Structural Catalytic Core of the Members of the Superfamily of Acid Proteases

Abstract

The superfamily of acid proteases has two catalytic aspartates for proteolysis of their peptide substrates. Here, we show a minimal structural scaffold, the structural catalytic core (SCC), which is conserved within each family of acid proteases, but varies between families, and thus can serve as a structural marker of four individual protease families. The SCC is a dimer of several structural blocks, such as the DD-link, D-loop, and G-loop, around two catalytic aspartates in each protease subunit or an individual chain. A dimer made of two (D-loop + DD-link) structural elements makes a DD-zone, and the D-loop + G-loop combination makes a psi-loop. These structural markers are useful for protein comparison, structure identification, protein family separation, and protein engineering.

1. Introduction

Earlier, we described structural catalytic cores in many serine and cysteine proteases and showed the presence of unique structure/functional environments, “zones”, around the catalytic sites in these proteins [1,2,3,4]. Each zone incorporated a segment of the catalytic core, connected to their respective element of protein functional machinery through a network of conserved hydrogen bonds and other interactions.

The four protease superfamilies studied earlier were (1) alpha/beta-hydrolases, (2) trypsin-like serine proteases, (3) cysteine proteinases, and (4) SGNH hydrolase-like proteins (SCOP (Structural Classification of Proteins, https://scop.mrc-lmb.cam.ac.uk/; accessed on 1 March 2024 [5]) IDs: 3000102, 3000114, 3001808, and 3001315, respectively). Each had only rare, structural exceptions, where aspartic acid could be found in place of the canonical catalytic serine or cysteine residues. At the same time, most of the proteases that predominantly use aspartic acid as a catalytic residue are grouped into the “acid proteases” superfamily (SCOP ID: 3001059). This superfamily belongs to the “all beta proteins” class (SCOP ID: 1000001) and includes four families, including the “pepsin-like” family (SCOP ID: 4002301). The 3D structure of a protein from the pepsin-like family consists of two similar beta barrel domains (N- and C-terminal) with one catalytic aspartate residue in each domain [6,7,8]. Aspartic proteases of this family use an activated water molecule bound to two conserved aspartate residues for hydrolysis of their peptide substrates. Enzymes of the pepsin-like family are synthesized as inactive zymogens (proenzymes), and later they are subsequently activated by cleavage of the N-terminal propeptide, and separate upon activation [9]. The protease 3D structures of the other three families resemble that of one of the structural domains of the peptidase from the “pepsin-like” family, and they become active when two monomers assemble to form the catalytically active dimer [10].

Here, we propose a general model of the conserved structural catalytic core (SCC) of aspartate proteases. Based on the “key” features of this model, we present a comparative structural analysis of 3D structures of superfamily representative domains in their zymogenic, free, and ligand-bound forms found in the Protein Data Bank (PDB [11,12]). In addition, we show a comparative structural analysis of SCC models obtained after dimerization of two identical amino acid chains of proteases or duplication of corresponding amino acid fragments within the same chain. Certain elements of catalytic mechanism are discussed only to highlight the role of shown residues, but the complete protein functional analysis is not within the scope of this manuscript.

2. Characterization of the Structural Catalytic Core of the Members of the Superfamily of Acid Proteases

2.1. Creating the Dataset of the Acid Proteases Superfamily Fold Proteins

The SCOP classification database [5] and the Protein Data Bank (PDB, http://www.rcsb.org/; accessed on 1 March 2024 [11,12]) were used to identify and retrieve 33 representative structures of proteins from the acid protease superfamily (SCOP ID: 3001059). Detailed descriptions of the protein structural information contained within this set of PDB files are given below.

Structure visualization and structural analysis of interactions between amino acids in proteins (hydrogen bonds, hydrophobic, other types of weak interactions) were performed using Maestro (Schrödinger Release 2023-1: Schrödinger, LLC, New York, NY, USA, 2021; https://www.schrodinger.com/user-announcement/announcing-schrodinger-software-release-2023-4; accessed on 1 March 2024) and software [13] to determine interatomic contacts, i.e., of ligand–protein contacts (LPCs) and contacts of structural units (CSUs).

Pairwise superpositions of representative structures were conducted using the Dali server (http://ekhidna2.biocenter.helsinki.fi/dali/; accessed on 1 March 2024) [14]. Weak hydrogen bonds from C-H•O contacts were identified, based on the criteria described in [15]. The π-π stacking and similar contacts were analyzed using the Residue Interaction Network Generator (RING, https://ring.biocomputingup.it/submit; accessed on 1 March 2024) [16]. Dimers were built using the “Protein interfaces, surfaces and assemblies” service PISA at the European Bioinformatics Institute (http://www.ebi.ac.uk/pdbe/prot_int/pistart.html; accessed on 1 March 2024) [17]. Figures were drawn with MOLSCRIPT [18].

Currently, according to the SCOP, the acid protease superfamily consists of four families: (1) Lpg0085-like (SCOP ID: 4001811), (2) retroviral protease (retropepsin) (SCOP ID: 4002288), (3) pepsin-like (SCOP ID: 4002301), and (4) dimeric aspartyl proteases (SCOP ID: 4004443), with more than 146 representative domains [5].

Representative 3D structures of this superfamily are tabulated in Table 1. Of the four families, only the pepsin-like family contains 3D structures of the zymogenic form of aspartic proteases. In addition to the SCOP database, we used data from the Proteopedia and the Uniprot databases (http://proteopedia.org/wiki/index.php/Main_Page; accessed on 1 March 2024 [19,20] and https://www.uniprot.org/; accessed on 1 March 2024 [21], respectively). Ten proenzyme structures were identified, and they are indicated with a “p” in Table 1. Since each 3D structure of the pepsin-like proenzymes contained two similar domains, both domains were separately analyzed at their catalytic regions, and thus Table 1 contains two lines for each PDB ID of a proenzyme labeled as “a” and “b”. For four proteins out of ten, in addition to coordinates of the zymogenic form, there were also available coordinates for both the ligand-free and ligand-bound forms, labeled in Table 1 with letters “c/d” and “e/f”, respectively. For three out of ten proteins, in addition to the coordinates of the zymogenic form, there were coordinates of only the ligand-bound form (i.e., “a”, “b”, “e”, and “f” only; rows N: 4, 6, and 7). And for the remaining three proteins, there were coordinates available only for the zymogenic form (i.e., “a” and “b” only; rows N: 8–10). In addition to these ten proteases from the pepsin-like family, three proteolytically nonfunctional proteins in one or two forms were also analyzed (rows N: 11–13). The proteolytic inactivity of the last three proteins is caused by the replacement of their catalytic aspartic acids in the C-domains with serine.

In SCOP, the retroviral protease (retropepsin) family is represented by the 3D structures of proteases from ten different organisms: HIV-1, HIV-2, HTLV-1, M-PMV, FIV, XMRV, SIV, RSV, MAV, and EIAV [5]. Of the ten proteases listed, only the 3D structure of the XMRV protease differs from that of the other retropepsins [22,23]. Therefore, only the 3D structures of HIV-1 and XMRV proteases in the free and ligand-bound forms were chosen for analysis (Table 1, rows 14 and 15).

