Characterizing high-affinity antigen/antibody complexes by kinetic- and equilibrium-based methods

Two biophysical methods, Biacore and KinExA, were used to kinetically and thermodynamically characterize high-affinity antigen/antibody complexes. Three to five independent experiments were performed on each platform with three different antigen/antibody complexes possessing nanomolar to picomolar equilibrium dissociation constants. By monitoring the dissociation phase on Biacore for 4 h, we were able to measure dissociation rate constants (kd) on the order of 1 x 10(-5)s(-1). To characterize high-affinity interactions by KinExA, samples needed to be equilibrated for up to 35 h to reach equilibrium. In the end, we show that similar kinetic rate constants and affinities were determined with both solution-phase and solid-phase methodologies. These results help further validate both interaction technologies and illustrate their suitability for characterizing extremely high-affinity interactions.     

Thermodynamic stability of the bacteriorhodopsin lattice as measured by lipid dilution

To determine the strength of noncovalent interactions that stabilize a membrane protein complex, we have developed an in vitro method for quantifying the dissociation of the bacteriorhodopsin (BR) lattice, a naturally occurring two-dimensional crystal. A lattice suspension was titrated with a short- and long-chain phosphatidylcholine mixture to dilute BR within the lipid bilayer. The fraction of BR in the lattice form as a function of added lipid was determined by visible circular dichroism spectroscopy and fit with a cooperative self-assembly model to obtain a critical concentration for lattice assembly. Critical concentration values of wild-type and mutant proteins were used to calculate the change in lattice stability upon mutation (DeltaDeltaG). By using this method, a series of mutant proteins was examined in which residues at the BR-BR interface were replaced with smaller amino acids, either Ala or Gly. Most of the mutant lattices were destabilized, with DeltaDeltaG values of 0.2-1.1 kcal/mol at 30 degrees C, consistent with favorable packing of apolar residues in the membrane. One mutant, I45A, was stabilized by approximately 1.0 kcal/mol, possibly due to increased lipid entropy. The DeltaDeltaG values agreed well with previous in vivo measurements, except in the case of I45A. The ability to measure the change in stability of mutant protein complexes in a lipid bilayer may provide a means of determining the contributions of specific protein-protein and protein-lipid interactions to membrane protein structure.     

Structural and computational characterization of the SHV-1 beta-lactamase-beta-lactamase inhibitor protein interface

Beta-lactamase inhibitor protein (BLIP) binds a variety of class A beta-lactamases with affinities ranging from micromolar to picomolar. Whereas the TEM-1 and SHV-1 beta-lactamases are almost structurally identical, BLIP binds TEM-1 approximately 1000-fold tighter than SHV-1. Determining the underlying source of this affinity difference is important for understanding the molecular basis of beta-lactamase inhibition and mechanisms of protein-protein interface specificity and affinity. Here we present the 1.6A resolution crystal structure of SHV-1.BLIP. In addition, a point mutation was identified, SHV D104E, that increases SHV.BLIP binding affinity from micromolar to nanomolar. Comparison of the SHV-1.BLIP structure with the published TEM-1.BLIP structure suggests that the increased volume of Glu-104 stabilizes a key binding loop in the interface. Solution of the 1.8A SHV D104K.BLIP crystal structure identifies a novel conformation in which this binding loop is removed from the interface. Using these structural data, we evaluated the ability of EGAD, a program developed for computational protein design, to calculate changes in the stability of mutant beta-lactamase.BLIP complexes. Changes in binding affinity were calculated within an error of 1.6 kcal/mol of the experimental values for 112 mutations at the TEM-1.BLIP interface and within an error of 2.2 kcal/mol for 24 mutations at the SHV-1.BLIP interface. The reasonable success of EGAD in predicting changes in interface stability is a promising step toward understanding the stability of the beta-lactamase.BLIP complexes and computationally assisted design of tight binding BLIP variants.     

