2.1 Structure characterizations of PSS
The commercial PSS sample was characterized by FT-IR, 1H-NMR, and viscosity molecular weight (Mƞ), and the results are presented in Fig. 1 (a, b and c). In Fig. 1a, the O-H stretching vibration is observed at band centered at 3436 cm-1, and the sulfonic acid group and the S=O characteristic absorption peaks in sulfonate are observed at 1182 cm-1 and 1035 cm-1 [36]. In Fig. 1b, the proton peak assignment from different carbon is presented in the chart. The resonance from 6.4 to 7.9 ppm is due to the para-substitution on the benzene ring [37]. From the 1H-NMR spectra, it can be seen that the PSS sample is quite pure. The MW of PSS was assessed by intrinsic viscosity [38]. The intrinsic viscosity of PSS was determined as shown in Fig. 1c, and the value was determined to be 0.2256 g/100mL. Mƞ of PSS was calculated based on Eq. (1) to be 61080, which agreed with the value (~70000) claimed by the supplier very well since viscosity-averaged molecular weight is usually less than the weight-averaged molecular weight. In sum, the structure of PSS used in this investigation was well-defined.
2.3 Chemical composition of poplar variants with green liquor pretreatment and the effect of PSS dosage on the enzymatic saccharification
As a mild alkali pretreatment method to remove some lignin and a small portion of hemicelluloses, GL pretreatment mainly used two chemical reagents, Na2S and Na2CO3 which are recycled from the kraft pulping process, to enhance the enzymatic hydrolysis productivity of hardwood [39, 40]. The benefits of this process are it reduces the amount of poisonous or erosive substances, such as furfural, acetic acid, and metal ions which may inhibit the subsequent fermentation or damage the pretreating equipment [41]. The enzymatic saccharification enhanced by LS is more pronounced for substrates subjected to kraft or alkaline pretreatment if compared with other pretreatments [13]. As shown in Table 1, the solid recovery was 84%, while most glucan (97%) in the poplar was reserved in the substrate. GL pretreatment reduced the crystalline index and polymerization degree of cellulose, ulteriorly increased substrate porosity to provide cellulase with the more contacted area. The enhanced cellulase accessibility benefited the bioconversion rate[41, 42].
Table 1 The main components of poplar and GL-P (based on raw material)
Materials
|
Solid recovery (%)
|
Carbohydrate (%)
|
Lignin (%)
|
Ash
(%)
|
Glucan
|
Xylan
|
Ara+Manb
|
KLc
|
ASLd
|
Poplar
|
/
|
43.0±0.2
|
15.0±0.2
|
2.1±0.1
|
22.0±0.9
|
2.4±0.5
|
0.69±0.01
|
GL-Pa
|
84.00
|
39.4±1.2
|
11.6±0.1
|
1.6±0.1
|
20.5±0.3
|
2.5±0.04
|
1.97±0.04
|
a Green liquor pretreated poplar
b Arabinan + Mannan
c Klason lignin.
d Acid-soluble lignin.
To verify whether the PSS has a similar function as LS or not, different loadings of PSS were added in the enzymatic hydrolysis system to evaluate its performance. In terms of its performance on pure cellulose enzymatic hydrolysis (i.e., Whatman filter paper), the substrate enzymatic digestibility (SED) or glucose yield was 32.0% without PSS or LS addition (Fig. 2a) at the enzyme loading of 6.3 FPU/g-glucan. When 0.05 g/g-substrate LS was put in, the glucose yield dropped to 29.3%. The enzymatic digestibility was decreased with the LS increment. It reduced to 23.5% and 23.2% for the LS dosages of 0.2 and 0.4 g/g-substrate, respectively. From the work by Wang et al.[12], their enzymatic hydrolysis efficiency after 72h decreased from 74.0% to 66.6% with the addition of 0.25 g/g-substrate commercial LS at an enzyme loading of 15 FPU/g-glucan. Therefore, the inhibition effect of LS on pure cellulose saccharification was consistent quite well with theirs in trend. Yet the effect of LS on enzymatic saccharification progress was closely bound up with MW and sulfonate group. As investigated by Lou et al. [31], LS with low MW and high sulfonate groups tends to promote the enzymatic hydrolysis of pure cellulose, whereas LS with high MW and low sulfonate groups induces the inhibition to some extent. However, PSS lowered the glucose yield dramatically to 9.5% even at the dosage as low as 0.025 g/g-substrate. With the PSS addition increased to 0.2 g/g-substrate, the glucose yield was gradually reduced to 7%. From this point of view, PSS exhibited a much severer inhibition on the enzymatic hydrolysis of pure cellulose than LS did. This may be attributed to its much higher MW (Mƞ =61080), compared with that of LS (~10000).