The dimeric aspartyl protease family contains seven representative protein 3D structures [5]. Six of the seven representative proteins are homologues of the DNA damage-inducible protein 1 (Ddi1) protease (PDB ID: 4Z2Z) [24]. The fold of the seventh representative protein, RC1339/APRc from Rickettsia conorii (PDB ID: 5C9F), does not form the mandatory homodimer like all other proteins in the dimeric aspartyl protease family [25]. Therefore, two 3D structures from this family, Ddi1 and APRc, were taken for conformational analysis. Finally, the Lpg0085-like family contains only one representative 3D structure (PDB ID: 2PMA) [26] and it was included in the analysis.

2.2. Structural Catalytic Core around the Catalytic Aspartates in Pepsin

Let us consider three variants of the pepsin 3D structure: the zymogenic propepsin (PDB ID: 3PSG), free pepsin (PDB ID: 4PEP), and ligand-bound pepsin (PDB ID: 6XCZ), which structurally define the pepsin-like family (SCOP ID: 4000470) (Table 1, rows 1a–1f).

Table 1. Structural amino acid alignment of the structural catalytic core (SCC) in the acid proteases superfamily proteins.

N	PDB ID and Chain	R(Å)	Protein	EC: Number	Propept. or N-Term Pept.	DD-Link	D-Loop	G-Loop	Mediator	Ref.
Superfamily: acid proteases
Family: pepsin-like
1a	3PSG_A,p	1.65	Propepsin	EC:3.4.23.1	7p VRK 9p	11 DTEY 14	31 FDTGSS 36	121 LGLA 124	Y125	[27]
1b	3PSG_A	1.65	Propepsin	-׀׀-		188 GYW 190	214 VDTGTS 219	301 LGDV 304
1c	4PEP_A	1.80	Pepsin	-׀׀-	7 ENY 9	12 TEY 14	31 FDTGSS 36	121 LGLA 124	Y125	[28]
1d	4PEP_A	1.80	Pepsin	-׀׀-		188 GYW 190	214 VDTGTS 219	301 LGDV 304
1e	6XCZ_A	1.89	Pepsin	-׀׀-	7 ENY 9	12 TEY 14	31 FDTGSS 36	121 LGLA 124	Y125	[29]
1f	6XCZ_A	1.89	Pepsin	-׀׀-		188 GYW 190	214 VDTGTS 219	301 LGDV 304
2a	3VCM_A,p	2.93	Prorenin	EC:3.4.23.15	14p KRM 16p	11 DTQY 14	31 FDTGSS 36	121 VGMG 124	F125	[30]
2b	3VCM_A	2.93	Prorenin	-׀׀-		188 GVW 190	214 VDTGAS 219	301 LGAT 304
2c	2REN_A	2.50	Renin	-׀׀-	13 TNY 15	18 TQY 20	37 FDTGSS 42	128 VGMG 131	F132	[31]
2d	2REN_A	2.50	Renin	-׀׀-		199 GVW 201	225 VDTGAS 230	315 LGAT 318
2e	3K1W_A	1.50	Renin	-׀׀-	13 TNY 15	18 TQY 20	37 FDTGSS 42	128 VGMG 131	F132	[32]
2f	3K1W_A	1.50	Renin	-׀׀-		199 GVW 201	225 VDTGAS 230	315 LGAT 318
3a	1PFZ_A,p	1.85	Proplasmepsin 2	EC:3.4.23.39	85p KVE 87p	12 QNIM 15	33 LDTGSA 38	124 LGLG 127	W128	[33]
3b	1PFZ_A	1.85	Proplasmepsin 2	-׀׀-		191 LYW 193	213 VDSGTS 218	301 LGDP 304
3c	1LF4_A	1.90	Plasmepsin 2	-׀׀-	9 VDF 11	14 IMF 16	33 LDTGSA 38	124 LGLG 127	W128	[34]
3d	1LF4_A	1.90	Plasmepsin 2	-׀׀-		191 LYW 193	213 VDSGTS 218	301 LGDP 304
3e	2BJU_A	1.56	Plasmepsin 2	-׀׀-	9 VDF 11	14 IMF 16	33 LDTGSA 38	124 LGLG 127	W128	[35]
3f	2BJU_A	1.56	Plasmepsin 2	-׀׀-		191 LYW 193	213 VDSGTS 218	301 LGDP 304
4a	3QVC_A,p	2.10	HAP zymogen	EC:3.4.23.39	84p NIE 86p	9 LANVL 13	31 FHTASS 36	121 FGLG 124	W125	[36]
4b	3QVC_A	2.10	HAP zymogen	-׀׀-		188 LMW 190	214 LDSATS 219	301 LGDP 304
4e	3QVI_A,B	2.50	HAP protein	-׀׀-	7_B K	12 VLS 14	31 FHTASS 36	121 FGLG 124	W125	[36]
4f	3QVI_A	2.50	HAP protein	-׀׀-		188 LMW 190	214 LDSATS 219	301 LGDP 304
5a	5N7N_A,p	2.30	Procathepsin D	N/A	7p TRF 9p	37 DVVY 40	57 FDTGSA 62	147 LGLA 150	Y151	[37]
5b	5N7N_A	2.30	Procathepsin D	-׀׀-		217 GYW 219	248 ANTGTS 253	336 LGDV 339
5c	5N71_A	1.88	Cathepsin D	-׀׀-	33 VNL 35	38 VVY 40	57 FDTGSA 62	147 LGLA 150	Y151	[37]
5d	5N71_A	1.88	Cathepsin D	-׀׀-		217 GYW 219	248 ANTGTS 253	336 LGDV 339
