JozefThesis.txt

under supervision of

 – 





 









: , ,  

[note:











]






 





For a research worker the unforgotten moments of his life are 
those rare ones,
which come after years of plodding work,
 when the veil over nature’s secret seems suddenly to lift,
and when what was dark and chaotic
appears in a clear and beautiful light and pattern


— Gerty Cori [103]





To my parents and teachers,
 who managed to sustain the curiosity in me.








Abstract

Almost every macromolecule in the cell or living organism has to 
interact transiently or permanently with other particles to 
fulfill their role. Around 30–40% of these macromolecules are 
envisioned to interact with metal ions, often the interaction 
being obligatory for the macromolecule to be biologically active. 
What is bizarre, given those numbers, is that to date, studies 
dedicated to the characterization of protein–protein networks do 
not overlap with the research devoted to metal–protein 
interactions, leaving the area of intersection a terra incognita. 
My interdisciplinary doctoral project had two phases. During the 
first phase, I attempted to explore the unknown area 
programmatically surveying the , for the —protein–protein 
interfaces involving the metal ions. The second phase was devoted 
to thorough characterization, by biophysical methods, one of the 
special cases of the —the zinc hook of , crucial for 
physiological dimerization of the  and, consequently, necessary 
for the proper functioning of the key player in  damage 
response—the .

The logic of the surveying was deployed using  programming 
language. One of the two generated in this process surveys 
present the physiological  (assessed manually). The second survey 
is constantly and regularly updated, as is it deployed as an 
online tool using the  framework—InterMetalDB.

The second phase of my doctoral project—the biophysical 
characterization of the zinc hook from the  had two parts. The 
first part was to describe the sequence–structure–stability 
relationship of the human . The study is the first study, 
exploring the stability of the eukaryotic , moreover, to my 
knowledge, the investigated complex forms the most stable 
Zn(II)-complex described in the human proteome, so far. The human 
 is phosphorylated in the near vicinity (structurally and 
sequentially) of the zinc-binding motif, at threonine 690. We 
have determined that, due to the extreme stability of the 
complex, the phosphorylation of the zinc hook cannot be a switch 
that controls the metallation state of the  under the 
physiological conditions.

Despite the extreme stability of the Zn()₂, the complex is not 
selective against Zn(II), which can be replaced by metal ions 
with higher affinities towards the zinc hook. The second part of 
the characterization of the zinc hook was performed using the 
model zinc hook domain from   and was devoted to the 
investigation of the structural, as well as stability properties 
of the mismetallated domain by heavy metal ions—Hg(II) and Ag(I). 
Those metals easily and readily displaced Zn(II) under the 
experimental condition, suggesting that the stabilities of the 
formed complexes were even higher those of the Zn(II)-complex. I 
was able to estimate these unimaginably high affinities using the 
least likely method—. 

The obtained results contribute to understanding the , including 
, as well as the relationship between the structure and the 
stability of the human , as well as the impact of the heavy metal 
ions on the stability and architecture of the  from 





 

Streszczenie

Prawie każda makrocząsteczka w komórce lub żywym organizmie musi 
przejściowo lub trwale oddziaływać z innymi cząsteczkami, aby 
spełnić swoją rolę. Przewiduje się, że około 30–40% tych 
makrocząsteczek oddziałuje z jonami metali, przy czym często 
oddziaływanie to jest niezbędne, aby makrocząsteczka była 
biologicznie aktywna. Biorąc pod uwagę te liczby, dziwi fakt, że 
jak dotąd badania poświęcone charakterystyce sieci oddziaływań 
białko–białko nie pokrywają się z badaniami poświęconymi 
interakcjom metal–białko, pozostawiając obszar przecięcia jako 
terra incognita. Mój interdyscyplinarny projekt doktorski miał 
dwie fazy. W pierwszej fazie podjęłam próbę eksploracji 
nieznanego obszaru programowo badając zasoby Research 
Collaboratory for Structural Bioinformatics Protein Data Bank, 
pod kątem interakcji białko–białko z udziałem metalu—a dokładniej 
interfejsów białko–białko z udziałem jonów metali. Drugi etap 
poświęcony był dokładnej charakterystyce, metodami biofizycznymi, 
jednego ze szczególnych przypadków interakcji białko–białko z 
udziałem jonu cynku—haczyka cynkowego białka naprawczego DNA 
Rad50, koronnego dla fizjologicznej dimeryzacji Rad50, a w 
konsekwencji niezbędnego do prawidłowego funkcjonowania 
kluczowego gracza w odpowiedzi na uszkodzenia kwasu 
deoksyrybonukleinowego—kompleksu Mre11–Rad50–Nbs1(Xrs2) (MRN(X)).

Logika badania została wdrożona przy użyciu języka programowania 
Python. Jeden z dwóch wygenerowanych w tym procesie przeglądów 
przedstawia fizjologicznych interakcji białko–białko z udziałem 
jonu cynku (ewaluowane ręcznie). Drugi przegląd jest stale i 
regularnie aktualizowany, ponieważ jest wdrożony jako narzędzie 
online z wykorzystaniem platformy programistycznej 
Django—InterMetalDB.

Drugi etap mojego projektu doktorskiego—biofizyczna 
charakterystyka haczyka cynkowego z białka Rad50 składał się z 
dwóch części. Pierwszą częścią było opisanie relacji 
sekwencja–struktura–stabilność ludzkiego białka Rad50. Powstała 
publikacja jest pierwszą publikacją, opisującą stabilność 
eukariotycznego Rad50, ponadto, według mojej wiedzy, badany 
kompleks tworzy najbardziej stabilny kompleks cynkowy opisany w 
ludzkim proteomie. Ludzkie białko Rad50 jest fosforylowane w 
bliskim sąsiedztwie (strukturalnie i sekwencyjnie) motywu 
wiążącego cynk, na treoninie 690. Pokazałem również, że ze 
względu na ekstremalną stabilność kompleksu fosforylacja haka 
cynkowego nie może być przełącznikiem kontrolującym stan 
dimeryzacji białka Rad50 w warunkach fizjologicznych. Pomimo 
ekstremalnej stabilności Zn(Rad50)₂, kompleks nie jest selektywny 
wobec Zn(II), który może być zastąpiony jonami metali o wyższym 
powinowactwie do haka cynkowego.

Druga część charakterystyki haczyka cynkowego została 
przeprowadzona z wykorzystaniem modelowej domeny haczyka 
cynkowego z Rad50 pochodzącego z z Pyrococcus furiosus (P. 
furiosus) i była poświęcona badaniu właściwości strukturalnych, a 
także stabilności domeny wysyconej przez jony metali 
ciężkich—Hg(II) i Ag(I). Metale te w warunkach eksperymentalnych 
łatwo i chętnie wypierały Zn(II), co sugeruje, że stabilność 
powstałych kompleksów była nawet wyższa niż kompleksu cynkowego. 
Te niewyobrażalnie wysokie powinowactwa udało mi się oszacować 
przy użyciu najmniej prawdopodobnej metody—izotermicznej 
kalorymetrii miareczkowej. Uzyskane wyniki przyczyniają się do 
zrozumienia interakcji białko–białko z udziałem metalu, w tym 
interakcji białko–białko z udziałem jonu cynku, jak również 
zależności pomiędzy strukturą i stabilnością ludzkiego białka 
Rad50, a także wpływu jonów metali ciężkich na stabilność i 
architekturę białka Rad50 z P. furiosus.












