SSF has lately proven to be a fascinating alternative to SLF and has demonstrated consistency in numerous industries . SSF is very successful in the synthesis of many novel antimicrobial agents, since its process conditions resemble the natural environment of Streptomyces sp. more closely than SLF . Different bacterial species have been reported in literature to produce diverse antimicrobial agents under SSF using different solid substrates [18, 36, 37].
The present study was targeted at the optimization of the nutritional and environmental conditions for paromomycin production in SSF. Choosing a suitable substrate is a crucial aspect of SSF since it represents both a nutrients source and a physical support . Substrate dependent bacterial product yield differences have been shown in previous studies and hence screening of several substrates is necessary . Therefore, six different substrates were screened for paromomycin production by S. rimosus NRRL 2455. As shown in the results, the highest production was obtained using corn bran as the solid substrate. Corn bran is the most abundant and low-value agro-industrial byproduct of the milling process of corn [39, 40]. After corn processing, its bran is generally discarded or used as animal feed . Corn bran is rich in carbohydrates (78%), proteins (3.5%), iron (16%) and fats (1%) . It has been successfully used in SSF of many metabolites including biosurfactants , enzymes  and antibiotics . Using agro-industrial by-products as carbon and energy sources is advantageous for two reasons: the use of a cheap substrate and a fascinating way of adding worth to a by-product . Hence, using this inexpensive agricultural residue will drastically reduce the costs of paromomycin production, and will also tile the way to efficient managing of solid wastes.
Production of antibiotics is significantly improved by the supplementation of numerous carbon and nitrogen sources in fermentation medium . Results from our previous study showed that aminoglycoside production medium (A6), consisting of glycerol and CaCO3 as carbon sources and soybean meal and NH4Cl as nitrogen sources, was the optimum medium for paromomycin production in SLF since it resulted in the highest specific productivity . Therefore, A6 media was selected in this study as the impregnating solution. Similar results were obtained in previous studies, where 1% w/w CaCO3 enhanced the tetracycline production in SSF using S. rimosus . Moreover, NH4Cl positively enhanced neomycin production , while soybean meal enhanced rifamycin B concentrations by Amycolatopsis sp. RSP 3 under SSF . The volume of impregnating solution used depended on the substrate’s liquid absorption capacity, which was different for each substrate. The absorption capacity is defined as the volume that can be added to 10 g of dry substrate without the emergence of free liquid . This was to ensure that optimum moisture levels were used, and to avoid excess moisture which might negatively affect antibiotic production.
After determining the best substrate, it was necessary to perform a kinetic study to investigate the best incubation time required for maximum paromomycin production. A short incubation period may result in incomplete antibiotic formation while excess incubation may cause nutrient exhaustion and accumulation of toxic metabolites which hinders further increase in antibiotic production . As depicted in the results, a maximum concentration of 593.35 µg/ml was obtained after 9 days of incubation. Therefore, extraction in ensuing runs was carried out at this time. Various researches have recorded different optimum incubation periods. Remarkably high levels of neomycin were achieved at the 10th day of fermentation in one study . Maximum rifamycin SV production was obtained on day 9, followed by a steady decline in production . Recently, maximum antibiotics concentration by Streptomyces sp. was achieved after 8 days of incubation .
A time course of paromomycin production under SLF using the impregnating solution as production media (A6 media) and the same cultivation conditions was also carried out to compare the production obtained with that obtained under SSF. As shown in the results, best incubation time was also found to be 9 days, however, a 5-fold improvement in production was observed in case of SSF, which highlights the superiority of this process. Similar results were obtained by Mahalaxmi et al.  who noticed that maximum rifamycin B production was obtained after 9 days of incubation under both SSF and SLF with Amycolatopsis sp. RSP 3, however, values were 4-fold greater in SSF with corn husk. In addition, Tabaraie et al. reported that higher levels of cephalosporin were obtained by A. chrysogenum in SSF than in SLF .
Upon reviewing literature on SSF for antibiotic production, it was found that temperature, pH and inoculum size were some of the factors having a strong impact on SSF . Therefore, these 3 factors were optimized using D-optimal experimental design in RSM. RSM is an efficient technique that can determine the best fermentation conditions for a multi-variable system mathematically and statistically . It is favorable over classical optimization methods because it is fast, reliable, helps understand the effect of varying concentrations of nutrients and leads to a substantial reduction in total number of runs therefore saving time, chemicals and manpower . RSM has been considerably used for the optimized production of many antibiotics [18, 34]. This technique comes with numerous types of designs for the optimization of important fermentation parameters, and D-optimal design is one of the most accurate ones . It has been used by many investigators in optimization studies [51, 52].