5e	5N7Q_A	1.45	Cathepsin D	-׀׀-	11 VNL 13	16 VVY 18	35 FDTGSA 40	125 LGLA 128	Y129	[37]
5f	5N7Q_A	1.45	Cathepsin D	-׀׀-		195 GYW 197	226 ADTGTS 231	314 LGDV 317
6a	1MIQ_A,p	2.50	Proplasmepsin	N/A	84p KVE 86p	13 NIM 15	33 FDTGSA 38	124 LGLG 127	W128	[38]
6b	1MIQ_A	2.50	Proplasmepsin	-׀׀-		191 LYW 193	213 VDSGTT 218	301 LGDP 304
6e	1QS8_A	2.50	Plasmepsin	-׀׀-	9 DDV 11	14 IMF 16	33 FDTGSA 38	124 LGLG 127	W128	[38]
6f	1QS8_A	2.50	Plasmepsin	-׀׀-		191 LYW 193	213 VDSGTT 218	301 LGDP 304
7a	5JOD_A,p	1.53	Proplasmepsin 4	EC:3.4.23.39	85p KID 87p	13 NLM 15	33 FDTGSA 38	124 LGLG 127	W128	[39]
7b	5JOD_A	1.53	Proplasmepsin 4	-׀׀-		191 LYW 193	213 VDSGTS 218	301 LGDP 304
7e	1LS5_A	2.80	Plasmepsin 4	-׀׀-	9 DDV 11	14 LMF 16	33 FDTGSA 38	124 LGLG 127	W128	[34]
7f	1LS5_A	2.80	Plasmepsin 4	-׀׀-		191 LYW 193	213 VDSGTS 218	301 LGDP 304
8a	1QDM_A,p	2.30	Prophytepsin	EC:3.4.23.40	11p KKR 13p	15 NAQY 18	35 FDTGSS 40	126 LGLG 129	F130	[40]
8b	1QDM_A	2.30	Prophytepsin	-׀׀-		195 GYW 197	222 ADSGTS 227	313 LGDV 316
9a	1HTR_B,p	1.62	Progastricsin	EC:3.4.23.3	8p KKF 10p	11 DAAY 14	31 FDTGSS 36	121 MGLA 124	Y125	[41]
9b	1HTR_B	1.62	Progastricsin	-׀׀-		189 LYW 191	216 VDTGTS 221	304 LGDV 307
10a	1TZS_A,p	2.35	Procathepsin E	EC:3.4.23.34	9p R	22 DMEY 25	42 FDTGSS 47	132 LGLG 135	Y136	[42]
10b	1TZS_A	2.35	Procathepsin E	-׀׀-		201 AYW 203	227 VDTGTS 232	317 LGDV 320
11c	1T6E_X	1.70	Xylanase inhib.	EC:3.2.1.8	8 TKD 10	14 SLY 16	28 LDVAGP 33	141 AGLA 144	NS146	[43]
11d	1T6E_X	1.70	Xylanase inhib.	-׀׀-		204 PAH 206	234 LSTRLP 239	348 LGGA 351
11e	1T6G_A	1.80	Xylanase inhib.	-׀׀-	8 TKD 10	14 SLY 16	28 LDVAGP 33	141 AGLA 144	NS146	[43]
11f	1T6G_A	1.80	Xylanase inhib.	-׀׀-		204 PAH 206	234 LSTRLP 239	348 LGGA 351
12c	3AUP_A	1.91	Basic 7S globulin	N/A	15 QND 17	21 GLH 23	40 VDLNGN 45	159 AGLG 162	HA164	[44]
12d	3AUP_A	1.91	Basic 7S globulin	-׀׀-		228 GEY 230	264 ISTSTP 269	361 LGAR 364
13c	3VLA_A	0.95	EDGP (Fragment)	N/A	14 KKD 16	20 LQY 22	39 VDLGGR 44	155 AGLG 158	RT160	[45]
13d	3VLA_A	0.95	EDGP (Fragment)	-׀׀-		235 VEY 237	270 ISTINP 275	374 IGGH 377
13e	3VLB_A	2.70	EDGP (Fragment)	-׀׀-	14 KKD 16	20 LQY 22	39 VDLGGR 44	155 AGLG 158	RT160	[46]
13f	3VLB_A	2.70	EDGP (Fragment)	-׀׀-		235 VEY 237	270 ISTINP 275	374 IGGH 377
Family: retroviral protease (retropepsin)
14c	3IXO_A	1.70	HIV-1 protease	N/A	N/A	8 R-P 9	24 LDTGAD 29	85 IGRN 88	N/A	[46]
14d	3IXO_B	1.70	HIV-1 protease	-׀׀-	N/A	8 R-P 9	24 LDTGAD 29	85 IGRN 88	N/A
14e	5YOK_A	0.85	HIV-1 protease	-׀׀-	N/A	8 R-P 9	24 LDTGAD 29	85 IGRN 88	N/A	[47]
14f	5YOK_B	0.85	HIV-1 protease	-׀׀-	N/A	8 R-P 9	24 LDTGAD 29	85 IGRN 88	N/A
15c	3NR6_A	1.97	XMRV protease	EC:3.4.23.-	N/A	15 E-P 16	31 VDTGAQ 36	93 LGRD 96	R95	[22]
15d	3NR6_B	1.97	XMRV protease	-׀׀-	N/A	15 E-P 16	31 VDTGAQ 36	93 LGRD 96	R95
15e	3SLZ_A	1.40	XMRV protease	N/A	N/A	15 E-P 16	31 VDTGAQ 36	93 LGRD 96	R95	[48]
15f	3SLZ_B	1.40	XMRV protease	-׀׀-	N/A	15 E-P 16	31 VDTGAQ 36	93 LGRD 96	R95
Family: dimeric aspartyl proteases
16c	4Z2Z_A	1.80	Ddi1 protease	EC:3.4.23.-	N/A	201 VPML 204	219 VDTGAQ 224	289 IGLD 292	N/A	[49]
16d	4Z2Z_B	1.80	Ddi1 protease	-׀׀-	N/A	201 VPML 204	219 VDTGAQ 224	289 IGLD 292	N/A
17c	5C9F_A	2.00	ApRick protease	EC:3.-.-.-	N/A	121 DGHF 124	139 VDTGAS 144	209 LGMS 212	N/A	[25]
Family: LPG0085-like
18c	2PMA_A	1.89	Protein Lpg0085	N/A	N/A	29 Y	46 LDTGAK 51	145 LGRD 148	RD148	[26]
18d	2PMA_I	1.89	Protein Lpg0085	-׀׀-	N/A	29 Y	46 LDTGAK 51	145 LGRD 148	RD148

