Streptomyces spp. were isolated following the protocol described by [15] from three distinct sources: the traditional starter Sta_96 produced in the laboratory according to [9], farm soils, and a cassava peel dump soil. For soil samples collection, 100 g of soil was collected from three different farms located in Yaoundé, Cameroon (3.848° N and 11.502° E). Soil samples were taken at a depth of 10–15 cm below the surface using sterile trowels to minimize contamination and ensure consistency. Sampling was conducted during the rainy season (April–May), when microbial activity is typically highest due to increased moisture and nutrient availability. At the time of collection, environmental conditions were recorded: average temperature ranged from 24–28 °C, relative humidity was between 75–85%, and the soil was moderately moist with visible organic matter. Samples were immediately placed in sterile plastic bags, labeled, and transported to the laboratory for processing within 24 hours to preserve microbial viability. The cassava peel dump sample was collected from a site with visible microbial colonization and was processed similarly to the soil samples. Once in the laboratory, samples were mixed with calcium carbonate (ACS reagent grade, $\ge$99.0% purity- Sigma-Aldrich) at a ratio of 1/10 (g/g) and kept in a dry, ventilated location for 48 h. Subsequently, 25 g of each sample was transferred into 225 mL of sterile physiological saline, vortexed for 5 min, and allowed to settle for an additional 5 min. The supernatants were serially diluted, and 0.1 mL of each dilution was spread onto Inorganic Salt Starch Agar (ISP4) previously poured into Petri dishes. Plates were incubated at 28 °C $\pm$ 2 for 7 days. After incubation, colonies exhibiting typical Streptomyces morphology (round and powdery, slightly embedded in agar, Gram-positive filamentous cells, catalase-positive) were presumptively identified as Streptomyces spp. These colonies underwent five successive subcultures to ensure purity, and culture purity was verified by Gram staining followed by microscopic examination, by confirming consistent colony morphology on solid medium, and by the absence of growth on non-selective contamination control plates. Following purity confirmation, biochemical characterization—including salt tolerance and sugar degradation—was systematically performed on all isolates (results for selected strains are presented in Supplementary Table 2). After this, these strains were then preserved in glycerol stocks (20%, v/v) at –80 °C and maintained on nutrient agar slants at 4 °C for short-term storage.

Cellulolytic and pectinolytic activities were assessed as described by [16] using mineral media supplemented with either Sodium-carboxymethyl cellulose (CMC 10 g/L, Sigma Aldrich CAS Number: 9004-32-4, Saint-Quentin-Fallavier, Isère, France) or pectin (10 g/L, Sigma Aldrich CAS Number: 9000-69-5, Saint-Quentin-Fallavier, Isère, France) as the sole carbon source. Isolates were first grown on ISP4 medium to obtain a confluent microbial lawn. Cylindrical plugs (6 mm diameter) were aseptically removed and placed onto CMC agar or pectin agar. Plates were incubated at 37 °C for 96 h, then flooded with 1% (w/v) potassium iodide solution and stained for 5 minutes. Clear halos surrounding the inoculum discs indicated extracellular cellulase or pectinase production. Hydrolysis halo diameters were measured using a caliper following the standard method described by [16]. All enzymatic activity tests were performed in triplicate for each isolate to ensure reproducibility. Only isolates exhibiting at least one of the enzymatic activities were retained for subsequent experiments. It is important to emphasize that, this study evaluates whole microbial retting performance rather than purified enzyme kinetics, reflecting real fermentation conditions where enzymatic cocktails, microbial succession, and matrix-substrate interactions determine efficacy.