 

Publications

Some ideas and figures have appeared previously in the following 
publications:







Bibliography



 







To succeed, planning alone is insufficient. 
One must improvise as well. 


 — Isaac Asimov [12] 





 

Acknowledgments

Plato, and his student, Aristotle, thought that  is by nature a 
political animal, i.e., man is made to live in a polis, and be an 
active member of society. A solitary man is not 
self-sufficient—he has various needs that cannot be fulfilled 
alone. An individual cannot even meet his material needs, let to 
say intellectual needs, such as play, art, music, sports, 
friendship, and learning. The same is true for conducting 
research—this doctorate would not have been possible without the 
contribution (direct or indirect) of others. These 
acknowledgments are not sorted by importance, I do believe that 
would be impossible to do objectively, however, I do hope that 
everyone important during my doctoral project is mentioned and 
recognized.

I would like to thank to:

• my supervisor Artur Krężel for being my supervisor and guidance 
  for past the five years,

• Violetta Trzyna for being the invaluable help with tedious and 
  complicated paperwork,

• Michał Padjasek for sharing various hobbies, support, and 
  inspiring,

• my parents and family for support,

• Adam Pomorski for insightful chats and brilliant remarks,

• Jakub Sławski, Michał Tracz and Marek Łuczkowski for common 
  lunches and discussions,

• Mateusz Krzyścik for technical expertise and providing a wrench 
  when needed,

• Anna Kocyła, Olga Kerber, Marek Łuczkowski, Aleksandra Chorążew
  ska, Alicja Misiaszek and other labmates, for successful 
  collaboration,

• students that have I worked with: Aleksandra Gędaj and Ewa 
  Olczak for patience,

• Krystyna Grzesiak for understanding, patience, and support, so 
  much needed when conducting research and writing articles,

• Czarek Grzesiak for short and long walks,

• Gabor Markowski, Wojciech Graf ,and Krzysztof Pukało for 
  showing me that sport can be fun,

• Krzysztof Pukało, Wojciech Graf, Gabor Markowski, Maciej Majkow
  ski and Piotr Bachry for giving me the opportunity to stress 
  out by playing Dungeons & Dragons,

• Katsuaki Inoue, Nathan Cowieson, Diego Gianolio for being a 
  local contact at Diamond Light Source.

• Alexandra Elbakyan, for her contributions to science.








 









































Introduction

Proteins and cofactors <chap:Proteins-and-cofactors>

A significant portion of organic compounds synthesized by living 
organisms are biopolymers; proteins, lipids, nucleic acids, and 
other macromolecules (e.g., lignins, gums, melanins). Among these 
macromolecules, proteins fulfill key roles in almost every 
biological process (catalysis, transport, mechanical-structural 
functions, and others). This extraordinarily rich diversity of 
functions is achieved despite the fact that proteins are linear 
polymers built only of 20 types of canonical amino acids[footnote:
Without taking into account  and amino acid residues that are 
inserted during the cotranslation process.
], those canonical amino acids linked sequentially into a linear 
polymer are responsible for a staggering variety of folds, 
structures, and assemblies. This gorgeous diversity of protein 
structures allows proteins to fulfill an astonishing number of 
functions, however, not every chemical (or physical) process is 
possible to conduct with only 20 canonical amino acids. One way 
evolution  has solved the problem of limited chemical groups that 
can facilitate important processes is the use of  and . Numerous 
publications state that roughly one-third of proteins interact 
with [footnote:
My duty to the reader is to point out that the statement that 
one-third of proteins interact with  is not very credible. This 
type of statement often appears in various papers discussing the 
, yet one looks in vain for a reference to the literature or 
information on what organism these proteins are from, or how the 
analysis was done.
]Rosenzweig 2002; Bushmarina et al. 2006; Cao and Li 2011; ?? ??. 
The  queried for structures of enzymes return over 30% of enzymes 
that contain , which points out that  are essential for the 
functioning of a vast number of proteins Mukhopadhyay et al. 2019
.  

, contrary to the  are not bound covalently to proteins, but 
associate with proteins by means of other interactions, e.g., 
hydrogen bonds, Van der Waals forces, etc., thus to interact with 
 proteins had to evolve to form sites that will attract and 
associate with . The residues (not the sequence!) that adopt 
particular conformation and arrangement in space, being able to 
accommodate a particular  are called collectively a binding site.

Proteins are able to interact with organic, non-protein , called 
, the inorganic molecules constituting  are mostly metal ions. 
The residues that binding sites that interact with metal ions are 
called a “metal binding sitea”. Among various , the Zn(II) cation 
is one of the most commonly found in proteins. The  interacting 
with Zn(II) are among the most diverse and widespread proteins 
found in nature. Genome sequencing projects have provided a great 
deal of information about the structure of primary information 
about the primary structure of proteins. The bioinformatic study 
of this information analyzed in terms of metal interaction 
suggests that about 10% of the human proteome may bind Zn(II) 
ions Andreini et al. 2006; Andreini et al. 2006; Andreini, Bertini, and Rosato 2009
. What is worth noting is the fact, that those predictions (based 
solely on the protein sequence) do not take into account the fact 
that the  may be bound by two or more macromolecules. This 
minuscule comment spans a variety of other, challenging questions 
— “How often this intermolecular metal binding is occurring?”, “
What are the properties of such sites and the binding proteins?” 
and many others. 

This PhD project was dedicated to the analysis of proteins 
binding metal ions in . The part of the project realized in 
silico concerns the survey of all known proteins that bind metal 
ions in an intermolecular fashion (including zinc ion), while ex 
silico studies regard the interaction of metal ions with   (See: [ch:Rad50]
).

1.1 Metal ions

1.1.1 Alkali and alkaline earth metal ions

Approximately 1% of the human body weight are made up of alkali 
and alkaline earth metal ions Joseph J. Stephanos 2014. The role 
of alkali and alkaline earth metal ions in organisms is diverse, 
sodium and potassium ions occur in all known organisms, and 
generally, those ions function as electrolytes. Calcium ions are 
known to function as messengers—used by living organisms to 
communicate and orchestrate intracellular processes. Magnesium 
ions are the most abundant divalent cations in the cell, the 
majority of it bound to , which is essential to utilize  Schwartz et al. 2014
. 



In proteins both alkali and alkaline earth metal ions commonly 
play a structural role in the stabilization of the , however, 
those metals can play a catalytic role as well (e.g. Mg(II) in T4 
ligase Cherepanov and de Vries 2002). Alkali and alkaline-earth 
compensate charge of highly acidic regions in macromolecules, 
e.g., polyphosphate backbone in nucleic acids Owczarzy et al. 2004; Zheng et al. 2015; Varnai and Zakrzewska 2004
. Despite the fact that the biology of alkali and alkaline earth 
metals is quite well understood, it still hides some of the 
secrets, for instance, the biological effects (as well as some of 
the side effects) of lithium are known; lithium carbonate is 
widely used to treat some of the mood disorders, however, the 
biological targets of Li(I) ions in the organisms are not known 
yet.

