Effects of Point Mutations on the Thermostability Ofb. Subtilis Lipase: Investigating

Effects of point mutations on the thermostability ofB. subtilis lipase: Investigating Nonadditivity

Bipin Singh†, Gopalakrishnan Bulusu†‡*, Abhijit Mitra†

†Center for Computational Natural Sciences and Bioinformatics (CCNSB), International Institute of Information Technology Hyderabad (IIIT-H), Gachibowli, Hyderabad, 500032, India

‡TCS Innovation Labs-Hyderabad (Life Sciences Division), Tata Consultancy Services Limited, Madhapur, Hyderabad, 500081, India

*Corresponding author E-mail: ;

Phone: +91-40-66673591

FAX: +91-40-66672222

SupplementaryMaterial

Text S1

(i) Conformational stabilities of in-silico mutants: RMSD, hydrophobic SASA (SASA-H), secondary structure and Free energy landscape (FEL) analyses

Broad structural parameters of conformational stability viz. C-α RMSD, SASA-H, number of residues in structural regions of protein (i.e. α-helices, β-sheets, β-bridges and turns) and number of hydrogen bonds between amino acid residues for all the systems are shown in Figs. S1, S2, S3 and S4.

Set 1

i. 12M with N166Yreversed (m12)

ii. 12M with N166Yand L114P reversed (m12_m20)

iii. 12M with A15S and A20E reversed (m51_m52)

iv. 12M with A15S, A20E and G111D reversed (m51_m52_m53)

v. 12M with M137Preversed (m62)

Except the mutants shown above, we have not observed any significant conformational changes in other in-silico mutants compared to the 12M mutant (Tables S1 (A), S1(B) and S1(C)). The mutants with serial no. iii-v show greater SASA-H (Table S2) compared to mutants with serial no. i-ii, suggesting their greater propensity towards aggregation at high temperature. Therefore, reversal of point mutations with serial no. iii-vnot only causes conformational destabilization but also increases the tendency towards possible aggregation at high temperature. To further support the above arguments, we have calculated the SASA-H (hydrophobic SASA) of the core β-strands (β1−β6) for all the systems (Fig.S5), which provide the information about the unfavorable contact of solvent molecules with the hydrophobic core structural regions of protein and thus the possible aggregation propensity of the protein at high temperature.

The in-silico mutants mentioned below show least effect on the conformational stabilization despite the reversal of point mutation(s), as suggested by different conformational properties of these systems (Figs. S1, S2, S3 and S4).

Set 2

i. 12M with A132D reversed (m11)

ii. 12M with I157M reversed (m30)

iii. 12M with N89Y, F17S and I157M reversed (m30_m41_m42)

iv. 12M with N89Y and F17S reversed (m41_m42)

v. 12M with A20E reversed (m52)

vi. 12M with G111D reversed (m53)

Fig. S1Mean C-α RMSDs for all the in-silico mutants. In-silico mutants showing higher RMSDs are shown in red color. Error bars show the standard deviations

Fig. S2Mean number of H-bonds formed between protein atoms for all the in-silico mutants. In-silico mutants showing significant reduction are shown in red color. Error bars show the standard deviations

Fig. S3Mean SASA-H for all the in-silico mutants. In-silico mutants showing higher SASA-H are shown in red color. Error bars show the standard deviations

Fig. S4Mean number of residues present in the structural regions (α-helices, β-sheets, β-bridges and turns) of all the in-silico mutants. In-silico mutants showing significant decrease in are shown in red color. Error bars show the standard deviations

Table S1(A) Structural properties of the 12M mutant. Standard deviations are shown inside the parentheses

12M / Values
RMSD (nm) / 0.34 (0.07)
SASA-H (nm2) / 34.5 (1.7)
No. of H-bonds / 132 (8)
No. of residues in structural regions / 102 (7)