Cassava roots from a 9-month-old bitter local variety (“Six-mois”) [17] selected for its prevalence in Cameroonian fermentation processes, highly cyanogenic glycoside content (400~655 mg HCN/kg fresh weight), and high starch content (70–74% dry matter basis), were peeled, washed, and cut into 1-cm-thick cylinders. Batches of 100 g were inoculated with Streptomyces isolates at 10⁶ CFU/g of root tissue, submerged in 100 mL of tap water, and incubated for natural fermentation. A non-inoculated batch was included as a negative control. Retting progression was monitored over 4 days. Every 3 h, six randomly selected cylinders from each batch were tested for firmness using a penetrometer. The softening level was expressed as the mean penetration depth (six measurements per cylinder) [14]. The penetration measurements were performed using the Penetrometer RPN10 Berlin model; the pressure was applied at a controlled rate of 10 mm/s to insure reproducibility and accuracy of the measurements. Penetration data obtained from each time point were compiled in Microsoft Excel 2016 and fitted using DM-fit software (https://browser.combase.cc/DMFit.aspx). The resulting sigmoid kinetics were modeled according to the equation described by [18], thereby, enabling calculation of the retting time as follows:

$$D_r=\mathrm{Lag}+\left(\frac{L}{\text{µ}}\right)$$

Here, D_r, Lag, L and µ are respectively the retting time (h), the time required to initiate fermentation (h), the critical penetrometry index (cm/h) and the softening speed (cm/h).

The percentage reduction of the retting time of each strain was calculated as follows:

$$\text{Percentage of retting time reduction (\%)}={\displaystyle \frac{\text{Retting time of strain}}{\text{Retting time of control}}}\times 100$$

Isolates demonstrating the shortest retting times were selected for further testing. They were incorporated into cassava roots at a concentration of 10⁶ CFU/g either individually or in combination with the traditional Sta_96 starter. The effectiveness of the enriched starters in accelerating retting was evaluated by comparing their retting duration with that of the non-enriched Sta_96 starter. Cassava roots were peeled, washed with tap water, and cut into 1-cm-thick cylinders. Batches of 100 g were prepared, and each selected isolate was mixed with Sta_96 starter at a load corresponding to 10⁶ CFU/g of starter. The mixtures were then applied to each batch, except for the controls (spontaneous fermentation and Sta_96-only control), and submerged in 100 mL of tap water. Where relevant, isolates showing retting-enhancing properties were also tested in combination with the best-performing isolate to further improve the efficiency of the Sta_96 starter. All treatments were incubated under laboratory conditions (28 °C $\pm$ 2), and retting progression was monitored as described in Section 2.3.

The effect of Streptomyces enrichment on the detoxification activity of Sta_96 was assessed by quantifying the total cyanide content of the retted cassava paste using the method of [11]. Briefly, 0.1 g of fermented cassava paste was placed into screw-cap vials containing 1 mL of distilled water and a strip of picrate paper. Vials were sealed and incubated at 30 °C for 18 h to allow cyanide interaction with the picrate paper. Thereafter, the picrate papers were transferred into test tubes containing 5 mL of distilled water and boiled for 5 min. Following boiling, the strips were removed and the solution was allowed to cool to room temperature (28 °C $\pm$ 3) before further analysis. The absorbance of the cooled solution was measured at 510 nm using a UV-Vis spectrophotometer Jenway 7305 model (Fisher Scientific SAS, Illkirch Cedex, France). The blank control consisted of the boiled solution from a picrate paper strip that had not been exposed to cyanide to ensure accurate baseline correction. Total cyanide content (ppm) was calculated by multiplying the absorbance by 396, as described by [19]. Cyanide reduction efficiency was expressed as the ratio between the cyanide concentration in retted paste and that of fresh roots.

All statistical analyses were performed using the IBM SPSS Statistics version 26.0 (IBM Corp., Chicago, IL, USA). The Duncan range test was applied to compare differences in pectinolytic and cellulolytic activities among Streptomyces isolates, retting times of Sta_96–Streptomyces combinations, and cyanide detoxification among treatments. Significant differences are represented by distinct letters in figures and tables. All data are expressed as arithmetic means $\pm$ standard deviations (Mean $\pm$ SD).