The alkali metals have a single s electron, and the alkaline 
earth metal ions have a filled outer s-orbital (by two 
electrons).  Both groups of metals are highly electropositive and 
reluctantly polarizable Joseph J. Stephanos 2014. Due to the low 
polarizability alkali and alkaline earth metal ions are 
considered to be a hard Lewis acid. In accordance with Ralph 
Pearson's  concept, hard Lewis acids should react more willingly 
with hard Lewis bases Pearson 1963. This behavior is easily 
confirmed by observation of the  that complex those metals, i.e., 
coordination sphere of those metal ions is usually consisting of 
hard Lewis acids (in proteins, by oxygen-containing acidic amino 
acids). Usually alkali and alkaline-earth metal ions are 
coordinated by macromolecules with six-coordinated octahedral 
geometry Kuppuraj, Dudev, and Lim 2009, however more than six 
donor atoms are sometimes found in the alkali or alkali-earth 
metal binding sites – mostly due to the bidentate  Zheng et al. 2008
.

1.1.2 Transition metal ions <subsec:Transition-metal-ions>

Transition metals are defined as elements that have partially 
filled d orbital. Transition metals due to the fact that d 
orbitals are quite close in energy levels (degenerated), can form 
compounds with a wide range of oxidation states, some of those 
states are quite stable. The elements across the d block tend to 
be more polarizable. Those two properties make the use of 
transition metals useful in catalysis, thus there is no wonder 
why evolution has utilized so widely transition metals in 
enzymes. The first property is used when there is need for 
switching the oxidation state of the enzyme's substrate. In such 
enzyme, the metal ion can cycle between oxidation states and play 
a role as an acceptor or donor of electrons (depending on the 
catalyzed reaction). The second property allow metal ions to 
function in the catalytic sites without changing the oxidation 
state. Such metal cations play a role of a Lewis acid, accepting 
pairs of electrons, e.g., like Zn(II) in catalytic site of  Krishnamurthy et al. 2008
. 

In the presence of  the orbitals of the transition metal ions 
break of degeneracy, this causes a variety of physio-chemical 
effects (i.e., the property of having a color), but also exerts a 
particular geometry of the complex. Unsurprisingly proteins have 
evolved to accommodate a various geometries of the metal ions—the 
metal binding sites have to match the geometry of the particular 
metal ion to bind it.

Out of ten first-row transition metals, five of them are 
essential to human health—manganese, iron, cobalt, copper and 
zinc (not being a transition metal per se, discussed in [sec:properties_of_zinc]
). The essentiality of those elements is nonnegotiable, as those 
elements are  or build some of the  of biologically important 
proteins or enzymes. 

Manganese is present as a  in several enzymes, which maintain the 
metabolism, for examples in the human bran manganese-dependent 
glutamine synthetase is responsible for glutamine synthesis in 
astrocytes Takeda 2003. 

Iron accounts for 0.05‰ of human body weight. Hemoglobin is a 
flagship protein that realizes the necessity of iron in human 
diet. Iron in  is often bound as a —heme, however,  where iron 
ions act as a  exists as well, e.g., ferritin or rubredoxin.

Cobalt is an essential part of vitamin B[subscript:12], which is used as a coenzyme in  synthesis, amino acid, and 
fatty acids metabolism. Similarly to iron, cobalt does not have 
to be only bound to corrin derivatives or vitamin B[subscript:12], but can be utilized by some proteins as a  directly (e.g., 
methionine aminopeptidase 2).

Copper  play essential roles in electron transport and oxygen 
metabolism. Copper is essential for aerobic respiration in 
eukaryotes, as it is found as a  of cytochrome c oxidase, where 
it serves the function of electron transporter. Copper is also a 
coenzyme of copper-zinc , an enzyme that catalyzes the reaction 
of the disproportionation of superoxides.



[float Table:
[Tabela 1.1: 
Simplified version of the Periodic Table of the Elements 
indicating biologically relevant metals. With green color 
essential for life metals were marked. Yellow color marks the 
elements that are essential for some species. Chromium is marked 
in orange color as the essentiality for life of this element is 
questionable.
]






]

Additionally, to the five first-row essential d-block elements, 
three more first-row d-block elements—chromium, vanadium, and 
nickel—show some biological effects, however, the essentiality 
and need for supplementation of those elements is debatable. 

Cr(III) is found in foods and dietary supplements, many of which 
are advertised to have a beneficial effect on glucose metabolism 
and regulation. This rationale is based on observations of 
patients receiving Cr(III)-free intravenous drip who developed 
symptoms similar to insulin-resistant diabetes. Symptoms were 
relieved after intravenous drip supplementation with a small 
amount of Cr(III) Stehle, Stoffel-Wagner, and Kuhn 2016. The 
efficacy of the supplementation of Cr(III) is characterized by 
evidence of low strength thus the rationale to recommend their 
use for glycemic control in patients with type 2 diabetes 
mellitus is precarious Costello, Dwyer, and Bailey 2016. In the 
United States of America, Cr(III) ions are considered an 
essential nutrient in humans ?? 2022, the European Food Safety 
Authority does not share this opinion, arguing that the evidence 
supporting this assumption is insufficient Agostoni et al. 2014. 
While the role of Cr(III) on the human body is uncertain, the 
toxicity and carcinogenic properties of Cr(VI) are known for a 
long time Långard and Norseth 1975. 

Vanadium seems to be more important for life in marine 
environments—the number of marine life forms like algae utilizes 
vanadium in enzymes Butler and Carter-Franklin 2004. Tunicates 
draw attention due to specialized blood cells types—vanadocytes, 
accumulating vanadium through vanadium interacting protein called 
vanabins, however, the role of those proteins and the vanadium in 
tunicates is still shrouded in mystery Ueki et al. 2003. 
Vanadium-related issues for human health are similar to the case 
of chromium—vanadium complexes may exert some antidiabetic 
effects Sakurai et al. 2002, however a functional role for 
vanadium in mammals and humans is not defined yet, thus it is not 
recognized as an essential nutrient Institute of Medicine 2001. 

Nickel is essential to some bacteria, archaea, fungi, and plants, 
where usually nickel plays a role in enzymatic processes, e.g., 
urease, a nickel enzyme is considered a virulence factor in some 
pathogens. The essentiality of Ni(II) in mammals is still 
discussed, on the one hand, rats grown without nickel (or in low 
abundance) exhibit depressed growth, and several outside-the-norm 
biochemical and morphology results. Nickel seemed to be involved 
in the reproduction of nickel-deprived rats, as the lack of it 
resulted in an increased perinatal mortality rate Zambelli and Ciurli 2013
. Though no physiological nickel targets in mammals and humans 
have been found, the vast usage of Ni(II) by prokaryotes suggests 
that the element might be essential for the symbiotic gut 
microbiota Zambelli and Ciurli 2013.