N/A—Not Available.

Table 1. Structural amino acid alignment of the structural catalytic core (SCC) in the acid proteases superfamily proteins.

N	PDB ID and Chain	R(Å)	Protein	EC: Number	Propept. or N-Term Pept.	DD-Link	D-Loop	G-Loop	Mediator	Ref.
Superfamily: acid proteases
Family: pepsin-like
1a	3PSG_A,p	1.65	Propepsin	EC:3.4.23.1	7p VRK 9p	11 DTEY 14	31 FDTGSS 36	121 LGLA 124	Y125	[27]
1b	3PSG_A	1.65	Propepsin	-׀׀-		188 GYW 190	214 VDTGTS 219	301 LGDV 304
1c	4PEP_A	1.80	Pepsin	-׀׀-	7 ENY 9	12 TEY 14	31 FDTGSS 36	121 LGLA 124	Y125	[28]
1d	4PEP_A	1.80	Pepsin	-׀׀-		188 GYW 190	214 VDTGTS 219	301 LGDV 304
1e	6XCZ_A	1.89	Pepsin	-׀׀-	7 ENY 9	12 TEY 14	31 FDTGSS 36	121 LGLA 124	Y125	[29]
1f	6XCZ_A	1.89	Pepsin	-׀׀-		188 GYW 190	214 VDTGTS 219	301 LGDV 304
2a	3VCM_A,p	2.93	Prorenin	EC:3.4.23.15	14p KRM 16p	11 DTQY 14	31 FDTGSS 36	121 VGMG 124	F125	[30]
2b	3VCM_A	2.93	Prorenin	-׀׀-		188 GVW 190	214 VDTGAS 219	301 LGAT 304
2c	2REN_A	2.50	Renin	-׀׀-	13 TNY 15	18 TQY 20	37 FDTGSS 42	128 VGMG 131	F132	[31]
2d	2REN_A	2.50	Renin	-׀׀-		199 GVW 201	225 VDTGAS 230	315 LGAT 318
2e	3K1W_A	1.50	Renin	-׀׀-	13 TNY 15	18 TQY 20	37 FDTGSS 42	128 VGMG 131	F132	[32]
2f	3K1W_A	1.50	Renin	-׀׀-		199 GVW 201	225 VDTGAS 230	315 LGAT 318
3a	1PFZ_A,p	1.85	Proplasmepsin 2	EC:3.4.23.39	85p KVE 87p	12 QNIM 15	33 LDTGSA 38	124 LGLG 127	W128	[33]
3b	1PFZ_A	1.85	Proplasmepsin 2	-׀׀-		191 LYW 193	213 VDSGTS 218	301 LGDP 304
3c	1LF4_A	1.90	Plasmepsin 2	-׀׀-	9 VDF 11	14 IMF 16	33 LDTGSA 38	124 LGLG 127	W128	[34]
3d	1LF4_A	1.90	Plasmepsin 2	-׀׀-		191 LYW 193	213 VDSGTS 218	301 LGDP 304
3e	2BJU_A	1.56	Plasmepsin 2	-׀׀-	9 VDF 11	14 IMF 16	33 LDTGSA 38	124 LGLG 127	W128	[35]
3f	2BJU_A	1.56	Plasmepsin 2	-׀׀-		191 LYW 193	213 VDSGTS 218	301 LGDP 304
4a	3QVC_A,p	2.10	HAP zymogen	EC:3.4.23.39	84p NIE 86p	9 LANVL 13	31 FHTASS 36	121 FGLG 124	W125	[36]
4b	3QVC_A	2.10	HAP zymogen	-׀׀-		188 LMW 190	214 LDSATS 219	301 LGDP 304
4e	3QVI_A,B	2.50	HAP protein	-׀׀-	7_B K	12 VLS 14	31 FHTASS 36	121 FGLG 124	W125	[36]
4f	3QVI_A	2.50	HAP protein	-׀׀-		188 LMW 190	214 LDSATS 219	301 LGDP 304
5a	5N7N_A,p	2.30	Procathepsin D	N/A	7p TRF 9p	37 DVVY 40	57 FDTGSA 62	147 LGLA 150	Y151	[37]
5b	5N7N_A	2.30	Procathepsin D	-׀׀-		217 GYW 219	248 ANTGTS 253	336 LGDV 339
5c	5N71_A	1.88	Cathepsin D	-׀׀-	33 VNL 35	38 VVY 40	57 FDTGSA 62	147 LGLA 150	Y151	[37]
5d	5N71_A	1.88	Cathepsin D	-׀׀-		217 GYW 219	248 ANTGTS 253	336 LGDV 339
5e	5N7Q_A	1.45	Cathepsin D	-׀׀-	11 VNL 13	16 VVY 18	35 FDTGSA 40	125 LGLA 128	Y129	[37]
5f	5N7Q_A	1.45	Cathepsin D	-׀׀-		195 GYW 197	226 ADTGTS 231	314 LGDV 317
6a	1MIQ_A,p	2.50	Proplasmepsin	N/A	84p KVE 86p	13 NIM 15	33 FDTGSA 38	124 LGLG 127	W128	[38]
6b	1MIQ_A	2.50	Proplasmepsin	-׀׀-		191 LYW 193	213 VDSGTT 218	301 LGDP 304
6e	1QS8_A	2.50	Plasmepsin	-׀׀-	9 DDV 11	14 IMF 16	33 FDTGSA 38	124 LGLG 127	W128	[38]
6f	1QS8_A	2.50	Plasmepsin	-׀׀-		191 LYW 193	213 VDSGTT 218	301 LGDP 304
7a	5JOD_A,p	1.53	Proplasmepsin 4	EC:3.4.23.39	85p KID 87p	13 NLM 15	33 FDTGSA 38	124 LGLG 127	W128	[39]
7b	5JOD_A	1.53	Proplasmepsin 4	-׀׀-		191 LYW 193	213 VDSGTS 218	301 LGDP 304
7e	1LS5_A	2.80	Plasmepsin 4	-׀׀-	9 DDV 11	14 LMF 16	33 FDTGSA 38	124 LGLG 127	W128	[34]
7f	1LS5_A	2.80	Plasmepsin 4	-׀׀-		191 LYW 193	213 VDSGTS 218	301 LGDP 304
8a	1QDM_A,p	2.30	Prophytepsin	EC:3.4.23.40	11p KKR 13p	15 NAQY 18	35 FDTGSS 40	126 LGLG 129	F130	[40]
8b	1QDM_A	2.30	Prophytepsin	-׀׀-		195 GYW 197	222 ADSGTS 227	313 LGDV 316
9a	1HTR_B,p	1.62	Progastricsin	EC:3.4.23.3	8p KKF 10p	11 DAAY 14	31 FDTGSS 36	121 MGLA 124	Y125	[41]
9b	1HTR_B	1.62	Progastricsin	-׀׀-		189 LYW 191	216 VDTGTS 221	304 LGDV 307
10a	1TZS_A,p	2.35	Procathepsin E	EC:3.4.23.34	9p R	22 DMEY 25	42 FDTGSS 47	132 LGLG 135	Y136	[42]
10b	1TZS_A	2.35	Procathepsin E	-׀׀-		201 AYW 203	227 VDTGTS 232	317 LGDV 320
11c	1T6E_X	1.70	Xylanase inhib.	EC:3.2.1.8	8 TKD 10	14 SLY 16	28 LDVAGP 33	141 AGLA 144	NS146	[43]
11d	1T6E_X	1.70	Xylanase inhib.	-׀׀-		204 PAH 206	234 LSTRLP 239	348 LGGA 351
11e	1T6G_A	1.80	Xylanase inhib.	-׀׀-	8 TKD 10	14 SLY 16	28 LDVAGP 33	141 AGLA 144	NS146	[43]
11f	1T6G_A	1.80	Xylanase inhib.	-׀׀-		204 PAH 206	234 LSTRLP 239	348 LGGA 351
12c	3AUP_A	1.91	Basic 7S globulin	N/A	15 QND 17	21 GLH 23	40 VDLNGN 45	159 AGLG 162	HA164	[44]
12d	3AUP_A	1.91	Basic 7S globulin	-׀׀-		228 GEY 230	264 ISTSTP 269	361 LGAR 364
13c	3VLA_A	0.95	EDGP (Fragment)	N/A	14 KKD 16	20 LQY 22	39 VDLGGR 44	155 AGLG 158	RT160	[45]
13d	3VLA_A	0.95	EDGP (Fragment)	-׀׀-		235 VEY 237	270 ISTINP 275	374 IGGH 377
13e	3VLB_A	2.70	EDGP (Fragment)	-׀׀-	14 KKD 16	20 LQY 22	39 VDLGGR 44	155 AGLG 158	RT160	[46]
13f	3VLB_A	2.70	EDGP (Fragment)	-׀׀-		235 VEY 237	270 ISTINP 275	374 IGGH 377
Family: retroviral protease (retropepsin)
14c	3IXO_A	1.70	HIV-1 protease	N/A	N/A	8 R-P 9	24 LDTGAD 29	85 IGRN 88	N/A	[46]
14d	3IXO_B	1.70	HIV-1 protease	-׀׀-	N/A	8 R-P 9	24 LDTGAD 29	85 IGRN 88	N/A
14e	5YOK_A	0.85	HIV-1 protease	-׀׀-	N/A	8 R-P 9	24 LDTGAD 29	85 IGRN 88	N/A	[47]
14f	5YOK_B	0.85	HIV-1 protease	-׀׀-	N/A	8 R-P 9	24 LDTGAD 29	85 IGRN 88	N/A
15c	3NR6_A	1.97	XMRV protease	EC:3.4.23.-	N/A	15 E-P 16	31 VDTGAQ 36	93 LGRD 96	R95	[22]
15d	3NR6_B	1.97	XMRV protease	-׀׀-	N/A	15 E-P 16	31 VDTGAQ 36	93 LGRD 96	R95
15e	3SLZ_A	1.40	XMRV protease	N/A	N/A	15 E-P 16	31 VDTGAQ 36	93 LGRD 96	R95	[48]
15f	3SLZ_B	1.40	XMRV protease	-׀׀-	N/A	15 E-P 16	31 VDTGAQ 36	93 LGRD 96	R95
Family: dimeric aspartyl proteases
16c	4Z2Z_A	1.80	Ddi1 protease	EC:3.4.23.-	N/A	201 VPML 204	219 VDTGAQ 224	289 IGLD 292	N/A	[49]
16d	4Z2Z_B	1.80	Ddi1 protease	-׀׀-	N/A	201 VPML 204	219 VDTGAQ 224	289 IGLD 292	N/A
17c	5C9F_A	2.00	ApRick protease	EC:3.-.-.-	N/A	121 DGHF 124	139 VDTGAS 144	209 LGMS 212	N/A	[25]
Family: LPG0085-like
18c	2PMA_A	1.89	Protein Lpg0085	N/A	N/A	29 Y	46 LDTGAK 51	145 LGRD 148	RD148	[26]
18d	2PMA_I	1.89	Protein Lpg0085	-׀׀-	N/A	29 Y	46 LDTGAK 51	145 LGRD 148	RD148

