The research, conducted at the ICAR-Central Plantation Crops Research Institute’s Regional Station at Vittal in coastal Karnataka, set out to answer a deceptively simple question: how much of the variation in flowering, fruit set, and cherelle wilt is written into a tree’s genes, and how much is governed by the nutritional status of its leaves? The site could hardly be more typical of cocoa’s expanding frontier in Asia. Vittal sits at low elevation in a warm, humid coastal zone, receiving an average of 3,670 millimeters of rain each year, almost entirely concentrated between June and November, followed by a pronounced dry season from December through May. Temperatures in the study area ranged from a minimum of 21 degrees Celsius to a maximum of 36 degrees, with relative humidity fluctuating between 61 and 94 percent. Twenty cocoa genotypes, including released hybrids, their parental lines, and a set of upper Amazon clones, all grown as mixed crops with arecanut, formed the experimental material.

The measurement protocol was elegantly direct. On each tree, researchers marked a one-meter length of stem just below the jorquette, the point where the branches fork from the trunk, and counted every flower as it opened in that defined area. They then tracked which of those flowers set fruit, calculated a fruit set percentage for each genotype, and counted the resulting cherelles, separating healthy ones from those that wilted. To assess nutrition, the team sampled the second leaf of the last maturing flush on each tree, oven-dried the material to constant weight, and ran a full panel of elemental analyses: nitrogen by micro-Kjeldahl digestion, phosphorus by the vanado-molybdate colorimetric method, potassium by flame photometry, calcium, magnesium, iron, manganese, copper, and zinc by atomic absorption spectrophotometry, and boron by the azomethine-H method. The whole dataset, gathered across three replications of twenty genotypes, was subjected to standard analysis of variance and Pearson’s correlation statistics.

The results revealed a staggering spread in reproductive behavior across the twenty genotypes, even though all of them grew under the same climate, soil, and management. Flower counts on the marked one-meter stem section ranged from just 16 in the hybrid VTLCH 2 to 196 in the clone VTLC 05, a more than twelvefold difference. Fruit set, expressed as the raw number of flowers converting to fruit in the marked area, peaked at 13.7 in VTLC 05, statistically on par with several other high performers including VTLCC 1, VTLC 61, VTLC 185, VTLCH 1, VTLCH 4, and VTLC 151, while the clone VTLC 154 managed only a single fruit set. When fruit set was expressed as a percentage of opened flowers, the picture shifted in an instructive way: the top performer was VTLC 182 at 13.8 percent, a genotype that produced only 31 flowers in the marked area, while VTLC 30A languished at 1.3 percent despite producing 152 flowers.

That inversion between flower abundance and conversion efficiency is one of the most biologically interesting findings of the study. Several genotypes, including VTLC 05, VTLC 61, and VTLC 30A, produced between 152 and 196 flowers on the marked stem length yet achieved fruit set percentages of only 1.3 to 7 percent. The authors connect this pattern to the internal economics of flower production. Because flowering represents a considerable investment of plant resources, trees that pour resources into enormous flower crops may dilute the auxin supply available to each individual flower. Previous work has shown that endogenous auxin concentrations are significantly higher in clones that produce few flowers and lower in prolific flower producers, which would help explain why a modest-flowering genotype like VTLC 182 can outconvert a floral overachiever like VTLC 05 when it comes to actually setting pods.

Cherelle wilt, the second major source of yield loss, proved equally variable. Across the twenty genotypes, the proportion of cherelles lost to wilting ranged from 22.9 percent in the upper Amazon genotype VTLC 148 to a punishing 72.9 percent in the upper Amazon clone VTLC 155. The Nigerian clone VTLCC 1 produced the highest total number of cherelles at 132 and retained 66 healthy ones, while VTLC 150 held onto only 16 healthy cherelles from a modest crop. Notably, the study found a strong positive correlation, with a coefficient of 0.874, between the total number of cherelles a tree produced and the number that wilted. In other words, trees that set more fruit also lost more fruit, a pattern consistent with the long-standing interpretation of cherelle wilt as a self-thinning mechanism by which the tree adjusts its pod load to what its resources can actually support.