Computational redesign of the SHV-1 beta-lactamase/beta-lactamase inhibitor protein interface

β-lactamases are enzymes that catalyze the hydrolysis of β-lactam antibiotics. β-lactamase/β-lactamase inhibitor protein (BLIP) complexes are emerging as a well characterized experimental model system for studying protein-protein interactions. BLIP is a 165 amino acid protein that inhibits several class A β-lactamases with a wide range of affinities: picomolar affinity for K1; nanomolar affinity for TEM-1, SME-1, and BlaI; but only micromolar affinity for SHV-1 β-lactamase. The large differences in affinity coupled with the availability of extensive mutagenesis data and high-resolution crystal structures for the TEM-1/BLIP and SHV-1/BLIP complexes make them attractive systems for the further development of computational design methodology. We used EGAD, a physics-based computational design program, to redesign BLIP in an attempt to increase affinity for SHV-1. Characterization of several of designs and point mutants revealed that in all cases, the mutations stabilize the interface by 10- to 1000-fold relative to wild type BLIP. The calculated changes in binding affinity for the mutants were within a mean absolute error of 0.87 kcal/mol from the experimental values, and comparison of the calculated and experimental values for a set of 30 SHV-1/BLIP complexes yielded a correlation coefficient of 0.77. Structures of the two complexes with the highest affinity, SHV-1/BLIP (E73M) and SHV-1/BLIP (E73M, S130K, S146M), are presented at 1.7 Å resolution. While the predicted structures have much in common with the experimentally determined structures, they do not coincide perfectly; in particular a salt bridge between SHV-1 D104 and BLIP K74 is observed in the experimental structures, but not in the predicted design conformations. This discrepancy highlights the difficulty of modeling salt bridge interactions with a protein design algorithm that approximates side chains as discrete rotamers. Nevertheless, while local structural features of the interface were sometimes miscalculated, EGAD is globally successful in designing complexes with increased affinity.     

Divalent metal dependence of site-specific DNA binding by EcoRV endonuclease.

Measurements of binding equilibria of EcoRV endonuclease to DNA, for a series of base-analogue substrates, demonstrate that expression of sequence selectivity is strongly enhanced by the presence of Ca2+ ions. Binding constants were determined for short duplex oligodeoxynucleotides containing the cognate DNA site, three cleavable noncognate sites, and a fully nonspecific site. At pH 7.5 and 100 mM NaCl, the full range of specificity from the specific (tightest binding) to nonspecific (weakest binding) sites is 0.9 kcal/mol in the absence of metal ions and 5.8 kcal/mol in the presence of Ca2+. Precise determination of binding affinities in the presence of the active Mg2+ cofactor was found to be possible for substrates retaining up to 1.6% of wild-type activity, as determined by the rate of phosphoryl transfer. These measurements show that Ca2+ is a near-perfect analogue for Mg2+ in binding reactions of the wild-type enzyme with DNA base-analogue substrates, as it provides identical DeltaDeltaG degrees bind values among the cleavable noncognate sites. Equilibrium dissociation constants of wild-type and base-analogue sites were also measured for the weakly active EcoRV mutant K38A, in the presence of either Mg2+ or Ca2+. In this case, Ca2+ allows expression of a greater degree of specificity than does Mg2+. DeltaDeltaG degrees bind values of K38A toward specific versus nonspecific sites are 6.1 kcal/mol with Ca2+ and 3.9 kcal/mol with Mg2+, perhaps reflecting metal-specific conformational changes in the ground-state ternary complexes. The enhancement of binding specificity provided by divalent metal ions is likely to be general to many restriction endonucleases and other metal-dependent nucleic acid-modifying enzymes. These results strongly suggest that measurements of DNA binding affinities for EcoRV, and likely for many other restriction endonucleases, should be performed in the presence of divalent metal ions.     

Halothane, an inhalational anesthetic agent, increases folding stability of serum albumin

Inhalational anesthetic agents are known to alter protein function, but the nature of the interactions underlying these effects remains poorly understood. We have used differential scanning calorimetry to study the effects of the anesthetic agent halothane on the thermally induced unfolding transition of bovine serum albumin. We find that halothane (0.6-10 mM) stabilizes the folded state of this protein, increasing its transition midpoint temperature from 62 to 71 degrees C. Binding of halothane to the native state of serum albumin thus outweighs any non-specific interactions between the thermally unfolded state of serum albumin and halothane in this concentration range. Based on the average enthalpy change DeltaH for unfolding of 170 kcal/mol, the increase from 62 to 71 degrees C corresponds to an additional Gibbs energy of stabilization (DeltaDeltaG) due to halothane of more than 4 kcal/mol. Analysis of the dependence of DeltaDeltaG on halothane concentration shows that thermal unfolding of a bovine serum albumin molecule is linked to the dissociation of about one halothane molecule at lower halothane concentrations and about six at higher halothane concentrations. Serum albumin is the first protein that has been shown to be stabilized by an inhalational anesthetic.     