N/A—Not Available.

The boundary between the N- and C-domains of the 3D structure of pepsinogen is in the vicinity of Gly₁₆₉ [9]. Asp₃₂ (N-domain) and Asp₂₁₅ (C-domain) are the two catalytically important aspartate residues. Each aspartate residue is positioned within the hallmark Asp-Thr/Ser-Gly (Asp₃₂-Thr₃₃-Gly₃₄ in 3PSG) motif which, together with a further Hydrophobic-Hydrophobic-Gly sequence motif, forms an essential structural feature known as a psi-loop motif [28,50,51,52,53]. Let us designate two fragments of the protease amino acid sequence involved in formation of the psi-loop motif as the D(Asp)-loop and G(Gly)-loop. In this section, the atomic structure of the D- and G-loops in the N- and C-domains and their position relative to each other in the 3D structures of pepsin will be analyzed in detail.

2.2.1. Propepsin

DD-Zone of Propepsin: A D-Loop_N - DD-Link_N - D-Loop_C - DD-Link_C Circular Motif

As noted above, the functional activity of pepsin is carried out simultaneously by both of the catalytic residues, Asp₃₂ and Asp₂₁₅. Therefore, two D-loops, D-loop_N for the N-terminal domain and D-loop_C for the C-terminal domain, were analyzed in detail (Table 1 and Table S1). It was found that the two domains of propepsin also contain structurally equivalent short peptides, which we call DD-link_N (Asp₁₁-...-Tyr₁₄) and DD-link_C (Gly₁₈₈-Tyr₁₈₉-Trp₁₉₀), where N and C also stand for the N-terminal domain and C-terminal domain, respectively (Table 1). These two special DD-link peptides “lock” the ends of the D-loop_N and D-loop_C to form a “circular” structure, which altogether we call the “DD-zone” (Figure 1A).

The DD-zone of propepsin consists of 19 amino acids in total from both D-loops and both DD-links and an additional residue Tyr₁₂₅. Tyr₁₂₅ serves as a structural mediator between the C-terminus of the D-loop_N and the N-terminus of the DD-link_C (Figure 1A); this residue directly follows Ala₁₂₄ from G-loop_N (Table 1).

Independently, in propepsin, residues Thr₃₃ and Thr₂₁₆ are located next to the two catalytic aspartates. Their side-chain OG1 atoms each make two hydrogen bonds with main-chain nitrogen and oxygen atoms of the opposite D-loop (Figure 1A, Table S1, last column). These interactions are known as the “fireman’s grip” motif [54,55].

The proenzyme segment in propepsin is Leu_1p-...-Leu_44p, where “p” indicates the proenzyme sequence region. The pepsin portion in 3PSG starts from Ile₁. Glu₁₃ and Phe₁₅ form a short β-sheet-like interaction with Lys_9p and Val_7p (Figure 1A, Table S2, last column). The residues of this β-sheet undergo a conformational change during the activation process [9].

The Psi-Loop_N and Psi-Loop_C Motifs: Interactions between the D-Loop and G-Loop in the N- and C-Domains

In 3PSG, the D-loop_N tetrapeptide, Asp₃₂ -...- Ser_35, contains a frequently occurring Asx-motif [56], where an aspartate (here, catalytic Asp₃₂) or an asparagine residue within a tetra- or pentapeptide forms two short-range (in terms of sequence location) main-chain and side-chain hydrogen bonds with the sequentially adjacent amino acids (Figure 1B). We observe a similar Asx-motif involving the catalytic Asp₂₁₅ from the D-loop_C tetrapeptide (Figure 1C). Additionally, there are four conserved long-range hydrogen bonds between the D- and G-loops in both N- and C-domains (Figure 1B,C). We will refer to the substructures shown in Figure 1B,C as the psi-loop_N and psi-loop_C motifs. Each psi-loop motif is an eight-residue 3D structure consisting of D- and G-loop residues that are held together by six hydrogen bonds. The geometric characteristics of these six hydrogen bonds are given in Table S2 (row 1a, columns 4–6).

Comparison of the Psi-Loop_N and Psi-Loop_C

Despite the apparent similarity, the psi-loop_N and psi-loop_C motifs are not identical. While making similar interactions, the D-loop_C is five amino acids long (Asp₂₁₅-...-Ser₂₁₉) and the D-loop_N has only four residues (Figure 1B,C). Moreover, the conformations of the two respective G-loops differ. The G-loop_C at its C-terminus contains a β-turn, which is stabilized by the hydrogen bond between O/Gly₃₀₂ and N/Phe₃₀₅, while the G-loop_N does not have a similar substructure. As a result, there is conformational difference between Phe₃₀₅ and its structural counterpart in the N-domain, Tyr₁₂₅, where Phe₃₀₅ takes part in the conformational arrangement of its respective psi-loop, while Tyr₁₂₅ does not. Still, the two psi-loop motifs are bound by a set of equivalent interactions, where the O/Asp₃₂-N/Leu₁₂₃ hydrogen bond in psi-loop_N is substituted by the O/Thr₂₁₈-N/Asp₃₀₃ hydrogen bond in psi-loop_C, and where the O/Ser₃₅-N/Ala₁₂₄ hydrogen bond in psi-loop_N is substituted by the O/Ser₂₁₉-N/Val₃₀₄ hydrogen bond in psi-loop_C (Figure 1B,C).