This interpretation is reinforced by earlier research the authors cite. Hand-pollinated cocoa trees, which set more pods than naturally pollinated ones, also suffer correspondingly higher levels of cherelle wilt, with the result that both categories of tree end up carrying roughly the same number of mature pods. Peaks of cherelle wilt also coincide with heavy leaf flushing, pointing to intense internal competition for photosynthates and mineral nutrients between the developing fruit and the expanding vegetative tissue. Leaf nutrient content in cocoa is known to decline during minor and major harvest periods, further implicating nutrient partitioning in the wilt phenomenon. Yet when the Indian team correlated cherelle wilting percentage against the full suite of leaf nutrients they measured, no significant association emerged, suggesting that the tree’s internal carbon and hormone economy, rather than leaf mineral levels alone, drives this thinning process.

The nutrient correlation analysis did, however, deliver one clear and actionable signal. Among all the elements tested, only calcium showed a significant positive association with fruit set percentage, with a Pearson correlation coefficient of 0.596. Nitrogen, phosphorus, potassium, magnesium, iron, manganese, zinc, copper, and boron all failed to reach statistical significance in their relationship with fruit set, despite the well-documented roles of phosphorus in stimulating cocoa flowering and boron in auxin metabolism and transport. The boron connection is particularly notable in the broader literature: boron deficiency has been shown to increase wilted cherelles, presumably by disrupting auxin movement within the developing fruit. But in this Indian dataset, under these specific soil and climatic conditions, calcium stood alone as the nutrient whose leaf concentration tracked with successful fruit set.

The significance of calcium fits with what is known about the demands of early fruit development. Calcium is essential for cell wall structure and cell division, and rapidly expanding young tissues such as newly set pods have high calcium requirements. Because calcium moves through the plant primarily in the transpiration stream and is relatively immobile once deposited, maintaining adequate leaf calcium levels may reflect a tree’s overall capacity to supply this structurally critical element to its developing fruit. For growers, the practical implication is straightforward: the study’s authors conclude that an optimum level of calcium should be maintained in the leaf to support better fruit setting in cocoa. In the coconut and arecanut gardens of coastal Karnataka and Kerala, where cocoa is rapidly becoming a favored intercrop, leaf nutrient analysis could become a routine diagnostic tool for identifying orchards where calcium nutrition is limiting pod set.

Beyond the nutrient finding, the study delivers a broader message for cocoa breeding and improvement programs in India and across the humid tropics. The existence of wide genotypic variation in flowering intensity, fruit set efficiency, and susceptibility to cherelle wilt, demonstrated here across twenty genotypes under identical conditions, gives breeders raw material to select for trees that convert flowers to pods efficiently rather than simply producing more blossoms. Genotypes such as VTLC 182, with its modest flower crop and exceptional 13.8 percent fruit set, and VTLC 148, with the lowest cherelle wilt at 22.9 percent, illustrate the kind of reproductive efficiency that future varieties could combine with high bean yield and quality. As global chocolate demand continues to rise and smallholder cocoa systems expand under shade trees across South and Southeast Asia, understanding the genetics and physiology of the flower-to-pod pipeline may prove just as important as fighting pests and diseases. This study shows that the answers lie partly in the tree’s genes and partly, quite literally, in its leaves.

Subject of Research: Genotypic variation in flowering, fruit set, and cherelle wilt and its relationship with leaf nutrient status in cocoa

Article Title: Genotypic variation in flowering, fruit set, and cherelle wilt, and their relationship with leaf nutrient status in cocoa ( Theobroma cacao L.) grown in humid tropics of India

Article References: Bhavishya, Bhat, R., Apshara, S. E., Pushpa, T. N., Prasad, D. S., Nayana, H., Thube, S. H., Pandian, R. T. P., & Ramesh, S. V. (2024). Genotypic variation in flowering, fruit set, and cherelle wilt, and their relationship with leaf nutrient status in cocoa (Theobroma cacao L.) grown in humid tropics of India. Innovations in Agriculture, 7, 1-5. https://doi.org/10.3897/ia.2024.124253

Image Credits: AI Generated

DOI: 10.3897/ia.2024.124253

Keywords: cocoa, Theobroma cacao, fruit set, cherelle wilt, calcium, leaf nutrient status, genotypic variation, flowering, plantation crops, India, plant nutrition, agroforestry

