The molecular pathways of biomass conversion by termites have long intrigued researchers of both basic and applied biosciences. Here, we described the first global study of the enzymatic repertoire involved in plant polysaccharide degradation by the lower termite, C. gestroi, a main urban pest occurring in Brazil. The goal of our work was to expand the knowledge on plant biomass degradation by termites through culture-independent approaches, based on biochemical and global proteome characterization of termite cell extracts [7, 29].
The study herein shows a comprehensive functional characterization of the crude body extracts from C. gestroi. We challenged the ability of the crude termite extract to degrade a wide collection of substrates, including natural polysaccharides and pNP derivatives. The C. gestroi digestome managed to break down all kinds of natural polysaccharides, highlighting the biotechnological potential for the discovery of new enzymatic components from termites [11, 12, 14, 30, 31]. Our data, evidencing that cell extracts from C. gestroi contained all enzymatic activities assayed, such as CMCase, β-glucosidases, endoglucanases, xylanases, manananases, pectinases and amylases, is in agreement with previous reports for various higher and lower termites [32–39].
The capacity of the termite to hydrolyze such a variety of glycoside bonds is a consequence of endogenous enzymes and lignocellulolytic microbiota living in the termite gut. Cellulose and xylan utilization has been described before in the digestive track of a number of lower and higher termite species [25, 33, 36, 39, 40]. Likewise, a number of cellulose and xylan degrading enzymes were previously purified and/or functionally characterized from termites, such as endogenous endoglucanases, classified as GH9, from C. formosanus, R. speratus and Nasutitermes spp. [7, 23, 38, 41], β-glucosidases (GH1) from R. flavipes  and Neotermes koshunensis [43, 44], an endo-β-glucanase (GH7) from a symbiotic protozoa in the hindgut of Coptotermes sp [45, 46] and an endo-β-xylanase (GH 11) from C. formosanus  and from Nasutitermes spp. .
Meanwhile, there is a paucity of studies describing enzymatic hydrolysis of mannan polysaccharides by termites or symbiotic microbiota . The detection of enzymatic hydrolysis of polysaccharides, such as laminarin, lichenan, xyloglucan, arabinogalactan, arabinan and pectin, by crude termite extracts, to the best of our knowledge, have not yet been reported. The detection of these enzymatic activities should be expected, based on the previous isolation of hemicellulolytic-degrading bacteria and yeast from the termite gut, which has already allowed the identification of enzymatic activities such as endo-xylanase, β-xylosidase, α-L-arabinofuranosidase, β-galactanase and β-galactosidases .
The capability to hydrolyze pNP-G and pNP-C indicated the occurrence of β-glucosidases and cellobiosidases, which were previously described in the genera Nasutitermes and Neotermes [36, 43, 44]. Enzymatic degradation of pNP-X and pNP-A, which were described for gut bacteria and yeast, have not been observed in Neotermes jouteli cell extracts , corroborating the data in the current study. However, β-xylosidase activity was identified in the midgut and hindgut portions of R. speratus . It is important to underline that there is a correlation between termite diet and the intestinal microbial population suitable for digestion. Therefore, it is expected that some enzymatic components may appear induced or repressed, which could explain some of the variation in enzymatic activities observed in our study .
The great efficiency of plant biomass degradation by termites is a consequence of the enzymatic action occurring throughout the termite gut, carried out by enzymes both endogenous and derived from symbiotic microbiota. The results for enzymatic assays carried out in different pHs drew our attention to a slight variation in enzyme extract effectiveness. Except for the arabinoxylan and synthetic substrate activity assays, enzyme optimal activity range was between pH 4 and pH 8. As described by others, a strong variation in pH occurs between gut compartments, consistent with the hydrogen potential profile of the termite gut , varying from very acidic (pH 3) to very basic (pH 12). However, the average pH in the gut is around pH 5 to pH 6, which correlated to the optimal results described herein, as has also been described by previous reports .
The proteomic strategy presented herein was successful, based on the number of glycoside hydrolase family members identified through our efforts. We identified 55 different GH enzymes and two CBM proteins, within 29 different CAZy families (Figure 2). Our data are similar to a previous report from N. corniger, which identified 48 proteins within 22 CAZy and Pfams , whereas Warnecke et al.  reported 13 CAZY proteins. Using a high confidence criterion, we identified protein matches for GH3, GH7 (two matches), GH9 and two CBMs from families 6 and 27. Previous proteomic efforts from Burnum et al.  retrieved members of six CAZy families (eight different protein targets), that passed filtering criterion of more than two unique peptide identifications. Several protein hits for polysaccharidases were found in our proteomic study with low significance, which may be justified by enzymes that were insoluble and/or remained firmly bound to lignocellulose or the unavailability of a representative genome, coupled by dealing with a microenvironment where proteins may be under-represented .
The proteomic study presented herein increased the range of known enzymes present in the lower termite digestome. We have found protein identifications for cellulose degrading enzymes (GH1, GH3, GH5, GH7 and GH16), xylan (GH10, GH11 and GH43) and mannan (GH2 and GH38) degrading components, pectinases (GH28 and GH29) and β-galactosidases (GH42), as well as amylases (GH13, GH31 and GH57) and chitinases (GH18 and GH85). Fewer than half these components were previously reported in the proteome of N. corniger; the exceptions being GH2, GH7, GH10, GH11, GH16, GH18, GH28, GH29, GH31, GH38 and GH57 [14, 17]. Metagenome sequencing of the hindgut of N. corniger have confirmed the coding genes for all protein hits identified by our study, but not for GH7, GH10 and GH29, which are virtually absent in higher termite metagenomes [14, 17]. On the other hand, confirming our proteomic data, carbohydrate-active transcripts for GH16 have been reported for termite symbionts and beetles [12, 50, 51] and GH7 have been retrieved in fungi and termite symbionts; GH 9 from C. gestroi was retrieved from the expressed sequence tag library, and GH2, GH3, GH11, GH18 and GH38 were represented in the host and symbiont sequence pools from Reticulitermes spp.
Our strategy not only enabled us to push forward the identification of glycoside hydrolases, but also validated our results for protein identification with both high and low confidence criteria (identification of only one unique peptide). The protein matches with low confidence, such as GH10, GH11 and GH 43, were derived from chromatographic fractions that degraded xylan, likewise proteins matching GH2, GH38 and CBM27 corresponded to fractions with hydrolytic activity on mannan. GH16 was identified from chromatographic fractions with high hydrolytic activity on lichenan. Moreover, the partial purification of two enzymes with low confidence followed by characterization using specific APTS-labeled oligosaccharides, made two low confidence protein matches for GH2 and GH7 more reliable. Collectively, our findings can offer a first glimpse of the GH range existing in the lower termite digestome, as well as help to suggest target genes for cloning and heterologous expression in future termite-based biomass conversion strategies. The use of robust LC-MS/MS analyzers and deep sequencing in a third generation sequencer could generate large data sets for termite digestome analysis, and further studies may be required to reveal additional components, such as lignin degrading proteins.