The structural changes described above appear to result in tighter binding of Asp₃₂ to the G-loop_N than of Asp₂₁₅ to G-loop_C, since the distance from Asp₃₂ to G-loop_N is shorter than that from Asp₂₁₅ to G-loop_C. It is possible that this structural fact is the main reason for the differences in functional activity between Asp₃₂ and Asp₂₁₅ in the proposed models of catalytic hydrolysis of peptide bonds by acid proteases [57,58,59]. If Asp₃₂ is more tightly bound with more potential hydrogen bonds as compared to Asp₂₁₅, then its nucleophilicity must be somewhat decreased. Thus, Asp₂₁₅ of the C-domain would play a more prominent role in the proteolytic cleavage of dipeptide substrates than Asp₃₂ of the N-domain.

The structural association of two psi-loops and the DD-zone allows us to obtain an assembly of structural elements of the structural catalytic core (SCC) of propepsin (Figure 2A). It includes all 28 amino acids listed in Table 1 (rows 1a and 1b).

2.2.2. Activation of Free Pepsin

The conversion of propepsin to active pepsin is achieved through proteolytic cleavage and subsequent removal of the N-terminal amino acid fragment. Here, we are mostly interested in changes that occur in the propepsin structural core, SCC. A structural comparison of propepsin (PDB ID: 3PSG) and mature pepsin (PDB ID: 4PEP) showed that rearrangements occur only in DD-link_N and its immediate environment. First, as described above, the length of the tetrapeptide Asp₁₁-...-Tyr₁₄ was reduced by one residue at its N-terminus (Table 1 and Table S1). Then, the two-stranded β-sheet (Glu₁₃-...-Phe₁₅)/(Val_7p-...-Lys_9p) is replaced with a structurally similar two-stranded β-sheet (Glu₁₃-...-Phe₁₅)/(Glu₇-...-Tyr₉) (Table 1 and Table S2). Thus, upon pepsin activation the architecture of the SCC remains largely unchanged.

2.2.3. Pepsin/Ligand Complex

During activation, the propepsin structure transforms into the active pepsin structure, ligand-free form. How does interaction with the ligand affect the SCC? Let us consider the 3D structure of the pepsin/saquinavir complex (PDB ID: 6XCZ). The key contacts between pepsin and the small-molecule ligand (saquinavir, ROC₄₀₁) are four hydrogen bonds (Figure 2B; Table S3, rows 1e and 1f). Two pairs of conserved residues from the D-loops of the N- and C-domains, Asp₃₂/Gly₃₄ and Asp₂₁₅/Gly₂₁₇, donate four oxygen atoms as part of the four hydrogen bonds. Each of the two aspartates forms an Asx-motif [56], and in addition to the four hydrogen bonds above, there are two additional hydrogen bonds via the mediator-waters HOH₅₂₇ and HOH₆₄₅ (Figure 2B), and also there is a hydrogen bond that involves the OH atom of Tyr₁₈₉, the central residue of the tripeptide DD-link_C. Thus, DD-link_C interacts with the inhibitor. Aside from the extensive hydrogen bonding inventory described above, binding of a ligand does not introduce any visible structural changes to the ligand-free form of the SCC of pepsin (Tables S1 and S2, rows 1c–1f).

The location of the structural catalytic core (SCC) in the 3D structure of propepsin is shown in Figure 3.

2.3. Structural Core in Proteins of the Pepsin-like Family

2.3.1. DD-Zones

Earlier, we showed that in propepsin the segment Asp₁₁-Phe₁₅, which includes DD-link_N, interacts with the pro-tripeptide Val_7p-Lys_9p (Figure 1A) by means of interactions listed in Table S2. During the transition from the inactive zymogenic form to the enzymatically active form, DD-link_N is slightly structurally modified as described above, and the pro-tripeptide is spatially substituted by the N-terminal tripeptide (Glu₇-Tyr₉; Table 1). Interactions between DD-link_N and the N-terminal tripeptide are shown in Table S2. We also observed similar structural rearrangements in the other members of the pepsin-like family although there are variations from the rule: with the histo-aspartic protease (HAP), DD-link_N is one amino acid longer, and with procathepsin E, only one amino acid, R_9P, of the propeptide, contacts DD-link_N (Table 1). However, the general structural trend for the pepsin-like family is the same.

In propepsin and pepsin, the contact between DD-link_N and D-loop_N involves a water molecule as an intermediary (Figure 1; Table S1). In the structure of ligand-bound pepsin, a water molecule does not participate in interactions as an intermediary. A similar water presence and functionality is observed for all of the remaining proteins of the pepsin-like family. However, considering differences in the resolution of structures (Table 1) and the associated difficulties in localization of the bound water molecules, it is not always possible to unambiguously correlate the presence or absence of a water molecule with any form of protein, and thus exceptions are possible.

In pepsin, the contact between D-loop_N and DD-link_C involves the amino acid Tyr₁₂₅ as a structural mediator (Figure 1; Table S1). In a number of proteins, there is also a mediating water molecule in addition to the aromatic amino acid (Table S1, column 5). In three proteins, xylanase inhibitor, basic 7S globulin, and EDGP, there are two mediator residues instead of a single Tyr₁₂₅. A hydrogen bond between the ends of DD-link_C and D-loop_C is, however, conserved and contains no mediator insertions in any of the analyzed structures (Table S1, column 6). The contact between D-loop_C and DD-link_N does not contain mediators, but can be variable in its nature, being a hydrogen bond, a weak hydrogen bond, or a hydrophobic interaction (Table S1, column 7).

Fireman’s Grip Motif Reflects Open/Close-Conformation Structural Change

In the pepsin-like family proteins, the open/close-conformation structural change during the transition from the inactive zymogen to the enzymatically active form can either lead to conformational changes in the DD-zone or not. In proteins, where the hallmark Asp-Thr/Ser-Gly sequence (see Section 2.2) in the C-terminal domain contains serine, the conformational change in the DD-zone does take place, and it is reflected by the change of the fireman’s grip motif (Table S1, column 8). In proteins, where the hallmark Asp-Thr/Ser-Gly sequence in the C-terminal domain contains threonine, the open/close conformational change in the DD-zone does not take place.

