Predatory phytoseiid mites associated with cassava in Kenya, identification key and molecular diagnosis (Acari: Phytoseiidae)

The objective of this study was to carry out a large survey in order to characterize Phytoseiidae species diversity on cassava in Kenya. A total of 29 species from 10 genera were identified in diverse ecological zones in Kenya. The warm-to-hot low midlands of eastern and the warm-humid coastal strip yielded over 70 % of the species identified. The dominant species were Euseius fustis (Pritchard & Baker) and Typhlodromalus aripo De Leon present in 37 and 34 % of samples, respectively. Typhlodromalus aripo was found persistent in coastal, eastern and western regions of Kenya while E. fustis was present in all sampled localities of the country. In addition to morphological diagnosis, molecular sequences for DNA fragments 12S, CytB, COI and ITS, were obtained for T. aripo and E. fustis. Molecular diagnosis has revealed the unexpected presence of Neoseiulus idaeus Denmark & Muma and indicated its misidentification as Neoseiulus onzoi (Zannou, Moraes & Oliveira) in a previous survey carried out in Kenya. Molecular sequences herein obtained will help further diagnosis of phytoseiid species and complement the international reference databases needed to assist molecular identification of Phytoseiidae species.


INTRODUCTION
Predatory mites of the family Phytoseiidae are the major natural enemies of the cassava green mites (CGM): Mononychellus tanajoa Bondar and M. progressivus Doreste (Acari: Tetranychidae) in Africa (Yaninek et al. 1993;Mutisya et al. 2015). The wet humid geographical regions are usually the most suitable for phytoseiid mites in cassava fields (Shipp and van Houten 1997;Walzer et al. 2007), as long dry periods cause high mortality of most phytoseiid species (De Courcy et al. 2004;Kariuki et al. 2005;Zundel et al. 2007). Bakker et al. (1993) reported that saturation deficit tolerance spectrum of phytoseiid species is an important factor for mite survival and breeding during periods of long drought. The phytoseiid mite Typhlodromalus aripo De Leon was introduced to Africa in the 1990's to control the CGM. Following its release, it quickly spread to most warm-humid regions of the continent (Hanna et al. 1998;Kariuki et al. 2002;Zannou et al. 2007a). Since the release of T. aripo in the mid-1990s in Kenya's coastal and western regions, full report on phytoseiid fauna on cassava was lacking in comparison to Malawi and Mozambique (Zannou et al. 2007a). Typhlodromalus aripo prefers inhabiting cassava apex and is reported to co-exist well with other native species present on leaves of the plant with no displacement impact of either species (Gnanvossou et al. 2005;Zannou et al. 2007a;Molo et al. 2016).
The objective of the present study was to explore phytoseiid diversity in different cassava production zones. Further, it offered an opportunity to carry out a molecular diagnosis study of two species common on cassava. Finally, an identification key to species found on cassava in Kenya based on works of Moraes et al. (2001Moraes et al. ( , 2007, Zannou et al. (2006Zannou et al. ( , 2007b and Chant and McMurtry (2006) was provided.

Survey and phytoseiid species identification
A survey was conducted (2011)(2012)(2013) in 166 cassava fields of three geographical zones: the hot-dry low midlands (LM), the cool-wet upper midlands (UM) and the warm-wet coastal lowlands (CL) of Kenya (Table 1, Figure 1). In each field, 15 cassava plants were sampled at random. The phytoseiid specimens were collected by beating the plants with a wooden stick (60 cm long) over a blue plastic board for approximately two minutes. This technique was chosen as it allows collecting mites from both apices  (2011), Moraes et al. (2001Moraes et al. ( , 2007, Chant and McMurtry (2005) using a phase contrast microscope (x 400).