To investigate the phenotypic characterization of all 33 Streptomyces isolates, the macroscopic and microscopic characteristics of strains from different sources were investigated. The summary reported in Table 1 shows that, among isolates originating from the starter Sta_96, 66.66% produced orange to brown pigments. In contrast, 58.82% of isolates from cassava field soils exhibited pigment production, predominantly pink. Additionally, 53.84% of isolates obtained from cassava peel soil produced pigments ranging from orange to brown.

Table 2 summarizes the pectinolytic and cellulolytic activities of the strains. A total of 24 out of 33 (72.7%) demonstrated pectinolytic activity, indicated by clear zones of pectin hydrolysis varying from 11.0 $\pm$ 1.4 mm (isolate CCS 3 from cultivated soil) to 42.5 $\pm$ 3.5 mm (isolate CDS 2 from cassava peel soil). Among these 24 pectinolytic isolates, 16 (66.7%) also produced cellulase, with hydrolysis halos ranging from 16.0 $\pm$ 0.1 mm (CDS 1) to 39.0 $\pm$ 0.1 mm (CCS 4). CCS 4 was the only cellulolytic isolate that lacked detectable pectinolytic activity.

The ability of the isolates to soften cassava roots presented in Fig. 1, was compared with that of the control, using rating time as evaluating evaluation metric. Retting time, ranged from 38.6 $\pm$ 0.4 h to 82.0 $\pm$ 10.4 h for STF 16 and CDS 1, as shown in Table 3 and Fig. 2, respectively. The 24 isolates can thus be classified into the following two main categories: isolates with the ability to soften cassava roots (i.e., category 1) and those without (i.e., category 2). Further, category 1, could subsequently subdivided into fast, moderately fast and slow, and respectively composed of isolates with the ability to reduce retting time by more than 50% (STF 16, CCS 4, CCS 20 and CDS 24); between 35 and 45% (CDS 31, CDS 15, CCS 21, CCS 6, STF 33 and CDS 2) and less than 30% (CDS 32, CCS 22 and CDS 12). Additionally, the duration of retting time under CDS 1 isolate did not differ significantly (p 0.05) from that of the control (natural retting). Category 2 on the other hand includes isolates that do not have the capacity to soften cassava roots.

Fig. 1.

Cassava roots: (left) non-retted cassava root and (right) retted-softened.

Fig. 2.

Kinetics of cassava roots softening by: starter isolates (A), soil from a cassava field (B) and soil from cassava peels (C) Control defined as sterile water-inoculated group.

Retting trials using Sta_96 enriched with the five most effective isolates (STF 16, CCS 4, CCS 20, CDS 24, and CDS 31), either individually or in combination, are presented in Table 4. Enrichment significantly reduced retting times (p 0.05), from 44.5 $\pm$ 0.6 h (Sta_96 alone) to 24.3 $\pm$ 0.4 h (Sta_96 + STF 16), corresponding to reductions of 45.5%, 40.5%, 39.6%, 38.1%, and 31.9% for STF 16, CDS 31, CCS 4, CCS 20, and CDS 24, respectively. STF 16 was the most efficient strain in enhancing the activity of Sta_96. When combined with other isolates, a reduction in its performance was observed, except with CCS 20, which did not significantly alter STF 16 retting efficiency (p 0.05).

A study of the detoxification efficiency of cassava roots retted with the best isolate yielded the data shown in Fig. 3. Fresh cassava roots contained 471.5 $\pm$ 7.5 mg/kg of total cyanide. Following spontaneous fermentation, cyanide levels decreased by 96.64%, reaching 15.4 $\pm$ 2.5 mg/kg in the retted paste. When Sta_96 was added, either alone or enriched with STF 16, cyanide was reduced to non-detectable levels.

Fig. 3.

Residual total cyanide content of fermented cassava. (FS) in the absence of the traditional starter Sta_96; (F_Sta_96) in the presence of the traditional starter Sta_96; (F_Sta_96_STF16) in the presence of Sta_96 with added Streptomyces. Sample size for each group is n = 3. Histograms represent the mean $\pm$ standard deviation. Averages followed by different letters (a, b, c) are significantly different at the p 0.05 level. The detection limit for undetectable levels is 0.4 mg/kg.