2.3.2. Psi-Loops

As noted above, the psi-loop motif includes amino acids from the D- and G-loops. In pepsin, both D-loops contain a catalytic aspartate. Of the thirteen proteins studied, eight are active hydrolases, and have both catalytic aspartates (Table 1). In the HAP protein, an evolutionary Asp₃₂His mutation did occur, which, however, did not lead to a loss of catalytic activity because the other Asp₂₁₅ was still present [36]. The remaining four proteins, cathepsin D, xylanase inhibitor, basic 7S globulin, and EDGP, lost their enzymatic activity due to the replacement of the catalytic aspartate with another amino acid in the C-terminal domain [37,43,44,45]. Loss of catalytic activity in these proteins versus the HAP protein is strong evidence that proteolytic activity requires the aspartate of the C-terminal domain, whereas the aspartate of the N-terminal domain may be dispensable.

Both psi-loop_N and psi-loop_C motifs are structurally identical among the thirteen proteins of the pepsin-like family in three different forms (proenzyme, mature enzyme, and enzyme/ligand complex) (Table S2, columns 4 and 5). That is, replacing the catalytic aspartate with another amino acid either does not affect the conformation of the psi-loop motifs or affects it insignificantly. Structural conservation of the psi-loop conformation also occurs despite structural rearrangement in the tetrapeptides forming the Asx-motif in some proteins (Table S2, column 6). For example, six proteins in one or several forms show a structural transition from the Asx-motif to a Asx-turn [60], which lacks the hydrogen bond between the atoms of the first and fourth residues of the tetrapeptide unlike the Asx-motif. The structures of these six proteins, the HAP protein, plasmepsin 4, phytepsin, xylanase inhibitor, basic 7S globulin, and EDGP, have geometrical parameters that formally exceed those of a canonical hydrogen bond [61].

2.3.3. Ligand Bound Pepsin-like Proteins

Section 2.2.3 identifies seven amino acids of the pepsin’s SCC that are responsible for ligand recognition. These are (1, 2, 3 and 4) catalytic Asp/Gly pairs of (Asp-Thr/Ser-Gly)_N and (Asp-Thr/Ser-Gly)_C, C-terminal and N-terminal Asp-Thr/Ser-Gly motifs; (5 and 6) two C-terminal serine residues of D-loop_N and D-loop_C; and (7) the Tyr₁₈₉, the central residue of the tripeptide DD-link_C. Of the thirteen pepsin-like representative structures listed in Table 1, only seven had a complex with a ligand close to or within the SCC. Six of these seven structures had similar D-loop/ligand contacts (Table S3). And, again, the HAP protein was unique, by lacking the expected contacts of Ala₂₁₇ and Ser₂₁₉ with the K95 inhibitor as seen in all of the other structures. With the HAP protein, instead of those contacts, Ala₂₁₇ and Ser₂₁₉ of chain_A formed hydrogen bonds with Asn₂₇₉ of chain_B, i.e., O/Ala_{217_A} - N/Asn_{279_B} at 2.9 Å and OG/S₂₁₉-ND2/N_{279_B} at 3.1 Å, respectively, and a weak hydrogen bond with Glu_278A of chain_B (designated as Glu_{278A_B} in the PDB file of 3QVI), O/Ala_{217_A} - CA/Glu_{278A_B} at 3.4 (2.6) 127° (for the definition of parameters of weak hydrogen bonds, see [15]). The changes in contact partners for Ala₂₁₇ and Ser₂₁₉ are due to the fact that in the inhibitor complex the enzyme forms a tight domain-swapped dimer, not previously seen in any aspartic protease [36]. As a result of such domain-swapped dimerization, Glu_278A of chain_B forms contacts with the inhibitor instead of Ala₂₁₇ and Ser₂₁₉ of chain_A (Table S3, row 4f and column 5).

Taken together, the pepsin-like family proteins from Table 1 have their SCC constructed from the same set of conserved amino acids in all three forms, i.e., proenzyme, ligand-free enzyme, and ligand-bound enzyme, while the most noticeable structural changes concern the transition of the DD-links and fireman’s grips from the zymogenic form to the enzymatic form. The DD-zones include the N-terminal and C-terminal D-loops, D-loop_N and D-loop_C, with their ends linked by the longer DD-link_N and a water molecule, and a shorter DD-link_C plus a mediator molecule (Figure 1A).

2.4. SCC in Hydrolases of the Retroviral Protease (Retropepsin) Family

2.4.1. DD-Zones

The retroviral protease (retropepsin) family is the second family of acid proteases listed in Table 1. Hydrolases of this family do not have a zymogenic form, and the enzyme is a dimer of two identical amino acid chains. Figure 4A shows a DD-zone of HIV-1 protease (PDB ID: 3IXO). The main differences between the DD-zones of pepsin and HIV-1 are the number of residues forming DD-links and an absence of mediators.

A change in the number of residues in the DD-links is usually associated with the presence or absence of the need to form a β-structural contact with either the propeptide or the N-terminal fragment (Figure 4A vs. Figure 1A). However, a decrease in the length of the DD-link by one amino acid does not necessarily lead to a change in the relative position of the D-loops relative to each other. Such is the case for the HIV-1 protease, where atoms of the long side-chain of Arg₈ (DD-link in HIV-1) interact with Asp₂₉ (D-loop in HIV-1) instead of the oxygen atoms of the shorter side-chains of Asp₁₁ (DD-link in pepsin) and Ser₂₁₉ (D-loop in pepsin) (Figure 4A vs. Figure 1A, Table S1).

In the XMRV protease (PDB ID: 3NR6), there is glutamate (DD-link in XMRV) in place of Arg₈ (DD-link in HIV-1) and glutamine (D-loop in XMRV) instead of Asp₂₉ (D-loop in HIV-1) (Table 1), which results in some changes in the architecture of the DD-zone in the XMRV protease compared to HIV-1 (Figure 4B, Table S1). In XMRV, there is an increase in the distance between the ends of the DD-link and the D-loop, which results in the absence of a direct contact between them. However, in XMRV, the D-loop/DD-link contact happens through the mediator residue Arg₉₅, which also participates in the formation of the psi-loop (Figure 4B).

Thus, the distinctive feature of the retroviral protease (retropepsin) family hydrolases is within the DD-zones, where the D-loops are bound by short DD-links of two residues plus a mediator residue. Additionally, in HIV-1 and XMRV, there is a separate residue Arg₈₇ (in HIV-1)/Arg₉₅ (in XMRV), which interacts with Asp₂₉ (in HIV-1)/Gln₃₆ (in XMRV) via a conventional hydrogen bond: NH2/R₈₇-OD1/D₂₉ (Table S1, column 5), and stabilizes the conformation of the D-loop. The function of this residue in HIV-1 and XMRV is unknown.

2.4.2. Psi-Loops in HIV-1 and XMRV

As noted above, a homodimer of two identical amino acid chains is the active form of a HIV-1 protease. Therefore, one can expect the conformation of the psi-loop motif in chains A and B to be identical. It was found out that HIV-1 and XMRV not only have similar psi-loop motifs, but they are also similar to that observed in the C-domain of pepsin (Figure 1C and Figure 4C). That is, the identical psi-loops in HIV-1 and XMRV have chosen a conformation that provides a catalytic aspartate with higher proteolytic efficiency in both subunits (Table S2). In Table S2, homodimer chains A and B in HIV-1 (and other retroproteases) are listed as the respective counterparts of the N- and C-domains in pepsin, but this is an arbitrary assignment.