Table S1(B)C-α RMSF (nm) for catalytically important residues of the 12M mutant

12M / C-αRMSF
I12 / 0.54
M78 / 0.14
S77 / 0.15
D133 / 0.25
H156 / 0.17

Table S1(C)Mean interatomic distances (nm) between catalytically important residues in the 12M mutant. Standard deviations are shown inside the parentheses

12M / Mean distance
I12-M78 / 0.89 (0.31)
S77-M78 / 0.32 (0.06)
D133-H156 / 0.98 (0.23)
S77-H156 / 0.60 (0.23)

Table S2In-silico mutants showing differential effects upon reversal of point mutations. Standard deviations are shown inside the parentheses

S. No. / Mutations Reversed in 12M / Mean Cα-RMSD in nm / Mean hydrophobic SASA in nm2 / Mean number of residues in structural regions
1* / N166Y (m12) / 0.37 (0.12) / 35.7 (2.01) / 102 (8)
2* / N166Y & L114P (m12_m20) / 0.38 (0.12) / 34.3 (1.58) / 97 (8)
3** / I157M & N89Y (m30_m42) / 0.23 (0.08) / 35.2 (1.79) / 104 (6)
4** / I157M, N89Y, F17S & A132D (m11_m30_m41_m42) / 0.31 (0.07) / 35.7 (1.72) / 103 (7)
5** / I157M, N89Y, F17S, A132D & N166Y (m11_m12_m30_m41_m42) / 0.32 (0.09) / 35.3 (1.47) / 100 (7)
6* / A15S & A20E (m51_m52) / 0.41 (0.17) / 36.1 (1.89) / 104 (7)
7* / G111D, A15S & A20E (m51_m52_m53) / 0.56 (0.23) / 38.8 (2.87) / 94 (12)
8* / M137P (m62) / 0.44 (0.16) / 36.0 (2.28) / 98 (8)

**System showing least effect upon reversal of point mutations

*Systems showing significant effect upon reversal of point mutations

Fig. S5 Distribution of SASA-H of the core secondary structural region (the six parallel β-strands β1−β6 consist of residues 6−9, 36− 38, 71−76, 96−102, 124−130, 147−151, respectively) for the in-silico mutants

The above observations suggest that only few combinations of point mutations show their effect additively in terms of conformational stabilization at high temperatures. Therefore, we suggest that despite their individual stabilizing nature, not all the point mutations cooperatively enhance the conformational stability, when combined and vice-versa. According to our MD simulation results, combination of mutations [G111D, A15S & A20E] is most important in terms of conformational stability of 12M followed by [M137P],[A15S & A20E], [N166Y & L114P] and [N166Y], because their reversal from the 12M, results in significant destabilization of the respective in-silico mutant(s) as suggested by different computed conformational properties (Figs. S1, S2, S3 and S4). It was also notable that, though the collective reversal of point mutations [G111D, A15S and A20E] causes significant destabilization, these mutations show least effect on conformational stability of the respective in-silico mutants (M17, M18 and M19) upon their individual reversal from the 12M mutant. These observations highlight the role of cooperative non-additive effects between these residues.Also, we have predicted the role of most robust point mutation M137P[1]accurately, as the reversal of this particular point mutation from 12M (the M25 in-silico mutant) results in significant loss of conformational stability.

To visualize the conformational landscape at high temperature, we have performed PCA based free energy landscape (FEL) analysis. As suggested by FEL plots, mutants with serial no. iii-vof Set 1may undergo native to aggregation prone non-native states, as shown by their respective FEL plots (Fig. S6).These systems sampled distantly located metastable states on FEL, which suggest less probability of accessing native states from non-native states, therefore the possibility of irreversible transition at high temperature. Other systems shown in Table S2 have sampled many metastable states well connected to each other, as shown by their respective FEL plots, thus provide them the capability of reversible transition to native state. They also show less RMSF for catalytic site residues compared to mutants with serial no. ii-v of Set 1 (systems 2, 6 7 and 8 in Fig. S7), suggesting less fluctuation in catalytic site geometry.