The glucose yield of enzymatic hydrolysis of GL-P is depicted in Fig. 2b. As can be seen, the yield was only 53.0% in the absence of any PSS or LS but increased significantly to 76.5% even at a very low dosage of PSS (i.e., 0.0125 g/g-substrate) under the enzyme loading of 13.3 FPU/g-glucan. Increasing PSS dosage led to the higher glucose yield and reached the highest (81.5%) at the PSS dosage of 0.1 g/g-substrate. However, when more PSS was added, the glucose yield started to decline, as observed in the cases of PSS dosages at 0.2 and 0.4 g/g-substrate. The promotion effect induced by PSS appeared to be similar to LS reported by Wang et al.[21], and their data are plotted in Fig. 2b for comparison. In their case, the highest glucose yield of GL-P reached 88.0% under the enzyme loading of 20 FPU/g-glucan with LS addition of 0.2 g/g-substrate. Even though their highest glucose yield was better than the value we obtained, it has to be noted their enzyme loading was 50% more than the enzyme loading in the current PSS system. When the dosage was less than 0.1 g/g-substrate, it can be observed the enzymatic hydrolysis rate induced by PSS was a little higher than that by LS at the same dosage, despite the enzyme loading difference. Under the same enzyme loading of 13.3 FPU/g-glucan, LS at its optimal dosage of 0.2 g/g-substrate led to the glucose yield of GL-P at 75.2% (Fig. 3b), which was inferior to the yield resulting from PSS at the half dosage. What’s more, the best glucose yield of GL-P obtained by PSS addition was much higher than that of poplar after SPORL pretreatment with SPORL hydrolysate addition (52.5%) [43]. In terms of overall performance and the reduced dosage, PSS exhibited better-promoting performance on the lignocellulose saccharification than LS.
2.4 The effect of pH and the dynamic enzymatic hydrolysis
The system pH is a key factor influencing enzymatic saccharification of all kinds of substrate, especially with the participation of promoting additives [12, 29]. As depicted in Fig. 3a, enzymatic hydrolysis of Whatman paper and GL-P with and without the addition of 0.1 g/g-substrate PSS were conducted in the pH range from 4.0 to 7.0. Firstly, for pure cellulose, PSS inhibited the SED of Whatman paper in the whole range of pH from 4.0 to 7.0, particularly pronounced in the low pH range. The work reported elsewhere on LS [31] also showed that the optimal pH was changed with the MW and sulfonate content of LS. Meanwhile, the enzymatic saccharification of pure cellulose was enhanced especially by the LS of low MW and high content of sulfonate (or hydrophilic) groups which facilitate the binding with cellulase. Compared with LS, the sulfur content of PSS used in this work is relatively high (9.68 wt%), while the MW is also higher, compared with those of LS. As a result, the better performance of PSS in cellulose enzymatic hydrolysis can be anticipated.
In terms of the promoting performance for GL-P enzymatic hydrolysis, it was found on the contrary PSS promoted the SED of GL-P in the pH range between 4.5 to 6, particularly pronounced at the pH close to 4.8. The SED was improved from 53.1%, 62.1% and 52.1% to 81.5 %, 66.5 % and 58.2 % when pH was 4.8, 5.0 and 5.5, respectively. Lou et al.[29] and Wang et al.[12] reported that LS exhibited a promotion effect at elevated pH, which was ascribed to the dissociation of phenolic groups of lignin (LS and residual lignin) and strong electrostatic repulsion between LS-cellulase complexes and residual lignin. In our case, there are no phenolic groups in PSS and the abundant sulfonate groups remain negative-charged in the pH range tested. Therefore, the pH effect on PSS promoting behavior is not significant as that for LS. The optimal pH range for celluases also remained unchanged.
To compare the effect of PSS and LS on enzymatic hydrolysis of lignocellulose, the glucose yield of GL-P after PSS/LS addition was plotted again enzymatic hydrolysis time and the results are shown in Fig. 3b. Since PSS processed a severer inhibition to the pure cellulose enzymatic hydrolysis, a lower PSS dosage at 0.05 g/g-substrate was used as well. The glucose yield of GL-P was increased obviously from 53.1% to 75.2% and 81.5%, respectively with LS and PSS applied. Even for the low PSS dosage (0.05 g/g-substrate), its glucose yield after 72 h also reached 76.5%, which was slightly higher than that of LS at a four-fold dosage. After fully enzymatic hydrolysis for 72 h, the enzymatic digestibility of GL pretreated substrate in the presence of additives followed the order of PSS 0.1 > PSS 0.05 > LS 0.2 > Control. Lou et al. [30] reported a much higher glucose yield when LS was applied as an additive. Compared with the results in our case, the substrate they used was the mixture of Whatman filter paper and enzymatic hydrolysis lignin. Even though their substrates contained similar chemical compositions as ours, the lignin chemistry and its distribution in the fiber should be quite different from the lignocellulose used in this work.