2.4.3. Ligand-Bound Forms of Retroviral Proteases

The DD-zones of ligand-bound pepsin and HIV-1 are very similar to each other (Figure 2B and Figure 4D). The main interactions are made by the three amino acids from each of the two D-loops, totaling six interacting residues (Table S3). In HIV-1, these residues are Asp₂₅, Gly₂₇, and Asp₂₉ from D-loop of chain A and, of course, identical residues are in D-loop of chain_B of the HIV-1 homodimer (Figure 4D). For comparison, in pepsin, those amino acids are Asp₃₂, Gly₃₄, and Ser₃₆ from D-loop_N and Asp₂₁₅, Gly₂₁₇, and Ser₂₁₉ from D-loop_C (Table S3). In addition, with pepsin, Section 2.2.3 describes the additional Tyr₁₈₉ from the DD-link_C that is involved in contacts with the ligand. In the ligand-bound HIV-1 protease (PDB ID: 5YOK), a combination of Arg₈ (DD-link)/Asp₂₉ (D-loop) performs an analogous role. Similar to HIV-1, in the ligand-bound XMRV (PDB ID: 3SLZ), the C-terminal position of the D-loop, Gln₃₆, also participates in ligand binding (Table S3, last column). Replacing Asp₂₉ (in HIV-1) with Gln₃₆ (in XMRV) also results in additional hydrogen bonds formed between XMRV and the inhibitor. Interaction with the ligand does not seem to affect the architecture of the DD-zone in the HIV-1 and XMRV proteases (Table S1).

The X-ray structure of the retroviral HIV-1 protease (Figure 4D) shows an identical mode of interaction between two catalytic aspartates, Asp₂₅ of chain_A and _B, and the bound ligand. However, if we take into account additional neutron crystallography data, we find that the catalytic aspartates are not identical in terms of their protonation state [62,63]. According to these data, one aspartate is protonated and the other is deprotonated at physiological pH. As a result, the two catalytic aspartates do interact differently with the same ligand. The deprotonated aspartate uses one of its deprotonated side-chain oxygens to interact with the hydrogen bound to the O2 atom of the ligand. At the same time, the protonated aspartate uses its protonated side-chain oxygen to interact directly with the same O2 oxygen atom of the ligand. These additional experimental data show the different roles that these two aspartates play in the catalytic mechanism of the HIV-1 protease.

The SCCs of the HIV-1 and XMRV proteases are shown in Figure 5A,B.

The location of the structural catalytic core (SCC) in the 3D structure of HIV-1 protease is shown in Figure 6.

2.5. SCCs of the Dimeric Aspartyl Proteases and Lpg0085-like Family Proteins

In HIV-1 and XMRV, we have shown how amino acid changes at the N-terminus of the DD-link and the C-terminus of the D-loop affect the structure of the DD-zone. The Ddi1 protease, like the XMRV protease, has glutamine as the C-terminal amino acid of the D-loop (Table 1 and Table S1, rows 16c and 16d). However, the DD-links of the Ddi1 and XMRV proteases differ in length. In Ddi1, the number of amino acids in the DD-link increases twofold (from 2 to 4 residues) compared to XMRV protease, while in Lpg0085 the DD-link is a single residue (Figure 7A,B; Table 1 and Table S1, rows 18c and 18d). To compensate for such a reduction in the DD-link length in Lpg0085, a mediator dipeptide Arg₁₄₇-Asp₁₄₈ is additionally present for DD-zone formation. Thus, the DD-zones of the dimeric aspartyl proteases and the Lpg0085-like proteins are characterized by the presence of either a longer DD-link of four residues or a shorter DD-link of one residue plus a separate two-residue mediator.

As in the case of retroviral proteases, Ddi1 and Lpg0085 use the psi-loop_C motif, which is equivalent to the C-terminal version of the psi-loop motif in pepsin-like family proteins (Table 1 and Table S2, rows 16c, 16d, 18c and 18d). The ApRick protease does not form a canonical dimer, as do Ddi1 and Lpg0085 [25]. However, the psi-loop in the ApRick protease monomer is still identical to that in Ddi1 and Lpg0085 (Figure 5C; Table 1 and Table S2, row 17c). Li et al. suggested that the ApRick protease “may represent a putative common ancestor of monomeric and dimeric aspartic proteases” [25]. The SCCs in Ddi1 and Lpg0085 are shown in Figure 8A,B.

3. Conclusions

Here, we have outlined the minimal conserved structural arrangement common to the acid protease superfamily of proteins, which we refer to as the structural catalytic core (SCC). We began with the pepsin-like family proteases, where we defined the DD-zone (Figure 1A). The DD-zone is a circular structural motif defined by substructures around the catalytic aspartates in the N- and C-terminal domains, D-loop_N and D-loop_C, and their interactions with the peptides DD-link_N and DD-link_C, which join the ends of D-loop_N and D-loop_C. Then, we increased the common substructure by defining the psi-loop_N and psi-loop_C motifs, where the DD-zone interacts through their D-loops with two external tetrapeptides, G-loop_N and G-loop_C, the residues of which intersect with the Hydrophobic-Hydrophobic-Gly sequence motif [51] (Figure 1B,C). While the two psi-loop motifs use the same logic in their formation, they differ in the environment around the catalytic aspartates, which may determine their different functional roles. Taken together, the psi-loops and the DD-zone define structural boundaries of the SCC in pepsin-like proteins.

The other families of acid proteases, retroviral proteases (retropepsin), dimeric aspartyl proteases, and Lpg0085-like proteins, also have the DD-zone and psi-loop substructures similar to pepsin. However, unlike pepsin—which can be very roughly described as a “hetero psi-loop” protein, where psi-loop_N and psi-loop_C are not structurally identical unlike the homodimer enzymes, with the psi-loop_C being more functionally active—the retroviral proteases, dimeric aspartyl proteases, and Lpg0085-like proteins can be described as having a “homo psi-loop” since they have two identical chains. The homo psi-loops are both structurally similar to psi-loop_C of pepsin. As with the pepsin-like proteases, the other three protein families use DD-links to form a DD-zone (Table 1). If a DD-link is equal to or shorter than two amino acids, there are additional mediator residues or water molecules filling the gap. Some mediator residues are located in sequence either at the C-terminus of the G-loop or immediately after it. Based on the structures seen so far, we can argue that a specific “long DD-link”, or “DD-link + mediator” or “DD-link + water” combination, is the same for a structural family within an acid protease superfamily, and may distinguish that family from the other proteins.

In summary, we can say that the SCC of the acid protease superfamily proteins consists of a dimer composed of a DD-link, D-loop, and G-loop blocks, where the D-loop plus DD-link forms a DD-zone, and the dimer of D- and G-loops forms two psi-loops. Defining the SCC in this way allows us to outline a minimal common substructure for the entire superfamily of proteins, such as acid proteases. This substructure combines amino acid conservation and protein functionality, which together can be used for protein comparison, structure identification, protein family separation, and protein engineering.