Cassava Root Necrosis Disease (CRND): A New Crop Disease Spreading in Western Democratic Republic of Congo and in Some Central African Countries

Cassava is consumed in the Democratic Republic of Congo (DRC) as a staple food for the majority of the Congolese population. This crop is used in several forms: as fufu, chikwangue and pondu; cassava leaves are the most consumed vegetable in the country. In 2002, cassava root symptoms similar to cassava brown streak disease (CBSD) were reported for the first time in western DRC. PCR assays, using primers specific to Cassava brown streak virus (CBSV), failed to detect or identify any viral pathogens in diseased cassava samples from western DRC. Therefore, next generation sequencing (NGS) techniques were used as they are able to sequence full organism genomes and are widely used for the identification of pathogens responsible for new diseases. The main objective of this study was to identify the pathogens causing root necrosis in western DRC. WhatmanFTATM cards were used to collect 12 cassava leaf samples from plants with symptoms indicative of very severe root necrosis, as well as two asymptomatic samples. These 12 samples were sent to Australia at the University of Western Australia in Perth for next generation sequencing (NGS) using the Illumina HiSeq platform. Additional bioinformatics tools included Geneious, CLC workbench, ParaKraken and Kaijou software for short DNA sequences. No viruses (including CBSV) were found in any of the DRC samples. These preliminary results confirm all the previous negative results obtained using PCR and CBSV primers. However, NGS analyses did reveal the presence of a number of bacterial and fungal taxa. These will require further investigation and tests such as the Koch Postulates, to establish their specific pathogenic role in cassava. This is the first scientific evidence that no currently known virus is responsible for the disease which had been referred to previously as ‘CBSD-like disease’. Consequently, the disease found in DRC cassava samples has been designated ‘Cassava Root Necrosis Disease’ or CRND.


Introduction
Cassava (Manihot esculenta Crantz, family Euphorbiaceae) produces carbohydrate-rich storage roots, which are a staple food crop for approximately 800 million people worldwide (Food and Agriculture Organization, 2013). In Africa, cassava is the second most important food staple in terms of per capita calories consumed (Nweke, 2004). The RNA pellet was then washed in 700 ml of 70% ethanol and the tubes vortexed briefly before being incubated at -20 °C for at least 10 min. The tubes were then centrifuged for 5 min at 13,000 rpm. The ethanol was then removed and the pellet was air-dried. Finally the dried RNApellet were re-suspended in 100 μl 1XTE/sterilized double distilled H20 on ice for about 30 min and stored at -20 °C before use. (Ndunguru et al., 2015) Total RNA extracts that presented 260/280 and 260/230 purity indices equal to or greater than 2.0 and integral RNA in electrophoresis and Bioanalyzer measurements (RIN > 8) were selected. The cDNA libraries were prepared from 1 μg of total RNA using the IlluminaTruSeq Stranded Total RNA Sample Preparation kit with Ribo-Zero TM Plant according to the manufacturer's instructions (Illumina, San Diego, California). Briefly, after rRNA depletion and RNA fragmentation, first and second strand cDNA was synthesized, adapters were ligated to the 50 and 30 ends of the fragments and the fragments enriched by PCR. cDNA libraries final size and concentration of each library was estimated using a Bioanalyzer (Agilent, Santa Clara, CA, USA) and the Qubit (Invitrogen, Carlsbad, CA, USA), respectively. Ten nM library pools were prepared by mixing the libraries to achieve an equal molar concentration of each. Libraries were normalized, pooled and sequenced using a 2 × 300 cycle PE V3 Illumina kit. Paired end reads were generated using the Illumina MiSeq System at the Biosciences Eastern and Central Africa-International Livestock Research Institute (BECA-ILRI) Hub in Nairobi, Kenya. (Ndunguru et al., 2015) For each sample, reads were first trimmed using CLC Genomics Workbench 6.5 (CLCGW) (CLC Bio) with the quality scores limit set to 0.01, maximum number of ambiguities to two and removing any reads with < 30 nucleotides (nt). Contigs were assembled using the de novo assembly function of CLCGW with automatic word size, automatic bubble size, minimum contig length 500, mismatch cost two, insertion cost three, deletion cost three, length fraction 0.5 and similarity fraction 0.9. Contigs were sorted by length and the longest subjected to a BLAST search (blastn and blastx). In addition, reads were also imported into Geneious 6.1.6 (Biomatters) and provided with reference sequences obtained from Genbank.