Interestingly, when the enzymatic digestibility was examined at the period less than 48 h, a different order was observed. For example, for the enzymatic digestibility of 24 h, the order followed LS 0.2 > PSS 0.05 > PSS 0.1 > Control, which was opposite with the order of 72 h. The results indicated both additives of LS and PSS could promote enzymatic efficiency when compared with the control. However, if comparing enzymatic digestibility between LS and PSS additives, it can be found that the enzyme cocktails of cellulases and PSS showed a lower enzyme activity in a short processing period but with a prolonged processive time. This unique and interesting phenomenon was noticed for the first time in the current, which has not been reported yet in previous work. However, it remains unclear why PSS lowers the enzyme activity but with a longer processive time than LS, which is worthy of further investigation.
2.5 Interactions between PSS and cellulase
To investigate the effect of PSS on cellulase, the zeta potentials of the cellulases and the mixtures of cellulase with various loadings of PSS were measured and the results are given in Table 2. The cellulase used in this investigation had a slight zeta potential of 1.89±0.05 mV. When PSS was added into the system and bound to cellulase molecules, all the mixtures of cellulase with various loadings of PSS exhibited negatively charged due to the strongly charged sulfonate groups of PSS, which rendered the complexes of cellulase and PSS with negative charges. With the increment of PSS addition from 0.05 to 0.10 g/g-substrate, the zeta potential of the mixtures dropped from -28.95 to -31.37 mV. Very interestingly, with further PSS increment to 0.15 and 0.20 g/g-substrate, the zeta potential of cellulase-PSS complex started to recover some extent to -25.55 and -21.53 mV, respectively. The most negatively charged mixture was achieved for the PSS dosage of 0.10 g/g-substrate. This agreed with the previous SED results; and the negatively supercharged cellulases rendered them lignin-resistant, as addressed by Whitehead et al. [44].
Table 2 Zeta potential (ζ) of cellulase and the mixtures of cellulase and PSS
Samples
|
Cellulase
|
Cellulase+0.05g/g-substrate PSS
|
Cellulase+0.10g/g-substrate PSS
|
Cellulase+0.15g/g-substrate PSS
|
Cellulase+0.20g/g-substrate PSS
|
ζ (mV)
|
1.89±0.05
|
-28.95±1.88
|
-31.37±0.17
|
-25.55±0.55
|
-21.53±1.37
|
The interaction between cellulase and PSS monitored by a QCM-D E4 is presented in Fig. 4a. Initially, the cellulase was immobilized on the gold surface of QCM sensors and then the interaction between PSS and cellulase was monitored in situ and in real-time (the data didn’t show in the chart). The cellulase solution was injected into the chamber at 119 min when the frequency (Δf) and the dissipation (ΔD) became smooth or steady. Once the PSS solution flowed into the chamber at 145.3 min, the Δf3 was dropped from +116.3 Hz to +100 Hz, and it recovered 5 Hz to +105 Hz after buffer rinse. In the meanwhile, the ΔD3 increased mildly by about 1.25×10−6. This adsorption process proceeded relatively fast, taking approximately 17.5 min. Regarding the interaction between cellulase and LS, the Δf3 dropped by ca. 7.8 Hz and the ΔD3 raise ca. 1.43×10−6 in the same medium [33]. It can be seen that PSS has a similar interaction as LS does to form complexes with cellulase when they contact in solution.
Based on the QCM-D overtone data, the thickness, viscositic, and shear moduli of cellulase and PSS films were fitted with Sauerbrey and viscoelastic model by Dfind (a software provided with QCM-D instrument by Biolin Scientific Corp.), and the results are depicted in Figs. 4b and 4c. Since Sauerbery always underestimates the film thickness when the film is viscoelastic and cellulase layer and the formed complex layer were indeed viscoelastic, thereafter the thickness was referred to as viscoelastic thickness. The thickness of cellulase film was 11.56 nm, which was quite close to the value (11.7 nm) reported previously by Wang et al. [33]. This demonstrated the cellulase immobilization on QCM sensors was repeatable and the protocol was reliable. When the stable complexes of cellulase and PSS were formed and subjected to buffer rinsing, the thickness of the PSS layer was 4.22 nm, which was a little bit thicker than that of LS film (i.e. 3.4 nm) adsorbed on the cellulase film (Table 3) [33]. Fig. 4c demonstrates the viscositic and shear moduli real-time change with the formation of the cellulase and PSS films. The viscositic and shear moduli of cellulase adlayer were 0.00165 kg/ms and 9.7 KPa, respectively. When the PSS solution was loaded, the viscositic moduli dropped slightly to 0.00157 kg/ms and the shear moduli increased significantly to 28.7 KPa. The viscositic moduli of PSS adlayer recovered to 0.00162 kg/ms and the shear moduli dropped to 12.66 KPa after buffer rinse. Compared with the values obtained for cellulase and LS, the viscositic moduli of both were quite close whereas the shear moduli of the complex of cellulase and PSS was much lower than that of the complex of cellulase and LS. This may be attributed to the difference in chemical structure between PSS and LS, since PSS exists as linear and random coils, whereas LS presents as rigid spheres in solution.