The scale of the investigation reflects the difficulty of working with organisms that are ubiquitous in soil and water. The researchers analyzed 193 clinical samples drawn from five distinct sources: 101 fecal samples, 21 nasal swabs, 13 trans-tracheal washes, 30 tissue specimens, and 28 blood samples. Each sample was processed through two parallel routes, direct polymerase chain reaction and culture-based isolation, a dual strategy designed to capture both organisms present in the sample at the time of collection and those capable of growing under laboratory conditions. Decontamination steps, essential when dealing with samples teeming with faster-growing environmental flora, were carefully applied to give the slow-growing mycobacteria a chance to emerge on culture media. The approach mirrors established diagnostic protocols for mycobacterial recovery from clinical material, where the balance between eliminating contaminants and preserving viable target organisms is notoriously delicate.

Identification proceeded in a tiered molecular fashion. Genus-level confirmation relied on PCR targeting the hsp65 gene, a workhorse marker in mycobacterial taxonomy that encodes a heat shock protein and allows discrimination of mycobacteria from other acid-fast organisms. Once the isolates were confirmed as members of the genus Mycobacterium and distinguished from the Mycobacterium tuberculosis complex, species-specific PCR assays were deployed to pin down exactly which non-tuberculous species had been recovered. This hierarchical approach, from broad genus screening to species-level resolution, is critical because the clinical and epidemiological significance of non-tuberculous mycobacteria varies enormously from one species to the next. In total, the workflow yielded eighteen confirmed non-tuberculous mycobacterial isolates, a modest but workable collection for downstream genetic analysis.

The species picture that emerged was strikingly lopsided. Mycobacterium vaccae accounted for two-thirds of the isolates, representing 66.7 percent of the confirmed collection. This organism, first described decades ago and long studied for its immunomodulatory properties, is a common inhabitant of soil, water, and dust, and its dominance in bovine samples underscores how readily environmental mycobacteria colonize or transit through livestock. The remaining isolates comprised other non-tuberculous species, though M. vaccae clearly set the tone of the collection. For veterinary diagnosticians, this predominance matters: when a bovine sample yields acid-fast organisms or tuberculosis-like lesions, the probability that the culprit is an environmental species rather than Mycobacterium bovis has direct consequences for how the finding is interpreted within national tuberculosis control frameworks.

With the isolates in hand, the team turned to the central question of the study: what resistance-associated genes do these bovine non-tuberculous mycobacteria carry? Four genes were selected for screening, each with a well-characterized role in drug resistance in Mycobacterium tuberculosis. The rpoB gene, encoding the beta subunit of RNA polymerase, is the principal target of rifampicin, and mutations within its resistance-determining region are the classic molecular signature of rifampicin resistance. The katG gene encodes catalase-peroxidase, the enzyme that activates the prodrug isoniazid, and disruptions in katG are strongly associated with high-level isoniazid resistance. The rpsL gene, encoding ribosomal protein S12, is implicated in streptomycin resistance, while gyrA, encoding the A subunit of DNA gyrase, harbors mutations in its quinolone resistance-determining region that confer fluoroquinolone resistance. Screening these four targets across the eighteen isolates provided a snapshot of the genetic resistance landscape in bovine non-tuberculous mycobacteria from the region.

The screening results revealed a clear hierarchy. The rpoB gene was the most frequently detected, present in 88.9 percent of the isolates, and its detection was mainly associated with Mycobacterium vaccae. The katG gene turned up in 16.7 percent of isolates, and the gyrA gene in just 5.6 percent, and notably both of these were found exclusively in M. vaccae. The rpsL gene, by contrast, was not amplified in any isolate, suggesting that streptomycin-associated genetic determinants, at least as targeted by the primers used, are absent from this collection. The pattern indicates that within this bovine sample set, genetic elements associated with resistance cluster heavily in one environmental species, raising questions about whether M. vaccae’s ecological lifestyle, including prolonged exposure to antimicrobial residues in agricultural environments, might shape its genetic repertoire.