Library Preparation and Illumina Sequencing
Total RNA and DNA extractions was carried out in the UWA from FTA samples and were sent to the Australian Genome Research Facility of the UWA for library preparation and sequencing on an Illumina HiSeq 2500.

Sequences Analysis
For each sample, reads were first trimmed using CLC Genomics Workbench 6.5 (CLCGW) (CLC Bio) with the following parameters: quality scores limit set to 0.01, maximum number of ambiguities set to twoand removal of any reads with < 30 nucleotides. Contigs were assembled using the de novo assembly function of CLCGW with automatic word size, automatic bubble size, minimum contig length 500, mismatch cost two, insertion cost three, deletion cost three, length fraction 0.5 and similarity fraction 0.9. Contigs were sorted by length and the longest subjected to a BLAST search (blastn and blastx) (Altschul et al., 1990). In addition, reads were also imported into Geneious 6.1.6 (Drummond et al., 2010) (Biomatters) and provided with reference sequences obtained from Genbank (NC012698 for CBSV, GQ329864 for CBSV-T and NC014791 for UCBSV). These methods have been used previously for the successful recovery of whole CBSV and UCBSV genome sequences (Ndunguru et al., 2015;Alicai et al., 2016;Ateka et al., 2017).
Mapping was performed using Kaiju software with minimum overlap 10%, minimum overlap identity 80%, allow gaps 10% and fine tuning set to iterate up to 10 times.
While recent taxonomic classification programs achieve high speed by comparing genomic k-mers, they often lack sensitivity for overcoming evolutionary divergence; these results in large fractions of the metagenomic reads remaining unclassified. Kaiju is a novel metagenome classifier, which finds maximum (in-) exact matches on the protein level using the Burrows-Wheeler transform (Menzel et al., 2016).
It has been shown that that Kaiju classifies reads with higher sensitivity and similar precision compared with current k-mer-based classifiers, especially in genera that are under-represented in reference databases. It has also been demonstrated that Kaiju classifies up to 10 times more reads in real metagenomes. Kaiju can also process millions of reads per minute and can run on a standard PC (Menzel et al., 2016).

Preliminary Results and Discussion
After trimming and assembling NGS data outputs using CLC workbench and Geneious software, sequences were processed using the Kaiju and outputs are presented in Figures 4 and 5   The figure 5 shows that viral sequences were quantified at 0.5%.
The list of all microorganisms identified in all 12 samples and those suspected to play a pathogenic role in plant diseases according to the literature are presented in Tables 2 and 3 below.   (Dreaden et al., 2011).
More researches are currently ongoing and each suspected microorganisms above needs to be confirmed by the Koch Postulates assays as causative pathogen(s) of CRND in western DRC.
Isolations of bacteria and fungi are currently ongoing with the partnership of the Plant Clinic of Kinshasa. Microorganisms that will be isolated from cassava roots necrotic tissues will be genetically characterized and sequenced.
Koch Postulates trials will be done with the involvement of the DSMZ (Deutsche Sammlung von Mikroorganismen und Zellkulturen) in Germany.
It is possible that the CRND root necrosis disease could be caused by the action of a bacterium-fungus complex.
The disease could be initiated by an initial attack of bacteria and root necrotic symptoms externalized by a secondary attack of fungi. Further studies are required to confirm or refute this hypothesis.

Conclusion and Perspectives
This study points to the apparent absence of CBSV in western region of DRC and suggests that CRND could be caused by other microorganisms such as bacteria, fungi or a combination of both. There appear to be two distinct diseases, namely CRND and CBSD which have similar root symptoms but different stem and foliar symptoms.
Since 2004, CBSD has been spreading from East Africa to Central Africa and was confirmed in 2012 in eastern DRC; it is expected to spread to western DRC and on to West Africa. At the same time, CRND is spreading from western DRC towards West Africa and eastern DRC.
If no control measures (quarantine, etc.) are put in place, there is a strong possibility that both diseases will spread to West Africa. Should this event cause cases of infections of both diseases, the results are likely to mean devastating cassava root crop losses and significant economic impacts on farmers' livelihoods. Ultimately, this has serious implications for food security in Central Africa.
We consider that further research on CRND pathogens identification is paramount. Koch's Postulates on isolated microorganisms from diseased plants and other biological assays will help to elucidate the causal pathogens of this disease. Information on disease etiology will allow for future disease epidemiology and genetic disease resistance research.