Zinc proteome <chap:Zinc proteome>

2.1 Properties of zinc <sec:properties_of_zinc>

As have been mentioned in the first sentence of [subsec:Transition-metal-ions]
 transition metals have partially filled d orbital, following 
this definition one cannot classify zinc as a transition metal ([fig:Bohr-model-diagram-zinc]
). The other definitions of transition metals (i.e. as chemical 
elements in the d-block of the periodic table), may include zinc 
as a transition metal. The goal of this subsection is not to 
discuss whether zinc is a transition metal— “transition metals” 
is a human concept, people are in the habit of categorizing and 
dividing, however, it does not matter for the chemical and 
physical properties of zinc whether one classifies zinc as a 
transition metal or not. The unique properties of zinc and the 
reason why chemists discuss where to categorize zinc is due to 
zinc's electron configuration ([fig:Bohr-model-diagram-zinc]) – 
the fully filled d orbitals of zinc are energetically stable, and 
usually, zinc can lose only 4s electrons, leading to the 
formation of stable Zn(II), which is entirely different to the 
biologically active transition metals. This property of zinc of 
occurrence in only one oxidation state (2+) prevents zinc from 
partaking directly in redox reactions. The total filling of 
zinc's d orbitals makes zinc a 'spectroscopically quiet' metal – 
zinc complexes have no color. Contrary to metals that have 
partially filled d-shells, zinc has almost no spectroscopic 
signature Penner-Hahn 2005. This limitation of Zn(II) being 
redox-inert in biology might produce a false view of its use by 
evolution in proteins.

[float Figure:


[Rysunek 2.1: 
<fig:Bohr-model-diagram-zinc>Bohr model diagram for of zinc.
]
]

Zinc, after iron, is the second most abundant d-block element 
found in proteins, Andreini et al. 2006; Andreini et al. 2006; Andreini, Bertini, and Rosato 2009
 moreover it is the only d-block element that is found in all 
classes of enzymes, Vallee and Auld 1990 this is possible due to 
being a Lewis acid by Zn(II).

In solution Zn(II) ions exist in equilibrium as an octahedral 
hexaaquo complexes ([Zn(H₂O)₆][superscript:2+]) and as a tetrahedral [Zn(H₂O)₄][superscript:2+], with equilibrium shifted towards the first one. Krężel and Maret 2016
 Due to the total filling of the d orbitals, the transition 
between the octahedral to tetrahedral coordination (typical for 
complexation by proteins) in the case of zinc does not entail an 
energetic penalty, Lachenmann et al. 2004 additionally the 
release of solvent from the aqua-complex involves an increase in 
degrees of freedom in the system, which is energetically 
favorable due to the increase of . Krężel and Maret 2016  that 
utilize this unique , are unique as well, and bind this metal ion 
in an unusual fashion.

2.2 zinc-binding Sites<sec:Zinc-Binding-Sites>

Zinc(II) binding proteins are extremely diverse in structure and 
function. Zn(II) is an intermediate Lewis acid in terms of 
hardness according to the , theory formulated by Ralph Pearson. Pearson 1963
 Because of this, Zn(II) can be coordinated by zarówno the sulfur 
atom of the cysteine (a soft Lewis base), the nitrogen atom of 
the histidine (an intermediate Lewis base), as well as by 
carboxyl anions derived from the aspartates and glutamates (a 
hard Lewis base). The most common  found in the zinc-binding 
sites is cysteine, followed by histidine and acidic residues, 
i.e. aspartic acid and glutamic acid [footnote:
Those findings are based on the structural data from the . Just 
because there are structures of such  does not necessarily mean 
that this is true in all existing zinc-binding sites existing.
]. Laitaoja, Valjakka, and Jänis 2013 For zinc-binding sites,  
ranging from two to seven were found in the literature Sousa et al. 2009; Andreini, Bertini, and Cavallaro 2011
, however, di- and tricoordinate are zinc complexes are highly 
reactive, meaning that the stable zinc-binding sites have  at 
least equal four. Thorough manual analysis performed by Laitaoja, Valjakka, and Jänis 2013
 ruled out of physiological existence of zinc-binding sites of  
higher than seven. Laitaoja, Valjakka, and Jänis 2013 The most 
common zinc-binding sites found in the  have =4, which 
corresponds to the tetrahedral geometry. The low occurrence of 
zinc-binding sites with >4 is probably due to the small size of 
Zn(II) (~74 pm and ~88 pm for =4 and =6, respectively), which 
causes molecular repulsion between the  Miessler, Fischer, and Tarr 2013
. The  equal to two in zinc-binding sites is usually caused by 
the not-fully resolved structure of zinc-binding site having 
higher  (e.g. structure of   zinc hook domain,  [https://www.rcsb.org/structure/6ZFF||6ZFF]
 Soh, Basquin, and Gruber 2021), likewise =3 in most cases is 
unresolved zinc-binding site with =4, which is a very common 
situation in enzymatic zinc-binding sites, where the missing, 
unresolved  is often a water molecule. The different number and 
types of  often translate into differences in the properties of 
the coordination sphere and zinc-binding site formed.

The process of Zn(II) complexation by polypeptide often entails 
the formation of a stable conformation. The zinc-binding sites 
can exist as a pre-made arrangement of amino acids in the space, 
where zinc-binding does not cause huge structural changes in 
protein (usually in enzymes), however, the zinc-binding can 
promote protein folding and formation of separate protein domains 
(e.g., zinc fingers), this process is often in structural zinc 
sites. Kochańczyk, Drozd, and Krężel 2015

Zinc-binding sites can be divided by function but also by their 
architecture. Functionally, zinc sites can be divided into:

catalytic Zn(II) can be found in the active center of any of the 
six classes of enzymes distinguished by the International Union 
of Biochemistry and Molecular Biology. In almost every enzyme 
Zn(II) acts as a  and one of the zinc  is a water molecule. In 
other proteins with no catalytic function, Zn(II) is usually 
bound by the atoms of the side chains of the amino acid residues. 
The interaction of the water molecule with the zinc cation allows 
for the transient replacement of the water molecule by the 
coordination of the substrate molecule. However in the case of 
the most studied carbonic anhydrase, , the water molecule is not 
replaced displaced from the zinc-binding site, Zn(II) lowers the 
water p, facilitating dissociation to a hydroxide, that reacts 
with .

structural zinc-binding sites are usually characterized by a high 
affinity towards Zn(II). A classical example of domains 
containing structural zinc-binding sites are . An important 
structural characteristic that accompanies structural binding 
sites is the presence of a hydrophobic core that stabilizes the 
zinc-protein complex Padjasek et al. 2020. The function of the 
zinc-binding site in such domains is to stabilize the  – a good 
example of this may be a  peptide, an example of de novo protein 
design that bases on the   of zinc finger Dahiyat, Sarisky, and Mayo 1997
. The de novo  peptide maintains the same  as a fragment of , 
however, the fragment compared with  shares only four amino acids 
in the same positions Dahiyat, Sarisky, and Mayo 1997. What is 
most notable is the change of the  sequence with hydrophobic and 
aromatic residues. Which stabilizes energetically the  due to the 
hydrophobic effect. Maybe even the greatest importance of 
structural zinc-binding sites can be seen in the case of chimeric 
 , where zinc hook  from  can be efficiently replaced the  hinge Tatebe et al. 2020
 while taking much less volume.