Fig.S6Free energy landscapes (FELs) for systems which show greater effect on conformational stabilization upon reversal of point mutations

Fig. S7 RMSF of catalytic site residues for all the in-silico mutants shown in Table S2

(ii) Hydrogen bonds

To explore the possible reasons for the differential effect of reversal of point mutations in the conformational stability of Set 1 and Set 2 in-silico mutants, we have compared the hydrogen bond occupancies of these in-silico mutants (Table S3). As expected, the more stable in-silico mutants possess higher number of hydrogen bonds with greater occupancies, which were either present with very less occupancy or absent in 12M mutant. Similarly the less stable in-silico mutants show decrease in occupancies of hydrogen bonds or loss of hydrogen bonds (due to in-silico mutation) compared to the 12M mutant (Table S1 (A)). It was also clear that the less stability of serial no. iv in-silico mutant of Set 1 was due to loss of many stable hydrogen bonds compared to the 12M mutant. In contrast, the higher stability of serial no. ii-vi in-silico mutants of Set 2 was due to stabilization of existing hydrogen bonds or formation of new hydrogen bonds. In serial no. iii of Set 2 in-silico mutant, the hydrogen bonds formed due to reversal of point mutations show greater occupancy (Table S3), explaining the reason for conformational stabilization.

Table S3Difference in H-bond occupancies in 12M with N89Y, F17S and I157M reversed (m30_m41_m42) and 12M systems

12M with N89Y, F17S and I157M reversed / 12M
ASN89N-TYR85O 89.706
LEU160N-ILE157O 76.236 / TYR89N-TYR85O 68.043
LEU160N-MET157O 63.955

(iii) Change in overall flexibility

We have analyzed RMSF of all the in-silico mutants to assess the change in flexibility due to reversal of point mutation(s) from the 12M mutant. The RMSF plots (Fig. S8) suggest the regions which show increase or decrease in overall flexibility due to reversal of mutation(s).The mutants with serial no. i-v of Set 1show overall large fluctuation including the N and C-terminal end of the protein. Another in-silico mutant with L114P mutation reversed in 12M; S. No. 4 in Table 1 also shows large fluctuation at the C-terminal end, but the other structural regions show lesser flexibility compared to the mutants with serial no. i-v of Set 1. Therefore, we consider the mutants with serial no. i-v of Set 1 as potentially unstable due to reversal of point mutation(s) from the 12M mutant. On the other hand mutants with serial no. i-vi of Set2 show very less fluctuation including the N and C-terminal end of protein, suggesting the decrease in conformational fluctuation and lesser impact of reversal of point mutation(s) (shown below) from the 12M mutant.

Fig.S8C-α RMSF for all the in-silico mutant lipases at 450 K

(iv) Impact on the catalytic site geometry

Structural properties and dynamic behavior of enzyme active sites are generally finely tuned for optimal catalytic efficiency. Enzyme inactivation can be due to unfolding (partial or complete) or due to minor structural changes in the active site of an enzyme. To assess the effect of reversal of point mutation(s) on the geometry of catalytically important residues (Fig. 3b), in those in-silico mutants which were found to be stabilized or destabilized upon reversal of point mutation(s) (viz. Set 1 and Set 2), we have analyzed Cα RMSF, all-atom RMSF (Figs. S9 and S10) and interatomic distances (Fig. S11) between these residues. Continuing the similar trend, the mutants with serial no. i-vi of Set2 show less fluctuation in catalytically relevant residues and maintain the favorable geometrical organization of active/binding site by maintaining the inter-atomic distance under constrain at high temperature. In contrast the mutants with serial no. i-v of Set 1 show large RMSF and inter-atomic distances for catalytically relevant residues, suggesting destabilization of active/binding site at high temperature.