Crystal structural analysis of protein-protein interactions drastically destabilized by a single mutation

The complex of barnase (bn) and barstar (bs), which has been widely studied as a model for quantitative analysis of protein-protein interactions, is significantly destabilized by a single mutation, namely, bs Asp39 --> Ala, which corresponds to a change of 7.7 kcal x mol(-1) in the free energy of binding. However, there has been no structural information available to explain such a drastic destabilization. In the present study, we determined the structure of the mutant complex at 1.58 A resolution by X-ray crystallography. The complex was similar to the wild-type complex in terms of overall and interface structures; however, the hydrogen bond network mediated by water molecules at the interface was significantly different. Several water molecules filled the cavity created by the mutation and consequently caused rearrangement of the hydrated water molecules at the interface. The water molecules were redistributed into a channel-like structure that penetrated into the complex. Furthermore, molecular dynamics simulations showed that the mutation increased the mobility of water molecules at the interface. Since such a drastic change in hydration was not observed in other mutant complexes of bn and bs, the significant destabilization of the interaction may be due to this channel-like structure of hydrated water molecules.     

LexA repressor forms stable dimers in solution. The role of specific dna in tightening protein-protein interactions

Cooperativity in the interactions among proteins subunits and DNA is crucial for DNA recognition. LexA repressor was originally thought to bind DNA as a monomer, with cooperativity leading to tighter binding of the second monomer. The main support for this model was a high value of the dissociation constant for the LexA dimer (micromolar range). Here we show that the protein is a dimer at nanomolar concentrations under different conditions. The reversible dissociation of LexA dimer was investigated by the effects of hydrostatic pressure or urea, using fluorescence emission and polarization to monitor the dissociation process. The dissociation constant lies in the picomolar range (lower than 20 pM). LexA monomers associate with an unusual large volume change (340 ml/mol), indicating the burial of a large surface area upon dimerization. Whereas nonspecific DNA has no stabilizing effect, specific DNA induces tightening of the dimer and a 750-fold decrease in the K(d). In contrast to the previous model, a tight dimer rather than a monomer is the functional repressor. Accordingly, the LexA dimer only loses its ability to recognize a specific DNA sequence by RecA-induced autoproteolysis. Our work provides insights into the linkage between protein-protein interactions, DNA recognition, and DNA repair.     

SUPREX (Stability of Unpurified Proteins from Rates of H/D Exchange) analysis of the thermodynamics of synergistic anion binding by ferric-binding protein (FbpA), a bacterial transferrin

SUPREX (stability of unpurified proteins from rates of H/D exchange) is a H/D exchange- and matrix-assisted laser desorption/ionization (MALDI)-based technique for characterizing the equilibrium unfolding/refolding properties of proteins and protein-ligand complexes. Here, we describe the application of SUPREX to the thermodynamic analysis of synergistic anion binding to iron-loaded ferric-binding protein (Fe(3+)FbpA-X, X = synergistic anion). The in vivo function of FbpA is to transport unchelated Fe(3+) across the periplasmic space of certain Gram-negative bacteria, a process that requires simultaneous binding of a synergistic anion. Our results indicate that Fe(3+)FbpA-X is not a so-called "ideal" protein system for SUPREX analyses because it does not exhibit two-state folding properties and it does not exhibit EX2 H/D exchange behavior. However, despite these nonideal properties of the Fe(3+)FbpA-X protein-folding/unfolding reaction, we demonstrate that the SUPREX technique is still amenable to the quantitative thermodynamic analysis of synergistic anion binding to Fe(3+)FbpA. As part of this work, the SUPREX technique was used to evaluate the DeltaDeltaG(f) values of four synergistic anion-containing complexes of Fe(3+)FbpA (i.e., Fe(3+)FbpA-PO(4), Fe(3+)FbpA-citrate, Fe(3+)FbpA-AsO(4), and Fe(3+)FbpA-SO(4)). The DeltaDeltaG(f) value obtained for Fe(3+)FbpA-citrate relative to Fe(3+)FbpA-PO(4) (1.45 +/- 0.44 kcal/mol), is in good agreement with that reported previously (1.98 kcal/mol). The value obtained for Fe(3+)FbpA-AsO(4) (0.58 +/- 0.45 kcal/mol) was also consistent with that reported previously (0.68 kcal/mol), but the measurement error is very close to the magnitude of the value. This work (i) demonstrates the utility of the SUPREX method for studying anion binding by FbpA, (ii) provides the first evaluation of a DeltaDeltaG(f) value for Fe(3+)FbpA-SO(4), -1.43 +/- 0.17 kcal/mol, and (iii) helps substantiate our hypothesis that the synergistic anion plays a role in controlling the lability of iron bound to FbpA in the transport process.     