regulatory zinc-binding sites usually have medium affinities 
towards Zn(II), this property allows them to associate with 
Zn(II) during the Zn(II) influx and dissociate during Zn(II) 
efflux. Zinc is an essential metal for life, however in case of 
Zn(II) excess, it exerts a toxic effect on the cell – thus the  
concentration in the cell needs to be regulated. So the cell 
regulates the  concentration by various mechanisms. The 
functional division of the zinc-binding sites proposed here 
distinguishes three classes that are involved in  regulation, 
which shows the importance of maintaining physiological  in the 
cell. An example of the regulatory zinc-binding site can be the  
from . Zinc-binding by  is characterized by remarkable lability 
observed both in vitro and in vivo Qiao et al. 2006.  This 
lability is essential for fulfilling its regulatory function.

transporting zinc-binding sites are responsible for cellular 
influx or efflux of Zn(II). Similarly to regulatory and buffering 
binding sites, the transporting zinc-binding sites are 
characterized by a medium affinity towards Zn(II). The examples 
of the transporting zinc-binding sites are  , , and , which are 
regulated by . The driving force for Zn(II) transport (which 
requires the conformational, and affinity towards Zn(II) change) 
may be a proton motive force, which is the case of human 
proton-coupled zinc antiporters (ZnT family), however, in the 
case of the bigger Zn(II) transporter family (ZIP), the driving 
force is unknown yet Coudray et al. 2013; Bafaro et al. 2017. 

buffering zinc-binding sites are found in proteins that bind 
Zn(II) in order to maintain physiological  in the cell. Those 
sites are characterized by a medium, but a broad range of 
affinities, which allows for maintaining adequate . A flagship 
example of buffering binding sites can be found in the 
metallothioneins, which may bind up to seven zinc, within a broad 
affinity range Krężel and Maret 2007.

The number of zinc-binding sites classes related to the 
regulation of the  in the cell together with the occurrence of 
zinc ions in cellular vesicles potentially could be explained by 
the fact that no Zn(II)-storage proteins (akin to the iron 
storage proteins – ferritin) was found. Ferritin for example may 
store several thousand iron ions trapped in its core Maret 2017. 

In terms of architecture, the zinc-binding sites can be divided 
into five classes:

intraprotein zinc-binding is the most common type of zinc 
coordination in proteins. The intraprotein binding is understood 
as a formation of a zinc-binding site only by singular 
polypeptide chains. 

clustered zinc-binding sites are multinuclear. The clustered 
zinc-binding sites have more than one Zn(II) per binding site, 
this is achieved by the presence of the bridging  (usually sulfur 
from cysteines). These sites are quite thermodynamically stable, 
while at the same time kinetically labile. No wonder why 
metallothioneins (mentioned in the buffering zinc-binding sites) 
utilize this type of binding. This type of binding is seen in 
most zinc enzymes, , etc. 

interprotein zinc-binding sites are not so common (or at least 
are not so commonly discovered and described) as intraprotein 
zinc-binding sites. Interprotein zinc-binding sites are formed by 
at least two polypeptide chains, so in this architecture, Zn(II) 
bound at the  of two proteins, takes a structural part in the 
formation of the quaternary protein structure. The formation of 
interprotein zinc-binding sites, referred to in the thesis as , 
may be obligatory to form a quaternary structure, or Zn(II) may 
bind to already formed protein-protein complex, to further 
stabilize it Kochańczyk, Drozd, and Krężel 2015. To date only a 
few  have been studied in terms of the complexes' stabilities, 
this includes - Kocyła and Krężel 2018and - Davis and Berg 2009 
complexes, and various  orthologs studied by me and my colleagues 
Padjasek et al. 2020; Tran, Padjasek, and Krężel 2022.

In terms of the time span of binding one can introduce two 
additional classifiers:

transient zinc-binding sites are characterized often by medium 
affinities toward Zn(II). The role of such transient zinc-binding 
sites is to bind Zn(II) temporally, which is the case in the 
buffering, transporting, and regulatory zinc-binding sites. Those 
sites need to bind Zn(II) in a kinetically labile way in order to 
fulfill its functions.

permanent zinc-binding sites are kinetically inert. This type of 
binding is characteristic of structural and enzymatic binding 
sites. The Zn(II) bound to those sites is not easily dissociated 
from the site. Of course, the reaction is in equilibrium so the 
subunit exchange still happens (see equations  and [eq:Ka_definition]
), however, the binding is characterized by a low value of .








Investigation of intermolecular metal-binding sites

Zn(II) ions were found to be essential for the growth of the 
toxic mold – Aspergillus niger in 1869, nearly one hundred years 
later, in 1961, it was postulated that zinc is essential also for 
humans.  Almost 20 years later the first structures of proteins 
containing Zn(II) started to appear. It will not be an 
exaggeration if I say that the number of scientific questions is 
proportional to the amount of new data—with the advent of the 
first Zn(II)-containing macromolecular structures, questions 
about the factors that govern the formation of the zinc-binding 
sites started to appear. With more Zn(II)-containing structures 
deposited in the  some tendencies became clear (see [sec:Zinc-Binding-Sites]
), however, the fact that the Zn(II) ions (and other metal ions) 
can be bound at the  of two or more macromolecules was rather 
ignored – one of the first reviews of structural knowledge of 
metal-binding sites in proteins appeared in 1992[a], Tainer, Roberts, and Getzoff 1992
 whereas the first review regarding the binding of metals at  
appeared in 2014[a]. Song et al. 2014 This Chapter will focus on 
the methodology of how one can investigate the metal-binding 
sites at .

3.1 In silico investigation of intermolecular metal-binding sites

3.1.1 The problems of in silico investigation of intermolecular 
  metal-binding sites

To date, there are no computational algorithms or other methods 
that are able to identify or classify intermolecular 
metal-binding based on protein sequence or structure. The problem 
of metal-binding sites prediction was undertaken in the past 
several times. Apart from the availability of such tools, which 
is often unacceptable Ye et al. 2022, the quality of the proposed 
classifiers is often poor. [margin:
Classifier is a tool that assigns an element to a certain class. 
An algorithm that tells whether a protein is a  based on the 
protein attributes is a classifier.
] Usually, the authors of such algorithms present the quality and 
efficiency of their classifier in a reliable way, e.g., by 
showing various kinds of statistics like receiver operating 
characteristic curve in the case of binary classifiers, the 
presented tools at first glance seem to work very well, however, 
to date, the use of such tools is limited. For sure one cannot 
accuse the authors of such articles of scientific misconduct or 
ill-faith – usually, the problem lies deeper. The are two main 
problems with the classifiers: the first one is the problem with 
the initial data set, which eventually is a source for creating 
the training and the test data sets, and the second problem is 
the choice of explanatory variables.

