Structural Bioinformatics Studies of Integral Transmembrane Enzymes pMMO Complex, C560, CYB, and DHSD and their AlphaFold3-Predicted Water-Soluble QTY Variants

The QTY code systematically replaces hydrophobic amino acids such as leucine (L), isoleucine (I), valine (V), and phenylalanine (F) into the hydrophilic amino acids glutamine (Q), threonine (T), and tyrosine (Y). Since the electron density maps of L vs Q, I/V vs T, and F vs Y are highly structurally similar, such conversion is able to preserve the functional structure of proteins while the transmembrane domain loses its hydrophobic characteristic.

We aligned the protein sequences of the native pMOA, pMOB, pMOC, CYB, C560, and DHSD and their QTY variants (Figure 2). Despite overall mutations of amino acids in the protein sequence (3.66–22.89%) and notable variation in the transmembrane domain (33.33–86.61%), the isoelectric-focusing points (pI) of native proteins and engineered QTY variants remain alike, with unit differences ranging between 0–0.10. This is due to the lack of charges in amino acids Q, T, and Y at neutral pH, which minimizes impact on a protein's pI. Such differences are rather insignificant regarding surface charges and unlikely to disrupt delicate structures and functions.

Figure 2.

Protein sequence alignments of six native protein sequences with their water-soluble QTY variants. The symbols | and ^*indicate whether amino acids are identical or different, respectively. The alpha helices (blue) are shown above the sequences. Alignments shown include a) pMOA vs pMOA^QTY, b) pMOB vs pMOB^QTY, c) pMOC vs pMOC^QTY, d) CYB vs CYB^QTY, e) C560 vs C560^QTY, f) DHSD vs DHSD^QTY.

After predicting the structures of water-soluble QTY variants with Alphafold3, we superposed the molecular structure of six CryoEM-determined transporters and their respective QTY variants in order to directly compare their structures. The molecular structures of native transporters are already experimentally determined for pMOA (PDB: 7EV9), pMOB(PDB: 7EV9), and pMOC (PDB: 7S4I), CYB (PDB: 8IOG), C560 (PDB: 8GS8), DHSD (PDB: 8GS8). Thus, the superpositions were carried out for pMOA vs pMOA^QTY, pMOB vs pMOB^QTY, pMOC vs pMOC^QTY, C560 vs C560^QTY, CYB vs CYB^QTY, and DHSD vs DHSD^QTY.

The experimentally determined native molecular structures and AlphaFold3-determined modified QTY structures superposed surprisingly well, with root mean square deviation (RMSD) values ranging below 1.00 Å at between 0.302 Å and 0.595 Å for all six protein pairs (Table 1 and Figure 3). Specifically, their RMSD values are as follows: pMOA vs pMOA^QTY (0.302Å), pMOB vs pMOB^QTY (0.460Å), pMOC vs pMOC^QTY (0.590Å), CYB vs CYB^QTY (0.595Å), C560 vs C560^QTY (0.489Å), and DHSD vs DHSD^QTY (0.343Å). Despite 33.33–86.61% amino acid substitutions in the transmembrane domain using the QTY code, these results demonstrate relatively similar and preserved 3-dimensional folds between the native transporters and their QTY variants.

Figure 3.

Superpositions of six Cryo-EM-determined structures of native oxidation enzymes and their AlphaFold3-predicted water-soluble QTY variants. The CryoEM-determined structures are obtained from the Protein Data Bank (PDB). The CryoEM structures (magenta) are superposed with their QTY variants (cyan) predicted by AlphaFold3. These superposed structures show that the native transporters and their QTY variants have very similar structures. For clarity of direct comparisons, unstructured loops in the CryoEM structures were removed in the QTY variants. a) pMOA vs pMOA^QTY (0.302 Å), b) pMOB vs pMOB^QTY (0.460 Å), c) pMOC vs pMOC^QTY (0.590 Å), d) CYB vs CYB^QTY (0.595 Å), e) C560 vs C560^QTY (0.489 Å), and f) DHSD vs DHSD^QTY (0.343 Å).

We also ask how well the AlphaFold3-predicted native transporters would superpose with the AlphaFold3-predicted structures of their water-soluble QTY variants. Therefore, we carried out the superposition of the following pairs of protein structures (Figure 4) and yielded relatively low RMSD values for pMOA^AF3 vs pMOA^QTY, pMOB^AF3 vs pMOB^QTY, pMOC^AF3 vs pMOC^QTY, CYB^AF3 vs CYB^QTY, C560^AF3 vs C560^QTY, and DHSD^AF3 vs DHSD^QTY. These results suggest AlphaFold3 predicted and experimentally determined structures share similarity between the native proteins and their QTY variants, although the QTY variant structures need to be experimentally verified. We also superposed the AlphaFold3-predicted native structures with their experimentally obtained CryoEM structures and calculated their RMSD to further test AlphaFold3's accuracy (Figure S8 in Supplementary Materials).

Figure 4.