Table 3 Thickness and coupled water of PSS/LS adlayer on cellulase film
Layer
|
QCM* (nm)
|
SPR** (nm)
|
Coupled water (%)
|
PSS
|
4.22
|
0.2
|
95.3
|
LS***
|
3.4
|
0.73
|
78.5
|
*The data presented in Fig. 4a fitted by the software Q-Dfind provided by Biolin Scientific Corp.; **The data presented in Fig. 4d simulated by the software WinSpall version 3.02; ***The data of LS were cited from ref.[45].
Based on the principles of QCM-D, it responds to the weight change on the surface of the sensor. As a consequence, the water molecules moved with the adsorbed layer are also considered. To further study the interaction between cellulase and PSS, surface plasmon resonance (SPR) based on optical technology was employed and it only measures the “dry” mass on the sensor. SPR angle changes with cellulase immobilization and PSS loading are depicted in Fig. 4d. The SPR angle increased with the celluase binding on the gold sensor. When the PSS solution was injected in, the SPR angle decreased obviously, which is attributed to the low refractive index of PSS (n=1.38). For MP-SPR data, the recorded full spectra were simulated by the software WinSpall (version 3.02). The fitted thickness of PSS with dry mass adsorbed on enzymes acquired by SPR was 0.2 nm (Table 3). This is much less than the value of LS adsorbed on cellulase film (0.73 nm). By comparing the thickness obtained by QCM-D and SPR techniques, the coupled water can be derived and the values are reported in Table 3. The results showed that the coupled water of PSS is as high as 95.3%, indicating it is super highly hydrated. While this value is only 78.5% for LS adlayer in the same condition. This can be attributed to the difference in sulfonating degree for PSS and LS samples. And the difference in hydration between PSS and LS layers also can explain why PSS layer possessed a much lower shear modulus than LS adlayer did.
2.6 Mechanism of PSS promoting enzymatic saccharification
The carbohydrate-binding module (CBM) plays an important role in cellulose enzymatic hydrolysis[46] and also in lignin-binding since lignin and derivatives prefer to bind to the CBM portion of cellulases[47, 48]. To wrap it up, the PSS promoting mechanism for lignocellulose enzymatic saccharification is illustrated in Fig. 6. In a normal enzymatic hydrolysis scenario without promoting agent presence (Fig. 6a), some cellulase molecules perform normally to bind to cellulose surface by their CBM, while others are inevitably and irreversibly bind to the residual lignin due to the strong interactions between lignin and CBM, and the called non-productive binding occurs[47, 48]. When PSS is introduced into the system, as depicted in Fig. 6b, PSS molecules preferentially bind to the CBM of cellulases, and the formed complexes carrying more negative charges inhibit their non-productive binding to residual lignin due to the electrostatic repulsion, since residual lignin after GL pretreatment carries negative charges. As such, the amounts of cellulases that can bind to the cellulose surface increase proportionally. If there is no residual lignin present as in the case of Whatman filter paper, no such competitive adsorption occurs and therefore, only inhibition will be observed. Both scenarios are shared by LS and PSS.
Regarding the difference between LS and PSS, many studies confirmed that LS molecules present spheroidicity or dish structure in solution due to its 3D networks[49-51]; while PSS molecules are extremely flexible in the unsalted aqueous solution [52]. PSS exhibits a much-extended conformation in aqueous solution than that of LS (Fig. 6c). Therefore, the adlayer of PSS exhibited much lower shear moduli than those of LS (Fig. 4c). In addition, the high sulfonating degree of PSS renders it highly hydrated (Table 3). Due to its soft and flexible conformation, PSS can provide more binding sites to cellulases than LS in a given mass, representing the much lower adsorbed thickness monitored by SPR (Fig. 4d and Table 3). That is the underlying reason why PSS promotes lignocellulose enzymatic saccharification more efficiently than LS, especially at a lower dosage. However, the prolonging processive action of cellulases with the addition of PSS remains unresolved, and this makes the PSS promoting mechanism for lignocellulose saccharification more profound and it is desired further investigation.