The availability of well-annotated (due to manual curation) 
protein sequence data is quite good. The primary source of such 
information is scientific articles. A publication that treats 
whether a protein binds metal or not has a number of methods that 
show and describe this binding. Such publication does not need to 
include information about the structure of the protein, this 
information is not necessary to prove the interaction of the 
metal with the protein. Typically, this type of data can be 
aggregated in , where metal-interacting residues are often 
assigned based on similarity and sequence. The problem with this 
type of data is the sparsity of annotations – not every known  is 
annotated in  to interact with metal. This leads to the attempts 
to create one's own data set, which in turn involves the use of 
the other type of input – the structural data deposited in the , 
however, the metal-binding sites in the  are not annotated to be 
a true physiological or adventitious metal-binding sites. This, 
in turn, leads to the use of simple algorithms that sort “true” 
from “false” metal-binding sites in the data set based on simple 
rules, i.e., the distance between the metal and , the number of , 
type of , etc., however, this approach is far from being 
accurate.

The problem with the selection of explanatory variables is most 
acute with classifiers that are based on the sequence. Whereas 
structural-based classifiers may involve the abundance of 
meaningful variables for the classifier (e.g., hydrophobicity, a 
charge of the amino acids, geometry, position in space, etc.), 
the effort of prediction of a metal-binding site based solely on 
the sequence might be similar to the prediction of the sales on 
stock exchange based using only stock shares name. Nevertheless, 
in some specific cases, where there is an existing pattern in the 
sequence it is possible to build sensitive and specific 
classifiers, e.g., the prediction of zinc fingers based on 
sequence is possible due to the existence of well-defined input 
data and the existence of patterns in the sequences Sathyaseelan, Patro, and Rathinavelan 2023
.

The above problems with the prediction of metal-binding sites are 
the same for intermolecular metal-binding sites, as well, 
moreover, the fact of two or more interacting macromolecules 
increases the complexity of the problem. So how to search for 
metals bound at  if there are no viable  in silico tools? The 
question is addressed below in [subsec:Obtaining-knowledge-about-InterMBS]
.

3.1.2 Obtaining knowledge about intermolecular metal-binding 
  sites<subsec:Obtaining-knowledge-about-InterMBS>

As in the case of information about the interaction of the metal 
with the protein, here also publications may also be the primary 
source of information about metal-binding on . However, the lack 
of recognition of inter-protein metal ion binding has partly 
contributed to the fact that even if a publication is accompanied 
by the deposition of a structure in the , the fact of 
inter-protein metal ion binding is not necessarily commented on 
in the publication in any way. An additional problem with 
extracting knowledge directly from publications is the fact that 
the search for this type of information can be time-consuming. 
There may be problems with the availability of the article (not 
every article is published as an open access article), and the 
fact that there is no tool that aggregates the articles based on 
the information on whether the described protein within the 
article has an inter-protein metal-binding site, makes this 
approach unreasonable. Nevertheless, it is not out of the 
question that this type of information aggregation will change 
(not necessarily regarding the ) with the increasing number of 
open access articles published each year and the development of 
technologies related to natural language processing.

To date, the best source of information on proteins that bind 
metal ions in an intermolecular manner are structural databases 
like . Since  contains all deposited structures (whether they 
contain metal or not) searching for information on metal-binding 
sites in  is not necessarily an easy task. For this reason, 
various secondary databases have been created that rely on the 
information contained in the .

MESPEUS was one of the first databases to include information on 
zinc sites, MESPEUS allowed filtering of data based on metal,  
type, and resolution, unfortunately, MESPEUS is now no longer 
available Hsin et al. 2008; Harding and Hsin 2014. 

ZifBASE is an even more specific database – it is dedicated only 
to zinc fingers, plus it contains far more information than just 
structural information – it contains sequence information, 
potential DNA interaction sequences, and more. However, as of 
today, despite the existence of the site, ZifBASE appears to be 
severely broken or non-functional at all, search and other 
functions do not work, additionally, it is not known how often 
the database is updated or whether longer updated Jayakanthan et al. 2009
. 

MetalPDB is the best-known derivative database based on  
regarding metal ions. MetalPDB has more advanced features 
compared to a simple database like MESPEUS, it has, for example, 
the ability to search for binding sites by geometric likeness. 
The big advantage of MetalPDB is the use of the “Minimal 
Functional Site” concept which allows the implementation of the 
above function. The disadvantage of MetalPDB is that MetalPDB is 
based on , not . This approach to building a database is 
incorrect because some of the binding sites (and especially the 
intermolecular binding) can only be seen after the symmetry 
operations are applied and the  is created. Like the rest of 
these tools, MetalPDB does not allow searching for inter-protein 
binding sites of metal ions. MetalPDB despite the fact that it is 
possible to automate this type of tool seems to be rarely 
updated, as of February 13, 2023, MetalPDB gives an update date 
as of April 5, 2022, which does not seem to be true. No 
metal-containing structures, but published close to the supposed 
MetalPDB update date are present in the database – for example, 
the structures [https://www.rcsb.org/structure/7V8G||7V8G] 
(published March 30, 2022) or [https://www.rcsb.org/structure/7TIH||7TIH]
 (published March 30, 2022) despite the fact that it is strewn 
with metals is also absent. The fact that derivative databases do 
not have all the records that are present in the primary database 
is understandable – for various reasons, over the years  files 
have all sorts of errors, inaccuracies, or format 
incompatibilities that cause problems in processing, 
nevertheless, the number of missing files in MetalPDB is puzzling 
Andreini et al. 2013; Putignano et al. 2018.

ZincBind is one of the newest databases, which includes 
information on “physiological” binding sites for Zn(II) ions, 
however,  it does not have information on intermolecular 
zinc-binding. Physiological sites in ZincBind are understood to 
be those with at least two binding residues and at least three 
binding atoms, additionally, ZincBind overcomes the biggest 
issues of MetalPDB: ZincBind is automatically updated, presented 
structures are already in the , and most of all the  of ZincBind 
is freely available under the MIT License Ireland and Martin 2019
.[margin:
The  availability makes the science more transparent.
] 

Using various types of databases, it is possible, after applying 
certain logic, to select  files that have metal ions at the 
interface of two macromolecules.[footnote:
This does not automatically mean that these sites will be 
physiologically relevant. 
] However, this type of manipulation may require some programming 
expertise to avoid the tedious manual review of hundreds or 
thousands of structures. To make this search possible for 
scientists without programming knowledge to find intermolecular 
metal-binding sites in proteins, I have decided to create a 
publicly available database – InterMetalDB[footnote:
https://intermetaldb.biotech.uni.wroc.pl/
]([chap:InterMetalDB]).