Superpositions of AlphaFold3-predicted structures of native and their QTY variants. Color code: green = AlphaFold3-predicted native structures; cyan = AlphaFold3-predicted water-soluble QTY variants. a) C560^AlphaFold3 vs C560^QTY, b) CYB^AlphaFold3 vs CYB^QTY, c) DHSD^AlphaFold3 vs DHSD^QTY, d) pMOA^AlphaFold3 vs pMOA^QTY, e) pMOB^AlphaFold3 vs pMOB^QTY, and f) pMOC^AlphaFold3 vs pMOC^QTY.

We also ask how the superposition will appear i) if we superpose the experimentally determined CryoEM transporter structures with ii) AlphaFold3-predicted native proteins and iii) AlphaFold3-predicted water-soluble QTY variant proteins. The superposition of the above three structures are shown in Figure 5. Structures in the superposition appear to align well, with minor differences in their residues and folding. To gain a clearer view of potential structural differences that would affect the proteins' abilities to function as transmembrane proteins, we also superposed only the transmembrane domains of the structures (Figure S9 in Supplementary Materials).

Figure 5.

Superpositions of CryoEM structures with AlphaFold3-predicted native enzymes and their water-soluble QTY variants. Superposition of i) the experimentally determined CryoEM structures (magenta) with ii) AlphaFold3-predicted native transporters (green) and iii) AlphaFold3-predicted water-soluble QTY variant transporters (cyan). a) C560^CryoEM vs C560^AlphaFold3 vs C560^QTY, b) CYB^CryoEM vs CYB^AlphaFold3 vs CYB^QTY, c) DHSD^CryoEM vs DHSD^AlphaFold3 vs DHSD^QTY, d) pMOA^CryoEM vs pMOA^AlphaFold3 vs pMOA^QTY, e) pMOB^CryoEM vs pMOB^AlphaFold3 vs pMOB^QTY, and f) pMOC^CryoEM vs pMOC^AlphaFold3 vs pMOC^QTY.

To validate the effectiveness of our protein engineering, we mapped the hydrophobicity of Cryo-EM-determined structures against AlphaFold3-predicted structures with the QTY code applied (Figure 6). The original proteins possess 2–8 transmembrane alpha-helical domains directly embedded in phospholipid bilayers, and their structures contain the highly hydrophobic amino acids L, I, V, and F. As desired, hydrophobic patches were largely reduced after systematic replacement of amino acids L, I, V, F with the more hydrophilic Q, T, and Y using the QTY code in said transmembrane regions, demonstrating the desired conversion.

Figure 6.

Hydrophobic surface of six native proteins and their water-soluble QTY variants. The native oxidation enzymes have many hydrophobic residues L, I, V, and F in the transmembrane helices. After Q, T, and Y substitutions of L, I and V, and F respectively, the hydrophobic surface patches (yellow) in the transmembrane helices become more hydrophilic (cyan). a) pMOA vs pMOA^QTY, b) pMOB vs pMOB^QTY, c) pMOC vs pMOC^QTY, d) CYB vs CYB^QTY, e) C560 vs C560^QTY, f) DHSD vs DHSD^QTY.

Structural biology researchers have sought to predict posttranslational protein folding accurately and efficiently for almost 70 years. With the advent of machine learning tool AlphaFold2 in 2021 (36, 37) and AlphaFold3 in 2024 (38), we are now able to study the structure of previously unattainable proteins, especially those with integral embedded transmembrane regions. Rather than undergo the difficult process of gene expression, protein production, detergent selection, purification, detergent exchange, maintaining stability and avoiding agglutination, scientists are able to generate fairly accurate predictions within hours, or even minutes depending on protein size.

In collaboration with the European Bioinformatics Institute (EBI), DeepMind has made over 214 million predicted protein structures freely available online (https://alphafold.ebi.ac.uk) within the three years since AlphaFold2's release. This speed and accuracy is unprecedented in the field of structural biology (36, 37, 38).

Particulate methane monooxygenase (pMMO) is one of the three enzymatic complexes known to carry out methane hydroxylation (8,9,10,11,12,13,14,15). The enzyme complex is very important for removing methane from the air and reducing the climate change damage. Its mechanism is not fully understood but thought to be related to the structurally similar cytochrome c oxidase II complex which also contains functionally important copper ions. The cytochrome c oxidase II complex found in the mitochondrial membrane is also involved in human diseases including mitochondrial encephalomyopathy, hypertrophic cardiomyopathy (HCM), sporadic mitochondrial myopathy (MM), pheochromocytoma syndrome, gastric stromal sarcoma, and nuclear type 3 mitochondrial complex II deficiency. We thus are interested to carry out the structural bioinformatic study of the six proteins in pMMO subunits and representative parts of the repository complex with known CryoEM structures to further validate the QTY code.

The water-soluble QTY variants may be useful in a variety of scenarios including i) using the purified soluble proteins to carry out ligand-binding drug discoveries, ii) determining the currently unknown detailed mechanisms and functions of methane monooxygenase and similar copper-center-containing proteins, and iii) potentially achieving industrial-scale use of methane as a biofuel by expressing-soluble pMMOs in filamentous algae which allows direct secretion of the protein into surrounding culture media without the need of cell break-down.