On the dynamic nature of the transition state for protein-protein association as determined by double-mutant cycle analysis and simulation

The process of protein-protein association starts with their random collision, which may develop into an encounter complex followed by a transition state and final complex formation. Here we aim to experimentally characterize the nature of the transition state of protein-protein association for three different protein-protein interactions; Barnase-Barstar, TEM1-BLIP and IFNalpha2-IFNAR2, and use the data to model the transition state structures. To model the transition state, we determined inter-protein distance-constraints of the activated complex by using double mutant cycles (DMC) assuming that interacting residues are spatially close. Significant DeltaDeltaG(double dagger)(int) values were obtained only between residues on Barnase and Barstar. However, introducing specific mutations that optimize the charge complementarity between BLIP and TEM1 resulted in the introduction of significant DeltaDeltaG(double dagger)(int) values also between residues of these two proteins. While electrostatic interactions make major contributions towards stabilizing the transition state, we show two examples where steric hindrance exerts an effect on the transition state as well. To model the transition-state structures from the experimental DeltaDeltaG(double dagger)(int) values, we introduced a method for structure perturbation, searching for those inter-protein orientations that best support the experimental DeltaDeltaG(double dagger)(int) values. Two types of transition states were found, specific as observed for Barnase-Barstar and the electrostatically optimized TEM1-BLIP mutants, and diffusive as shown for wild-type TEM1-BLIP and IFNalpha2-IFNAR2. The specific transition states are characterized by defined inter-protein orientations, which cannot be modeled for the diffusive transition states. Mutations introduced through rational design can change the transition state from diffusive to specific. Together, these data provide a structural view of the mechanism allowing rates of association to differ by five orders of magnitude between different protein complexes.     

DeltaG-based prediction and experimental confirmation of SYCRP1-binding sites on the Synechocystis genome

DNA-binding sites for SYCRP1, which is a regulatory protein of the cyanobacterium Synechocystissp. PCC6803, were predicted for the whole genome sequence by estimating changes in the binding free energy () for SYCRP1 for those sites. The values were calculated by summing DeltaDeltaG values derived from systematic single base-pair substitution experiments (symmetrical and cooperative binding model). Of the calculated binding sites, 23 sites with a value <3.9kcal.mol(-1) located upstream or between the ORFs were selected as putative binding sites for SYCRP1. In order to confirm whether SYCRP1 actually binds to these binding sites or not, 11 sites with the lowest values were tested experimentally, and we confirmed that SYCRP1 binds to ten of the 11 sites with a DeltaDeltaG(total) value <3.9kcal.mol(-1). The best correlation coefficient between and the observed DeltaDeltaG(total) for binding of SYCRP1 to those sites was 0.78. These results suggest that the DeltaDeltaG values derived from systematic single base-pair experiments may be used to screen for potential binding sites of a regulatory protein in the genome sequence.     

Tissue transglutaminase modulates alpha-synuclein oligomerization

We have studied the interaction of the enzyme tissue transglutaminase (tTG), catalyzing cross-link formation between protein-bound glutamine residues and primary amines, with Parkinson's disease-associated α-synuclein protein variants at physiologically relevant concentrations. We have, for the first time, determined binding affinities of tTG for wild-type and mutant α-synucleins using surface plasmon resonance approaches, revealing high-affinity nanomolar equilibrium dissociation constants. Nanomolar tTG concentrations were sufficient for complete inhibition of fibrillization by effective α-synuclein cross-linking, resulting predominantly in intramolecularly cross-linked monomers accompanied by an oligomeric fraction. Since oligomeric species have a pathophysiological relevance we further investigated the properties of the tTG/α-synuclein oligomers. Atomic force microscopy revealed morphologically similar structures for oligomers from all α-synuclein variants; the extent of oligomer formation was found to correlate with tTG concentration. Unlike normal α-synuclein oligomers the resultant structures were extremely stable and resistant to GdnHCl and SDS. In contrast to normal β-sheet-containing oligomers, the tTG/α-synuclein oligomers appear to be unstructured and are unable to disrupt phospholipid vesicles. These data suggest that tTG binds equally effective to wild-type and disease mutant α-synuclein variants. We propose that tTG cross-linking imposes structural constraints on α-synuclein, preventing the assembly of structured oligomers required for disruption of membranes and for progression into fibrils. In general, cross-linking of amyloid forming proteins by tTG may prevent the progression into pathogenic species.     