Fig.S9C-αRMSF of catalytic site residues

Fig.S10All-atom RMSF (in nm) of catalytic site residues

Fig. S11Mean interatomic distances (in nm) for catalytically relevant residues. The pair I12-M78 corresponds to the distance between the peptidic nitrogens of the two residues, S77-M78 pair corresponds to the side chain hydroxyl oxygen atom of S77 and peptidic nitrogen atom of M78,D133-H156 corresponds to distance between side chain delta oxygen atom of D133 and the delta nitrogen atom of H156, S77-H156 pair corresponds to the side chain hydroxyl oxygen atom of S77 and epsilon nitrogen atom of H156. Error bars show the standard deviations

Table S4Comparative analysis of H-bond occupancies between the 12M and differentin-silico mutants.The mutation(s) which were reversed inthe 12M mutant are show inside the parentheses

M1 (A132D)
>50% in M1 and <25% in 12M
HIS10 N VAL39 O
ILE123 N PRO115 O
>75% in M1 and <50% in 12M
VAL6 N LYS35 O / 12M
>50% in 12M and <25% in M1
GLN121 N ASP118 O
SER141 OG GLY104 O
M7 (I157M)
>50% in M7 and <25% in 12M
HIS10 N VAL39 O
ILE123 N PRO115 O
LEU160 N ILE157 O
>75% in M7 and <50% in 12M
VAL6 N LYS35 O / 12M
>50% in 12M and <25% in M7
LEU160 N MET157 O
M12 (N89Y & F17S)
>50% in M12 and <25% in 12M
ASN18 N SER15 O
ASN89 N TYR85 O
HIS10 N VAL39 O
ILE123 N PRO115 O
PHE19 N SER16 O
>75% in M12 and <50% in 12M
LYS35 N SER32 O
VAL6 N LYS35 O / 12M
>50% in 12M and <25% in M12
TYR89 N TYR85 O
M13 (I157M, N89Y and F17S)
>50% in M13 and <25% in 12M
ASN89 N TYR85 O
HIS10 N VAL39 O
ILE123 N PRO115 O
LEU160 N ILE157 O
SER130 OG VAL154 O
>75% in M13 and <25% in 12M
ASN89 N TYR85 O
HIS10 N VAL39 O
LEU160 N ILE157 O / 12M
>50% in 12M and <25% in M13
GLN121 N ASP118 O
LEU160 N MET157 O
TYR89 N TYR85 O
M18 (A20E)
>50% in M18 and <25% in 12M
ASN51 N THR47 O
GLY52 N ASN48 O
HIS10 N VAL39 O
SER24 N ALA20 O
>75% in M18 and <50% in 12M
VAL6 N LYS35 O / 12M
>50% in 12M and <25% in M18
SER24 N GLU20 O
M19 (G111D)
>50% in M19 and <25% in 12M
ASN18 N SER15 O
ASN51 ND2 ASN82 OD1
GLY52 N ASN48 O
HIS10 N VAL39 O
ILE123 N PRO115 O
>75% in M19 and <50% in 12M
LYS35 N SER32 O
