smith and Couche (1991); it is deduced in our study that both the vegetative cells and spores of B. thuringiensis and B. pumilus persisted under field conditions in coffee trees.

CONCLUSIONS
This study contributes with knowledge about the persistence of two Bacillus isolates, B.
thuringiensis 3971 and B. pumilus 4185, with potential for biocontrol of coffee berry borer.
Thanks to their ability to form endospores, both isolates can persist up to six months under field conditions of Puerto Rico.
The advantage of using isolates of Bacillus that are adapted to field conditions where they strive is reflected in a greater abundance which give them competitive advantage over other microbiota for biological control of pest.