Molecular characterization of phytoseiid mites
Specimens from plant apex in the coastal Kenya at Mtwapa (assumed to belong to T. aripo) and specimens from the middle canopy leaves at Embu (upper eastern region) were collected in 100 % alcohol vials for DNA analyses. Genomic DNA was individually extracted from four mite specimens using a Qiagen DNeasy tissue kit (Qiagen, Hilden, Germany), according to the DNA extraction protocol « Purification of Total DNA from Animal Blood or Cells » (Spin-Column Protocol) adapted for extracting total DNA from mites (Kanouh et al. 2010). After extraction, specimens were retrieved and mounted on slides according to the method developed by Tixier et al. (2010). A nuclear ribosomal gene section including ITS1-5.8S-ITS2 (reported as ITSS) and three mitochondrial markers (COI, 12S rRNA, Cytochrome B mtDNA) were used. For the COI mtDNA, two fragments were amplified, one using the former's primers and one using the primers proposed by Navajas et al. (1996Navajas et al. ( , 2000 and used in Okassa et al. (2009) andTixier et al. (2006). Primers and thermal cycling are those reported in Tixier et al. (2012) and in Okassa et al. (2009). PCR products were sequenced along both strands using Dynamic ET Terminator Cycle Sequencing kit, and puri-fied using ExoSAP-IT (Amersham). Sequences were aligned and analysed with Geneious v3.5.4 (Drummond et al. 2007). The genetic distances were calculated using the Kimura 2-parameter model using Mega 6.0.6® (Tamura et al. 2013).

Diversity of phytoseiid mites
Within the 166 cassava plots sampled, 528 mites belonging to 29 species and 10 genera were identified ( Table 2). The highest number of species belonged to genus Euseius Wainstein (12). The most frequent and numerous species were: E. fustis and T. aripo (present in 37 and 34% of sampled fields, respectively). All species found are reported below with some comments on DNA sequences when obtained. Information on the known distribution of the species is taken from the online (http://www.lea.esalq.usp.br/phytoseiidae/) Phytoseiidae catalogue (Demite et al. 2014).
1 female: (04°08'56"S, 543  (Kreiter et al. 2016), Saudi Arabia, Senegal, Spain, Tanzania, Turkey, USA, Yemen, Zimbabwe. Remarks. This species has been described from Israel and mainly recorded from Mediterranean Basin and Africa. It is a very efficient natural enemy commercialized and released in greenhouses for controlling phytophagous mites and thrips (Buitenhuis et al. 2015). It has been recorded once on cassava in Ghana (Zannou et al. 2007b).
Previous records. Congo, Kenya, Malawi, Mozambique, South Africa. Remarks. This species is only recorded from Africa. Rodrigues (1968) reported it on cassava plants in Mozambique. Zannou et al. (2005) showed that this species was abundant on cassava plants in that country.
Euseius fustis (Pritchard & Baker, 1962 (Table 3). The genetic distances between the specimens of E. fustis are very low (12S rRNA: 0-0.002; CYTB mtDNA: 0-0.002; COI (LCOHCO): 0-0.032; ITSS: 0) and much lower than those previously reported from two species of a same genus (Tixier et al. 2006(Tixier et al. , 2008(Tixier et al. , 2011(Tixier et al. , 2012(Tixier et al. , 2014. Sequences of this species are not reported from Genbank. Remarks. Euseius fustis was collected in 62% of the cassava fields sampled in Kenya. This species was mostly observed in eastern, western and coastal areas covering most agro-ecological zones of upper and low midlands (UMs & LMs) as well as the coastal lowlands (CLs). This species was already reported from different plants in western Kenya including cassava (Moraes and McMurtry 1988;Moraes et al. 1989;El-Banhawy and Kapp 2011). It is only present in Africa and has been reported on cassava also in Benin (Moraes et al. 1989), Malawi and Democratic Republic of Congo (Gutierrez and Bonato 1994;Zannou et al. 2005). The high densities presently observed as well as its high frequency indicate this is the most common phytoseiid on cassava in Kenya, even after the introduction of T. aripo as suggested by Onzo et al. (2014) due to par-tial niche separation among predator species (Onzo et al. 2010). Its ability to control M. tanajoa alone has been shown to be limited (Onzo et al. 2014;Mutisya et al. 2015).
Euseius lokele (Pritchard & Baker, 1962 Vantornhout et al. 2004). This is the first report on cassava in Kenya; it has been already observed on this plant in Benin, Burundi, Malawi and Uganda (Zannou et al. 2005). Several studies report its low efficiency in controlling M. tanajoa because of the lack of spatial persistence in cassava fields (Nwilene and Nachman, 1996).  (Table 3). DNA sequences of this species exist in Genbank but not for the fragment of COI herein sequenced. Remarks. This species was observed in 0.7 % of the cassava fields sampled, with a total of 19 specimens. Based on the morphological details given by El-Banhawy and Knapp (2011) for specimens from Kenya reported as Neoseiulus onzoi Zannou Moraes & Oliveira, we consider that those mites instead belong to N. idaeus. The latter species has been introduced to Africa between 1985 and 1993 to control M. tanajoa (Yaninek et al. 1992(Yaninek et al. , 1993 without promising results (Yaninek et al. 1993;Kariuki et al. 2005). During 1988-1992, N. idaeus was released in coastal, western and eastern Kenya to control CGM, but recovery efforts did not yield any specimens (Yaninek and Herren, 1988;Yaninek et al. 1993). The specimens herein reported have been collected less than 500 meters from where N. idaeus was released in 1992 at KALRO Katumani field plot.  (Moraes et al. 1989), Mozambique (Rodrigues 1968) and Malawi (Zannou et al. 2006). Nwilene and Nachman (1996) studied the reproduction characteristics of this species on M. tanajoa. It was more efficient than I. degenerans, but seems not efficient enough in field conditions (Nwilene and Nachman 1996). Transeius msabahaensis (Moraes & McMurtry, 1989) Specimens observed.