3.1.3 How one does build a database?

The first step in building a database is data acquisition. In the 
case of creating a primary database, data should be obtained 
directly from a suitable source. For scientific databases, the 
source of such data is often publications. For example, the 
NIST46[footnote:
https://www.nist.gov/srd/nist46
] database of affinity constants for chelator complexes with 
metals was built by searching and evaluating data from 
publications. In the case of , the situation is somewhat 
different nowadays – journals now often require the structure to 
be deposited prior to the article acceptation or release. This is 
a very smart approach, for several reasons – it reduces the 
workload on the  side, and guarantees that virtually every 
publication presenting a new structure will simultaneously 
deposit the structure in the . However, this does not mean that 
every structure described by humankind will be available in the . 
In 2023, private companies still do not publish macromolecular 
structures in the , while structures deposited in the  contribute 
to the development of new drugs by public entities Westbrook and Burley 2019
. 

In the case of the derivative databases, the first step o 
acquiring a data set is to query a primary database. If  is 
available the acquisition is usually trivial, as it requires only 
the application of the proper querying logic. If the database 
does not provide the  one still can obtain the information – 
manually or by constructing a web crawler [margin:
A web crawler is a program that scans the Internet. The main 
purpose of web crawlers is to collect data for various purposes, 
such as search engine indexing, data mining, and content 
monitoring.
] provided the database has an online version.

If the data was not filtered in the previous step the next step 
is the data selection, annotation, organization, and storage. The 
goal of this step is to ensure that the data is accurate and 
consistent. This step is particularly important in scientific 
research, where data is often complex, and diverse and requires 
long-term preservation. This can be done manually, by computer 
programs or scripts, or by a combination of these approaches. In 
case of the curation of the macromolecular data, it is important 
to take into account various information that accompanies the 
structure: the sequence of the macromolecules in the  files, the 
coordinates in the file, gene source organism, the expression 
mechanism, ways of structure acquisition, etc. An important issue 
with the data aggregated in the  is the fact of the redundancy[margin:
In the case of the  redundancy is understood as the existence of 
many similar macromolecular structures, i.e., two  files of the 
same protein can be almost identical, however, one file may have 
a better resolution.
] of structural files, which should be taken into account during 
the construction of the derivative database. This redundancy is 
dealt with usually by clustering based on the sequence.

During the data organization it is important to design a proper 
logic of the database – how the records of the databases are 
connected together within the database. For example 
macromolecular structure may contain multiple metal-binding 
sites, this simple and hardly revealing fact implies several 
questions, e.g.,: how the records of the metal-binding sites 
should be connected with the records of the structures?; how to 
choose a representative metal-binding site from a structure, or 
maybe are there multiple representative metal-binding sites?. If 
the answers to those and other questions are known it is possible 
to implement the logic behind the database.

The final step of the data curation is the data presentation and 
sharing, in the age of the Internet it is normal to present the 
curated data over the Internet as a hosted web application that 
allows for the querying, and viewing the aggregated results. This 
step largely relies on the previous step and the logic behind the 
database, if the database has poorly done bindings the querying 
of the database for the end user will be cumbersome. 

  Use of programming and markup languages in the creation and 
  presentation of databases

Programming languages play an important role in the creation of 
modern databases. There are several programming languages that 
are commonly used for creating the back-end[margin:
The back-end is understood as everything that is invisible to the 
end user. It contains all the application logic, bindings, etc.
] databases including , , Java, and others.

 is the most widely used language for managing and querying 
relational databases. It is used to create, modify, and delete 
database objects, such as tables, views, and indexes.  is also 
used to insert, update, and retrieve data from tables.

 is a popular general-purpose programming language that is widely 
used for data processing and analysis. A variety of libraries 
allow for easy object manipulation in , which allows for managing 
and building the logic in the database. For example,  is a 
popular web framework that is built on  and is often used for 
creating web applications that are operating on databases. The 
use of  allows working with databases using high-level 
abstractions instead of writing raw  queries. With , it is 
possible to define the logic and models of the database tables as 
regular  classes.  provides a convenient  for querying the 
database allowing it to retrieve and manipulate the data with  
objects and methods, which allows finally for a presentation as a 
front-end. [margin:
Front-end is everything that the end-user is able to see – user 
interface; buttons, images, etc. 
] 

The front-end of the web application is usually built with , the 
website is stylized using Cascading Style Sheets, and the 
interactivity is added using JavaScript (e.g., the ability to 
view the protein structure).

  Clustering

The part of data curation is the clustering of the primary data. 
One of the most common ways of dealing with redundancy in the 
protein structures is the use of a sequence clustering algorithm. 
This can be done by several programs, however,  in the 
InterMetalDB ([chap:InterMetalDB]) I have decided to utilize 
CD-HIT Li and Godzik 2006; Fu et al. 2012. Similar algorithms and 
programs may be used, e.g., BLAST Camacho et al. 2009 or mmseqs2 Mirdita, Steinegger, and Söding 2019; Mirdita et al. 2021
. To cluster  files one typically follows several steps:

1. Convert the  files into a FASTA format (a text file extending 
  ).

2. Use CD-HIT to cluster the protein sequences at a desired 
  sequence identity threshold.

3. Map the clustered sequences back to the original  structures.

4. Remove redundant structures in each cluster and select a 
  representative structure for each cluster. This step can be 
  performed by choosing as a representative a structure with the 
  best resolution or newest structure.

The resulting output is a set of non-redundant (or with 
diminished redundancy) representative structures for each 
cluster. Depending on the clustering threshold it may be required 
to cluster the structures based on another factor, e.g., based on 
the structure.



3.2 Ex silico investigation of intermolecular metal-binding sites

What if one already have some preliminary information (for 
example, obtained through database research, [subsec:Obtaining-knowledge-about-InterMBS]
) or hunch about the intermolecular metal-binding at ? With 
preliminary knowledge (or suspicion) of intermolecular metal 
binding, one can proceed with ex silico testing, where ex silico 
I will understand as all experiments performed without the use of 
a computer as the main working tool. In general, verification of 
intermolecular metal-binding is not significantly different from 
the study of metal–protein interaction, although there are 
existent some important differences that should be taken into 
account during the design of the experiment.

3.2.1 Proteomic tools for the detection of metal-involved 
  protein-protein interfaces

At the outset, it is worth noting that the issues addressed below 
concern both intermolecular metal ion binding and metal-induced  
formation. The verdict, which process is dealt of may require a 
more thorough investigation, for example, the involvement of 
structural and biophysical methods (see bellow). To my knowledge 
no , to detect complexes with an intermolecularly bound metal 
ion, have been performed so far, thus leaving the landscape of  
hidden in the fog —for this reason, considerations in this 
Subsection are purely theoretical.

State of art proteomic technologies already may give some 
information about , however, the detection of such interactions 
by those methods requires to take into experimental framework 
some important ideas regarding the complexes stabilities (see [subsec:Equilibria]
). One of the most important issues to take care for is the 
availability of the free metal (e.g., ) in the experiment. The 
care should be taken to supply used buffers with an appropriate 
metal, preferably in an controlled manner ([subsec:Equilibria]). 
Unfortunately the most of the modern proteomic technologies are 
not suitable for the investigation of transient (or weak) 
interactions, mostly due to the washing steps aimed at removal of 
nonspecific interactions. Furthermore, the subsequent detection 
of interaction partners with  techniques efficiently removes 
metals due to the low pH during sample preparation, thus 
additionally to the control experiments, experiments that 
stabilize the complexes (e.g, by crosslinking) should be 
prepared.