How protein stability and new functions trade off

Numerous studies have noted that the evolution of new enzymatic specificities is accompanied by loss of the protein's thermodynamic stability (DeltaDeltaG), thus suggesting a tradeoff between the acquisition of new enzymatic functions and stability. However, since most mutations are destabilizing (DeltaDeltaG>0), one should ask how destabilizing mutations that confer new or altered enzymatic functions relative to all other mutations are. We applied DeltaDeltaG computations by FoldX to analyze the effects of 548 mutations that arose from the directed evolution of 22 different enzymes. The stability effects, location, and type of function-altering mutations were compared to DeltaDeltaG changes arising from all possible point mutations in the same enzymes. We found that mutations that modulate enzymatic functions are mostly destabilizing (average DeltaDeltaG = +0.9 kcal/mol), and are almost as destabilizing as the "average" mutation in these enzymes (+1.3 kcal/mol). Although their stability effects are not as dramatic as in key catalytic residues, mutations that modify the substrate binding pockets, and thus mediate new enzymatic specificities, place a larger stability burden than surface mutations that underline neutral, non-adaptive evolutionary changes. How are the destabilizing effects of functional mutations balanced to enable adaptation? Our analysis also indicated that many mutations that appear in directed evolution variants with no obvious role in the new function exert stabilizing effects that may compensate for the destabilizing effects of the crucial function-altering mutations. Thus, the evolution of new enzymatic activities, both in nature and in the laboratory, is dependent on the compensatory, stabilizing effect of apparently "silent" mutations in regions of the protein that are irrelevant to its function.     

Prediction of protein mutant stability using classification and regression tool

Prediction of protein stability upon amino acid substitutions is an important problem in molecular biology and the solving of which would help for designing stable mutants. In this work, we have analyzed the stability of protein mutants using two different datasets of 1396 and 2204 mutants obtained from ProTherm database, respectively for free energy change due to thermal (DeltaDeltaG) and denaturant denaturations (DeltaDeltaG(H(2)O)). We have used a set of 48 physical, chemical energetic and conformational properties of amino acid residues and computed the difference of amino acid properties for each mutant in both sets of data. These differences in amino acid properties have been related to protein stability (DeltaDeltaG and DeltaDeltaG(H(2)O)) and are used to train with classification and regression tool for predicting the stability of protein mutants. Further, we have tested the method with 4 fold, 5 fold and 10 fold cross validation procedures. We found that the physical properties, shape and flexibility are important determinants of protein stability. The classification of mutants based on secondary structure (helix, strand, turn and coil) and solvent accessibility (buried, partially buried, partially exposed and exposed) distinguished the stabilizing/destabilizing mutants at an average accuracy of 81% and 80%, respectively for DeltaDeltaG and DeltaDeltaG(H(2)O). The correlation between the experimental and predicted stability change is 0.61 for DeltaDeltaG and 0.44 for DeltaDeltaG(H(2)O). Further, the free energy change due to the replacement of amino acid residue has been predicted within an average error of 1.08 kcal/mol and 1.37 kcal/mol for thermal and chemical denaturation, respectively. The relative importance of secondary structure and solvent accessibility, and the influence of the dataset on prediction of protein mutant stability have been discussed.     

Fluorous effect in proteins: de novo design and characterization of a four-alpha-helix bundle protein containing hexafluoroleucine

Several studies have demonstrated that proteins incorporating fluorinated analogues of hydrophobic amino acids such as leucine and valine into their hydrophobic cores exhibit increased stability toward thermal denaturation and unfolding by guanidinium chloride. However, estimates for the increase in the thermodynamic stability of a protein (DeltaDeltaG(unfold)) afforded by the substitution of a hydrophobic amino acid with its fluorinated analogue vary quite significantly. To address this, we have designed a peptide that adopts an antiparallel four-helix bundle structure in which the hydrophobic core is packed with leucine, and investigated the effects of substituting the central two layers of the core with L-5,5,5,5',5',5'-hexafluoroleucine (hFLeu). We find that DeltaDeltaG(unfold) is increased by 0.3 kcal/mol per hFLeu residue. This is in good agreement with the predicted increase in DeltaDeltaG(unfold) of 0.4 kcal/mol per residue arising from the increased hydrophobicity of the hFLeu side chain, which we determined experimentally from partitioning measurements on hFLeu and leucine. The increased stability of this fluorinated protein may therefore be ascribed to simple hydrophobic effects, rather than specific "fluorous" interactions between the hFLeu residues.     