A total of 16 runs were conducted to study how the 3 factors influenced paromomycin production. To investigate the significance of the design, we used ANOVA which provides a better understanding about the sources of variation and is the most exploited statistical tool to evaluate the variables impact over a process . The attained F-value (163.16) proved that the model developed in this study was significant and may be used to explain paromomycin production by SSF. Alternatively, the lack of fitness should be non-significant for the model to fit well with the experimental design. In our results, the non-significant lack of fit (P value = 0.1299) showed that the model was appropriate for the current study. The CV indicates the precision level with which the treatments are compared, and model reliability usually declines as the CV value rises . Our low CV indicates adequate reliability of the experimental values. In addition, the obtained R2 demonstrates a close agreement between the experimental and the predicted values. The ability of the model to precisely predict a response value can be expressed as the predicted R2, which should be in good agreement with the adjusted R2, and difference between both shouldn’t exceed 0.2 . In our study, the Pred R2 and Adj R2 were in fair agreement with each other. Adequate (Adeq) Precision assesses the signal to noise ratio, and a ratio greater than 4 is usually desirable . Our Adeq precision ratio of 37.35 implied a satisfactory signal and that the model may be used to navigate the design space and may effectively be used to explain paromomycin production by SSF with S. rimosus.
For a better understanding of the variables’ effects on paromomycin production, the predicted model was presented as 3D plots. The 3D plots can directly reflect the effect of different levels of the factors on the response and therefore pinpoint their optimum levels . Using these plots together with numerical optimization function, optimum conditions for highest production were found to be a pH of 8.5, temperature of 30 °C and inoculum size of 5% v/w, yielding a maximum IZ diameter of 27.25 mm which was nearly equal to the value predicted by the model proving the soundness of the model. The constructed model diagnostic plots further verified the validity of the model built.
P value was also used to evaluate the significance of each of the tested variables. The smaller the P-value and larger the sum of squares, the more significant the corresponding factor is . Results revealed that the factors A (pH) and B (temperature) had a significant effect on paromomycin concentration (P < 0.0001), temperature being more significant. However, factor C (inoculum size) had the least effect on paromomycin production. Similar findings were reported by Mahalaxmi et al.  who showed that inoculum level exhibited the least influence on antibiotic production. Temperature and pH are vital physiological parameters influencing the metabolic pathways, hence the generation of various metabolites . In our study, maximum paromomycin concentration was obtained using pH 8.5. A pH 8.5 was also optimum for the maximum production of rifamycin B in SSF in a previous study . On the other hand, maximum rifamycin SV production was obtained at pH 7 in a previous study . This indicates that the best pH for maximum antibiotic production was strain dependent.
Another critical factor affecting antibiotic production is temperature. As depicted in the results, optimum temperature was found to be 30 °C. This is similar to results obtained in our previous study, where optimal temperature for highest paromomycin production under SLF was found to be 28 °C . Results by other researchers were variable. Some studies showed that 30 °C was ideal for antibiotic production under SSF including neomycin , rifamycin SV  and cephalosporin C . Others reported that maximum rifamycin B production in SSF by A. mediterranei strain MTCC 14 was obtained at 32 °C . In our study, a decrease in paromomycin concentration was detected when the incubation temperature was lower than the optimum temperature. It has been reported by several researchers that low temperatures tend to slow down the metabolic activities of the microorganisms . Moreover, production was completely abolished at a temperature of 37 °C in acidic or neutral conditions. This may be because heat evolved during SSF process is poorly dissipated and therefore gets accumulated in the medium, resulting in decreased microbial activity and growth, thus reducing the product yield .
Moreover, ANOVA results revealed that the model term AB was significant, meaning that the interaction between pH and temperature was significant, while the other interaction terms were insignificant. A significant interaction between 2 factors means the effect of one factor is dependent on the level of the other . Contour plots can reveal the significance of the interaction between two factors: an elliptical contour implies a significant interaction between the two factors, whereas a circular contour implies that the interaction between the two factors is weak . As depicted in the results, the contour plot obtained for AB was oval in shape indicating significant interaction between these variables.
Therefore, optimization of paromomycin production by D-optimal design resulted in a maximum IZ diameter of 27.25 mm which was equivalent to 2.21 mg /g IDS or 1.47 mg/ml A6. Therefore, a 4.3-fold enhancement in production was attained in comparison to production obtained using unoptimized conditions.
It is interesting to note that in our previous study, paromomycin production was optimized under SLF, resulting in a maximum paromomycin concentration of 1.58 mg/ml A6 . Upon comparing these results with results obtained in the current study using SSF, it was found that comparable antibiotic levels were obtained using SSF in nearly the same period of time as SLF (9 days) and nearly the same inoculum size, however, using cheaper substrates and a relatively simpler technique. In addition, Streptomyces mycelium morphology is well-matched to invasive growth on solid culture. This morphology accounts for substantial problems in SLF, including sheer forces, increased viscosity due to the metabolic secretion, and a reduced metabolic stability, which leads to very high mixing requirements and oxygen transfer efficiency in addition to product recovery complications . Therefore, SSF process may be utilized as a substitute, permitting better oxygen circulation, less waste water production and reduced energy requirements for stirring and sterilization making it a more attractive technique for paromomycin production.