Sequencing of selected amplicons added an important layer of nuance. When the researchers examined the hotspots of rpoB and katG, the regions where resistance-conferring mutations most commonly occur in Mycobacterium tuberculosis, they found no mutations at all. This is a meaningful negative result: the mere presence of a gene does not imply the presence of resistance-conferring variants within it. However, the story differed for gyrA. Sequence analysis of the quinolone resistance-determining region of gyrA revealed non-synonymous substitutions, meaning nucleotide changes that alter the encoded amino acid sequence of DNA gyrase. Whether these substitutions affect fluoroquinolone susceptibility in these isolates remains an open question, but their presence in the canonical resistance region of the enzyme is exactly the kind of finding that warrants closer phenotypic scrutiny.

The authors are careful, and rightly so, about what these results do and do not demonstrate. The detection of resistance-associated genes and even of amino acid substitutions reflects genotypic variation only. Because drug susceptibility testing was not performed on the isolates, the study cannot confirm phenotypic resistance to rifampicin, isoniazid, streptomycin, or fluoroquinolones. This distinction is fundamental in mycobacteriology. Genes such as rpoB and katG are essential housekeeping and metabolic genes present in virtually all mycobacteria; their detection by PCR says nothing about function unless the specific mutational hotspots are altered. Moreover, resistance mechanisms in non-tuberculous mycobacteria differ in important ways from those in the tuberculosis complex, involving intrinsic factors such as impermeable cell walls, efflux pumps, and species-specific enzymatic machinery, which means that extrapolating interpretive criteria from M. tuberculosis to environmental species is fraught with uncertainty.

Nevertheless, the study carries practical implications for animal health and public health alike. Non-tuberculous mycobacteria can produce tuberculosis-like lesions in bovines, and such lesions can trigger false-positive reactions in tuberculin skin testing, the backbone of bovine tuberculosis surveillance programs. Every misattributed lesion represents a potential misdiagnosis, with consequences ranging from unnecessary slaughter to missed transmission of true Mycobacterium bovis infection. By documenting which non-tuberculous species circulate in Indian cattle and buffaloes, and by beginning to map their resistance gene content, the researchers are building the species-level baseline that surveillance programs need to distinguish environmental noise from genuine tuberculosis signals. The finding that M. vaccae dominates the isolates gives diagnosticians a concrete expectation about what they are most likely to encounter when non-tuberculous organisms appear in bovine material.

The road forward, as the authors indicate, runs through phenotypic and species-specific analyses. Establishing the functional significance of the gyrA substitutions observed in M. vaccae will require correlating genotypes with measured minimum inhibitory concentrations of fluoroquinolones, and ideally with functional studies of the mutant gyrase enzymes. Broader panels of resistance genes, and ultimately whole-genome sequencing, could reveal the full complement of intrinsic and acquired resistance determinants in these organisms, an approach increasingly advocated in the literature on non-tuberculous mycobacterial susceptibility prediction. For now, the study stands as a careful first census: eighteen isolates, one dominant species, one frequently detected resistance-associated gene without hotspot mutations, and a handful of amino acid changes in a quinolone target that demand follow-up. In a world where antimicrobial resistance knows no boundaries between human clinics, farms, and soil, even the humblest environmental mycobacterium from a cow’s gut deserves that kind of attention.

Subject of Research: Molecular identification of non-tuberculous mycobacteria from bovine clinical samples and profiling of resistance-associated genes

Article Title: Molecular identification and resistance-associated gene profiling of non-tuberculous mycobacteria isolated from bovine clinical samples

Article References: Panigrahi, B., Narang, D., Chandra, M., Narang, A., Gupta, K., & Singh, S. T. (2026). Molecular identification and resistance-associated gene profiling of non-tuberculous mycobacteria isolated from bovine clinical samples. Molecular Biology Reports, 53(1), Article 1640. https://doi.org/10.1007/s11033-026-12814-7

Image Credits: AI Generated

DOI: 10.1007/s11033-026-12814-7

Keywords: non-tuberculous mycobacteria, Mycobacterium vaccae, bovine tuberculosis, rpoB, katG, gyrA, antimicrobial resistance, PCR, cattle, diagnostics, genotypic variation, India