Fluorescence assay for the binding of ribonuclease A to the ribonuclease inhibitor protein

Ribonuclease A (RNase A) and the ribonuclease inhibitor protein (RI) form one of the tightest known protein-protein complexes. RNase A variants and homologues, such as G88R RNase A, that retain ribonucleolytic activity in the presence of RI are toxic to cancer cells. Herein, a new and facile assay is described for measuring the equilibrium dissociation constant (K(d)) and dissociation rate constant (k(d)) for complexes of RI and RNase A. This assay is based on the decrease in fluorescence intensity that occurs when a fluorescein-labeled RNase A binds to RI. To allow time for equilibration, the assay is most readily applied to those complexes with K(d) values in the nanomolar range or higher. Using this assay, the value of K(d) for the complex of RI with fluorescein-labeled G88R RNase A was determined to be 0.55 +/- 0.03 nM. In addition, the value of K(d) was determined for the complex of RI with unlabeled G88R RNase A to be 0.57 +/- 0.05 nM by using a competition assay with fluorescein-labeled G88R RNase A. Finally, the value of k(d) for the complex of RI with fluorescein-labeled G88R RNase A was determined to be (7.5 +/- 0.4) x 10(-3) s(-1) by monitoring the increase in fluorescence intensity upon dissociation. This assay can be used to characterize complexes of RI with a wide variety of RNase A variants and homologues, including those with cytotoxic activity.     

Evaluation of protein-protein association energies by free energy perturbation calculations

The association energy upon binding of different amino acids in the specificity pocket of trypsin was evaluated by free energy perturbation calculations on complexes between bovine trypsin (BT) and bovine pancreatic trypsin inhibitor (BPTI). Three simulations of mutations of the primary binding residue (P(1)) were performed (P(1)-Ala to Gly, P(1)-Met to Gly and P(1)-Met to Ala) and the resulting differences in association energy (DeltaDeltaG(a)) are 2. 28, 5.08 and 2.93 kcal/mol for P(1)-Ala to Gly, P(1)-Met to Gly and to Ala with experimental values of 1.71, 4.62 and 2.91 kcal/mol, respectively. The calculated binding free energy differences are hence in excellent agreement with the experimental binding free energies. The binding free energies, however, were shown to be highly dependent on water molecules at the protein-protein interface and could only be quantitatively estimated if the correct number of such water molecules was included. Furthermore, the cavities that were formed when a large amino acid side-chain is perturbed to a smaller one seem to create instabilities in the systems and had to be refilled with water molecules in order to obtain reliable results. In addition, if the protein atoms that were perturbed away were not replaced by water molecules, the simulations dramatically overestimated the initial state of the free energy perturbations.     

Structural characterisation and functional significance of transient protein-protein interactions

Protein-protein complexes that dissociate and associate readily, often depending on the physiological condition or environment, play an important role in many biological processes. In order to characterise these "transient" protein-protein interactions, two sets of complexes were collected and analysed. The first set consists of 16 experimentally validated "weak" transient homodimers, which are known to exist as monomers and dimers at physiological concentration, with dissociation constants in the micromolar range. A set of 23 functionally validated transient (i.e. intracellular signalling) heterodimers comprise the second set. This set includes complexes that are more stable, with nanomolar binding affinities, and require a molecular trigger to form and break the interaction. In comparison to more stable homodimeric complexes, the weak homodimers demonstrate smaller contact areas between protomers and the interfaces are more planar and polar on average. The physicochemical and geometrical properties of these weak homodimers more closely resemble those of non-obligate hetero-oligomeric complexes, whose components can exist either as monomers or as complexes in vivo. In contrast to the weak transient dimers, "strong" transient dimers often undergo large conformational changes upon association/dissociation and are characterised with larger, less planar and sometimes more hydrophobic interfaces. From sequence alignments we find that the interface residues of the weak transient homodimers are generally more conserved than surface residues, consistent with being constrained to maintain the protein-protein interaction during evolution. Protein families that include members with different oligomeric states or structures are identified, and found to exhibit a lower sequence conservation at the interface. The results are discussed in terms of the physiological function and evolution of protein-protein interactions.