We previously showed that the QTY code can also be reversible because of their simple pairwise changes, L <=> Q, I/V <=> T, and F <=> Y. This is similar as the DNA double-helix, A <=> T, G <=> C and vice versa T <=> A, C <=> G. These QTY or reverse QTY changes can be results of a single base pair mutation (Figure S1).

Karagöl et al. reported in several glutamate transporters and (29), serotonin, dopamine and norepinephrine transporters (30) that underwent such L <=> Q, I <=> T, and F <=> Y mutations. Some of them result diseases and others are silent, all depending on the specific location of mutations.

Membrane transporter proteins are conserved from bacteria to mammals (39). This remarkable conservation makes them ideal subjects for evolutionary studies. For instance, evolutionary profiling potentially allows a better understanding of systematic amino acid substitutions, including the QTY code (29, 30). The phenomena perhaps also took place in other transmembrane helical proteins including pMOA, pMOB, pMOC, CYB, C560, and DHSD. Applied the same idea to the transmembrane helices of pMMO complex and others, QTY code-based studies are likely to provide insights into biological foundations of chemically distinct alpha-helices. Understanding the alpha-helical evolution has a potential to provide crucial information on protein molecular evolution and sequence-based structural predictions. Moreover, by combining this information with variant based approaches, we analyzed to what extent Darwinian natural selection and mutationally favored biases affected the diversification of QTY-code pairs (L/Q, I/T, V/T, and F/Y).

The chemical properties of the helical structures are governed by the genetic code, as genetic code's second position is determinant of the chemical nature of amino acids (1, 2, 3, 4, 19). For instance, i) amino acids with U at the second position are hydrophobic (Phe, Leu, Ile, Val, and Met); ii) amino acids with C at the second position are less hydrophobic (Pro and Ala), or with a hydroxyl -OH group (Ser and Thr); iii) amino acids with A at the second position are hydrophilic and water-soluble (Asp, Glu, Asn, Glu, Lys, His and Tyr), and 2 stop codons Ochre (UAA) and Amber (UAG); iv) amino acids (Arg and Ser) with G at the second position are water-soluble, Cys is partially water-soluble and Gly is achiral and has a hydrogen as the side chain (1, 2, 3, 4, 19). The stop codon is UGA. In general, pyrimidine U and C at the second position confer hydrophobicity; in contrast, purine A and G at the second position confer hydrophilicity (Figure S1).

The variations all occur in the second position of the genetic code, including transition mutation, i.e., purine to purine (A->G, G->A) and pyrimidine to pyrimidine (C->U, or U->C); or transversion mutation (U->A, U->G, C->A, C->G, A->U, A->C, G->U, G->C). In the case of L->Q, I->T, and F->Y substitutions. i) in L (leucine), two codons are CUA and CUG, and in Q (glutamine), two codons are CAA and CAG; in these cases, the second position of U is mutated to A, which is a transversion mutation (U->A). ii) In I (isoleucine), three codons are AUU, AUC, and AUA, in T (threonine), four codons are ACU, ACC, ACA, and ACG; which is a transition mutation (U->C). iii) In F (phenylalanine), two codons are UUU and UUC, in Y (tyrosine), two codons are UAU and UAC, and the second position of U is mutated to A which is a transversion mutation (U->A). The same logic applies in the case of the reverse QTY mutations of Q->L, T->I, Y->F. Transitions are expected to be more frequent than transversions, with a ratio much higher in coding regions. F<=>Y and L<=>Q changes that required same mutations (U->A and A->U) observed in similar ratios.

The native protein sequences for transporters including pMOA, pMOB, and pMOC, CYB, C560, and DHSD, were obtained from UniProt (https://www.uniprot.org). The sequences for the QTY variants were aligned using the same methods as previously described. The MWs and pI values of the proteins were calculated using the Expasy (https://web.expasy.org/compute_pi/).

The protein structures of the QTY variants were predicted using AlphaFold3 (https://github.com/sokrypton/ColabFold) and by following the instructions at the website. PBD files for the predicted native protein structures were obtained from The EBI (https://alphafold.ebi.ac.uk), which contains all AlphaFold3-predicted structures for native proteins. The UniProt website (https://www.uniprot.org) provided protein ID, entry name, description, and FASTA sequence for each native protein. The QTY code can be applied to FASTA sequences by manually replacing amino acids in the TM domains (found on the Protter 2D diagram-plotting website, http://wlab.ethz.ch/protter/start/) but can also be done through the QTY method website (https://pss.sjtu.edu.cn/). The website also provides MWs, pI values, TM variation, and overall variation.

PBD files for native protein structures experimentally determined by CryoEM were taken from the PDB include pMOA, pMOB, and pMOC, CYB, C560, DHSD. Predictions for the QTY variants were carried out using the AlphaFold3 program, which can be found at https://github.com/sokrypton/ColabFold. These structures were superposed and the RMSDs were calculated using PyMOL (https://pymol.org/3/). For detailed method information, please see references 25–30.

PyMOL (https://pymol.org/3/) was used to superpose the native protein structure and the QTY variant. UCSF Chimera (https://www.rbvi.ucsf.edu/chimera) was used to render each protein model with hydrophobicity patches. For detailed method information, please see references 25–30.