Neoseiulus tekeleius
1 female: (03°15'41"S, 39°32'10"E, Alt. 38m-Chumani Malindi, 16-XI-2011). Previous records. Kenya. Remarks. This species was described from Kenya and it is the third report after its description. It is the first report on cassava.  Table 3). The sequences of T. aripo match with those already in hand by French contributors of the paper were collected at Kiboko: 02°07'52"S, 37°25'57"E and Mtwapa: 03°10'01"S, 40°12'55"E , south east lowlands and coastal region of Kenya respectively. Remarks. Typhlodromalus aripo occurs in western, eastern and coastal regions and was absent in the cool upper eastern midlands (UM). Kariuki et al. (2002Kariuki et al. ( , 2005 has reported its continued occurrence on cassava since its release in Africa in 1990's. Earlier reports have shown this species to be the most efficient in controlling CGM (Kariuki et al. 2005;Mutisya et al. 2014). This species was described from the Neotropical region and has been introduced to Africa for the control of the CGM where it suppresses pest populations to < 2 mites per leaf in most fields in optimum climatic conditions (Mutisya et al. 2015;Molo et al. 2016). In the absence of CGM, this species can survive on alternative preys such as whitefly and mealy bug nymphs (Gnanvossou et al. 2005;Mutisya et al. 2014). 1 female: (01°59'31"S, 37°33'32"E, Alt. 1339m-Mwala Machakos, 31-V-2011). Previous records. Kenya. Remarks. This is the second report of this species from Kenya, the first one being on coffee (El-Banhawy and Knapp 2011).

Phytoseiid species distribution and abundance
The most frequent and numerous species in this study were E. fustis (indigenous species) and T. aripo (introduced species), accounting respectively for 37 and 34 % of the specimens collected. They were present in more than 70% of the fields sampled. Euseius neofustis followed as a numerous species (13 % of the specimens collected). Each of the other species corresponded to less than 4 % of the specimens collected, most of them represented by a single specimen. Most of the species reported in this paper are indigenous to Africa and some of them are being reported first time on cassava in the present work. This study shows the spread of T. aripo and success of its introduction throughout Kenya.

Geographical distribution and impact of climatic conditions
Phytoseiid mites were observed from all climatic areas sampled. The highest number of plots occupied by Phytoseiidae was observed in the CL and LM zones corresponding to the most humid areas. The percentage of plots where phytoseiids were found were highest for the warm-wet low midlands and coastal lands; in the warm-drier coastal lowlands, 28 % of the plots were occupied by Phytoseiidae species. This observation validates the hypothesis that phytoseiids are susceptible to dry conditions. Only two species among the 29 sampled are observed in all regions (Euseius fustis and E. africanus) whereas 19 species were only found in one region, nine in the warm-wet coastal lowlands (CL) and ten in the hot-dry low midlands (LM). Only E. fustis was observed in the cool-wet upper midlands (UM). Typhlodromalus aripo prevailed in the CL (warm-wet coastal lowlands) region whereas E. fustis prevailed in the UM (cool-wet upper midlands) region.