VAL6 N LYS35 O / 12M
>50% in 12M and <25% in M19
ASP144 N ASP111 O
>90% in 12M and <75% in M19
ARG107 N LEU140 O
M31 (M137P, S163P & L114P)
>50% in M31 and <25% in 12M
HIS10 N VAL39 O
TYR125 OH LEU114 O
>75% in M31 and <50% in 12M
TYR125 OH LEU114 O
VAL6 N LYS35 O / 12M
>50% in 12M and <25% in M31
LEU140 N PRO137 O
TYR125 OH PRO114 O
>75% in 12M and <50% in M31
LEU140 N PRO137 O
M2 (N166Y)
>50% in M2 and <25% in 12M
GLY52 N ASN48 O
ILE123 N PRO115 O
LYS170 N ASN166 O / 12M
>50% in 12M and <25% in M2
LYS170 N TYR166 O
TYR166 OH LEU160 O
>90% in 12M and <75% in M2
ARG107 N LEU140 O
M6 (L114P & N166Y)
>50% in M6 and <25% in 12M
HIS10 N VAL39 O
ILE123 N PRO115 O
LYS170 N ASN166 O
TYR125 OH LEU114 O
>75% in M6 and <25% in 12M
TYR125 OH LEU114 O / 12M
>50% in 12M and <25% in M6
ASP64 N GLN60 O
GLN121 N ASP118 O
GLU65 N LYS61 O
HIS76 NE2 HIS156 O
LYS170 N TYR166 O
THR66 N VAL62 O
THR66 OG1 VAL62 O
TYR125 OH PRO114 O
TYR166 OH LEU160 O
>75% in 12M and <25% in M6
LYS170 N TYR166 O
>90% in 12M and <25% in M6
THR66 OG1 VAL62 O
M20 (A15S & A20E)
>50% in M20 and <25% in 12M
ILE123 N PRO115 O
SER131 N ILE151 O
>75% in M20 and <25% in 12M
ILE123 N PRO115 O / 12M
>50% in 12M and <25% in M20
GLN121 N ASP118 O
LYS23 N PHE19 O
SER24 N GLU20 O
TYR166 OH LEU160 O
VAL27 N LYS23 O
>75% in 12M and <50% in M20
LEU143 N ASN148 OD1
M23 (G111D, A15S & A20E)
NONE / 12M
>50% in 12M and <25% in M23
ALA113 N ASP144 O
ARG107 NH1 ARG142 O
ASN106 ND2 SER141 O
ASN148 ND2 LEU143 O
ASN174 N GLU171 O
ASP144 N ASP111 O
GLY172 N LEU168 O
HIS76 NE2 HIS156 O
LEU108 N ALA105 O
LEU143 N ASN148 OD1
LEU173 N ILE169 O
LYS23 N PHE19 O
SER141 N ASN138 O
SER141 OG GLY104 O
SER24 N GLU20 O
THR109 N ASN106 O
THR110 N ASN106 O
TYR166 OH LEU160 O
VAL27 N LYS23 O
>75% in 12M and <25% in M23
ALA113 N ASP144 O
ASN148 ND2 LEU143 O
GLY172 N LEU168 O
LEU143 N ASN148 OD1
>90% in 12M and <25% in M23
ASN148 ND2 LEU143 O
M25 (M137P)
>50% in M25 and <25% in 12M
HIS10 N VAL39 O
SER131 N ILE151 O / 12M
>50% in 12M and <25% in M25
GLN178 N TYR125 O
LEU140 N PRO137 O
TYR125 N GLN178 O
>75% in 12M and <25% in M25
LEU140 N PRO137 O

Fig.S12Predicted amylogenic regions using the amino acid sequences of 12M and the M1 in-silico mutant. The plot suggests that the reversal of A132D mutation increases the intrinsic aggregation propensity of the protein

Intrinsic aggregation propensity of the stable in-silico mutants

To examine the intrinsic aggregation propensity of the in-silico mutants, we have used the Waltz program[2]. Waltz uses an algorithm for prediction of amylogenic regions in protein sequences, trained from a large set of experimentally characterized amyloid forming peptides. Except for the M1 in-silico mutant (Fig.S12), none of the identified stable in-silico mutants show increase in intrinsic aggregation as analyzed using their respective amino acid sequences. Therefore, the stable in-silico mutants M7, M12, M13, M18 and M19 seem to be both conformationally stable and intrinsically aggregation resistant. The increase in intrinsic aggregation upon reversal of A132D mutation in 12M also supports the earlier experimental observation[3]regarding the stabilizing role of A132D mutation through alteration in aggregation kinetics rather than conformational stabilization of the protein.

Table S5Experimental data[1,3]for the single point mutants generated using 9M,showing the importance of M134E and M137P point mutations

Mutant / T50 (°C) / T1/2 in min / Tm(app) / ΔG (kcal mol-1) / Kcat/Km in mM- 1min-1
WT / 52.8 / 3.1
(at 55°C) / 56.3 / 10.39 / 2.2
9M / 68.4 / 4.4 (at 75°C) / 71.2 / 13.75 / 3.7
9M+M134E / 93.0 / 38.8 (at 75°C) / 72.9 / 14.41 / 13.8
9M+M137P / 93.0 / 101.2 (at 75°C) / 74.1 / 14.73 / 8.5
9M+S163P / 72.0 / 22.2 (at 75°C) / 72.2 / 13.50 / 9.14
12M / 93.0 / 430.5 (at 75°C) / 78.2 / 15.1 / 8.1

Table S6Experimental data[1,3]for single point mutants generated using WT, showing the importance of M134E and M137P point mutations

Mutant / T50 (°C) / T1/2 in min at 55°C / Tm(app) / ΔG (kcal mol-1)
WT+M134E (1M) / 57.8 / 196.0 / 60.4 / 11.03
WT+M137P(1M) / >65.0 / 2617.0 / 62.4 / 11.55

Table S7 Comparison between the H-bond occupancies for 9M and 12M mutants suggesting the H-bonds responsible for the observed difference in flexibility of regions involving H156 and residues 40-45

(A)12M_2 has these additional H-bonds compared to 12M_1
VAL27 N LYS23 O 51.451
HIS76 NE2 HIS156 O 51.201
GLN121 N ASP118 O 66.203
LEU140 N PRO137 O 82.198
SER141 N ASN138 O 57.727
LEU160 N MET157 O 63.955
TYR166 OH LEU160 O 50.051 / (B)12M_1 has these additional H-bonds compared to 12M_2
SER141 N PRO137 O 67.311
LEU140 N VAL136 O 55.557
TYR139 N VAL136 O 63.291
ILE123 N PRO115 O 66.889
GLY52 N ASN48 O 63.561
ASN51 N THR47 O 50.761
HIS10 N VAL39 O 65.171
(C) 12M_2 has these additional H-bonds compared to 9M
LYS23 N PHE19 O 65.687
SER24 N GLU20 O 51.721
HIS76 NE2 HIS156 O 51.201
GLN121 N ASP118 O 66.203
LEU140 N PRO137 O 82.198
LEU160 N MET157 O 63.955
TYR161 N GLY158 O 59.849
VAL165 N SER162 O 54.833
TYR166 OH LEU160 O 50.051 / (D) 9M has these additional H-bonds compared to 12M_2
ILE123 N PRO115 O 66.321
ASN82 ND2 GLY46 O 51.199
GLY52 N ASN48 O 58.369
ASN51 N THR47 O 53.033
HIS10 N VAL39 O 56.473
(E) 12M_1 has these additional H-bonds compared to 9M
SER141 N PRO137 O 67.311 / (F) 9M has these additional H-bonds compared to 12M
NONE

Table S8 Average no. of H-bonds formed in different systems. Standard deviations are shown inside the parentheses

System / H-bonds
WT+S163P / 113 (9)
WT+M134E / 122 (9)
WT+M137P / 114 (11)
WT+M134E+M137P / 123 (11)
WT+ M134E+M137P+S163P / 125 (9)

Fig. S13 DCC map showing the difference in dynamic cross correlation for 9M and 12M mutants

References

[1]S. Ahmad, N.M. Rao, Thermally denatured state determines refolding in lipase: Mutational analysis, Protein Sci. 18 (2009) 1183–1196. doi:10.1002/pro.126.

[2]S. Maurer-Stroh, M. Debulpaep, N. Kuemmerer, M. Lopez de la Paz, I.C. Martins, J. Reumers, K.L. Morris, A. Copland, L. Serpell, L. Serrano, J.W.H. Schymkowitz, F. Rousseau, Exploring the sequence determinants of amyloid structure using position-specific scoring matrices, Nat. Methods. 7 (2010) 237–242. doi:10.1038/nmeth.1432.

[3]S. Ahmad, V. Kumar, K.B. Ramanand, N.M. Rao, Probing protein stability and proteolytic resistance by loop scanning: a comprehensive mutational analysis, Protein Sci. 21 (2012) 433–446. doi:10.1002/pro.2029.

S1