  Mass spectrometry

 is the gold standard method that is utilized in proteomics. 
Using  it is possible to tell what is the mass of the measured 
protein, or the fragments of the protein – peptides. Seemingly 
this information is of low value to the researcher, however, the 
the number of articles published each year proves different – 
with known protein mass it is possible to identify modifications 
of the protein. Tandem  and/or enzymatic fragmentation can give 
information about identity of the protein (provided the genome of 
the investigated organism is known).  combined with crosslinking 
and computational modeling can provide information about protein 
topology or even the structure Petrotchenko and Borchers 2022. So 
it is not without reason why  techniques are used so often in 
proteomics, including the investigation of . Unfortunately the 
investigation of  with  is not straightforward, as the sample 
prepared for the  usually have different pH than the 
physiological, and combined with protein fragmentation (enzymatic 
or in tandem ), and harsh ionization causes loosing of the bound 
metal ion. Combining regular  with native  can partly facilitate 
the detection of metal, however, native  comes with its own 
limitations.

  Affinity chromatography

The identification of novel  can be realized by means of affinity 
purification methods – based on the selective interaction between 
a  and a target proteins of peptides followed by identification 
using a coupled .[footnote:
Note that sample preparation and ionization in  usually leads to 
loss of the metal.
] From various affinity purification techniques two techniques 
stand out:  and . Both techniques rely on recombinant fusion 
proteins with affinity labels. The affinity labels used in those 
methods are used to pull-down the recombinant protein in complex 
with the interaction partner. The binding and washing steps 
determine the composition of eluted sample so during the 
experiment care should be taken to perform those steps in the 
buffers containing metal ions of interest, preferably in a 
controlled manner. The control experiments should be made that 
are conducted in metal of interest free environment. 

Additionally improvements, mostly due to the use of crosslinking 
(which can be made in vivo) in those techniques have been made to 
detect transient binding. The crosslinking allows for use of more 
rigorous washing steps, however, at cost of loosing bound metal 
ions.

Unfortunately this approach seems unsuitable for homomeric 
interactions.  

  Label transfer

In label transfer, the protein of interest is tagged with a 
reagent that after the crosslinking reaction is able to transfer 
the label to an interacting partner. The first step of the label 
transfer is the crosslinking of the complex composed of tagged 
protein of interest with an interaction partner, this is done 
usually using the photoreactive mechanism. The label transfer can 
be performed by another type of chemistry, e.g., by reducing the 
disulfide bond in the label. The transferred label can be used in 
affinity purification and detection. The obtained sample can be 
analyzed by downstream methods like Western Blot or  Liu et al. 2007
. Potentially this system could be used to detect  by performing 
crosslinking in the presence of metal of interest and comparison 
with a control crosslinking in metal-free buffers. The proteins 
that are found in the experiment with metal, but are missing from 
the control, the metal-free experiment could interact with the 
protein of interest due to the presence of particular metal.

  Crosslinking

Crosslinking can assist or be a part of the above-mentioned 
methods. The development of reactive groups that allow in vivo 
crosslinking certainly can facilitate the discovery of new , even 
those that are transient. The cell environment contains the 
physiological concentrations of free metals, thus providing the 
ideal environment for the metal-protein interaction. The 
crosslinking in vivo can be performed using cell-permeable 
chemicals or using genetically encoded noncanonical amino acids 
containing the reactive crosslinking groups. The crosslinked 
proteins can be detected using . 

3.2.2 Equilibria of metal-protein complexes<subsec:Equilibria>

Biophysical tools that are suitable for the investigation of  
generally answer the question “How strong is the interaction?”. 
The answer to the question can be expressed as a value, namely  ( 
or ). This thermodynamic property of a metal–protein system is a 
key element, that allows one to understand the interaction in a 
biological systems. So how to obtain those values and how are 
they defined? To simplify the discussion presented below I will 
focus only on the , which is sufficient to obtain the  value as 
it is defined as K_{\text{d}}=\frac{1}{K_{\text{a}}}. As the 
reader reads through the following somewhat lengthy introduction, 
it should become clear to the reader that, along with choosing 
the right experiment, the right models must be chosen to describe 
metal binding—each case may require a slightly different 
approach.

For reaction:



in which theoretical particles A and B form the complex  the  is 
defined as:

K_{\text{a}}=\frac{[A_{x}B_{y}]}{[A]^{x}\cdot[B]^{y}}=\frac{k_{\text{on}}}{k_{\text{off}}}
,where  is the  and  is the . All the equilibria reactions that 
concern the interaction of metal ions with proteins are happening 
in water, thus one should define a new constant—:

K_{\text{ex}}=\frac{[\text{MeL}]\cdot[\text{H}_{2}\text{O}]^{x}}{[\text{M\ensuremath{\text{e}_{\text{aq}}}}]\cdot[\text{L}]}=K_{\text{a}}[\text{H}_{2}\text{O}]^{x}
,where  is a metal ion and  is a . Note that in reaction:

the  does not change due to the fact that in most systems , the  
does not change making the  constant. Due to this fact in 
biochemistry,  is usually defined as in [eq:Ka_definition]. 
Following this logic,  for the processis defined as:

K_{\text{a}}=\frac{[\text{MeL}]}{[\text{Me}]\cdot[\text{L}]}So 
the  for the homodimeric  complex, which formation is defined by:
is defined as:

K_{\text{a}}=\frac{[\text{MeL}_{2}]}{[\text{Me}]\cdot[\text{L}]^{2}}
In case of , (¹ and ² are two different types of ) which forming is 
described by equation  the  assume the same unit as  defined in 
equation [eq:Ka_ML2], however, have different form (equation [eq:Ka_ML1L2]
).



K_{\text{a}}=\frac{[\text{Me}\text{L}^{1}\text{L}^{2}]}{[\text{Me}]\cdot[\text{L}^{1}]\cdot[\text{L}^{2}]}
Note that with more interacting partners the  definition, thus as 
well as the unit, will change. This should be taken into 
consideration when comparing complexes of various 
stoichiometries. Depending on the system the  can be determined 
with direct titration of the metal to the protein or competition 
experiments. 

  Direct titration experiments

In direct titrations, the only competing  for the metal is only 
water. Thus in the case of the  complex, equation [eq:KaML] 
defines the . In such conditions the total concentration of  is 
equal to the sum of protein-bound  and free , thus:

\frac{[\text{MeL}]}{[\text{Me\ensuremath{]_{\text{t}}}}}=\frac{1}{1+\frac{1}{K_{\text{a}}\cdot[\text{L}]}}
, where [\text{Me\ensuremath{]_{\text{t}}}} is total  
concentration. So in situations where \frac{1}{K_{\text{a}}}>>[L] 
the ratio of \frac{[\text{MeL}]}{[\text{Me\ensuremath{]_{\text{t}}}}}\rightarrow0
, which means that the affinity is too low to be observable (or 
nonexistent) during the titration, the similar situation is 
observed when \frac{1}{K_{\text{a}}}<<[L], in such case \frac{[\text{MeL}]}{[\text{Me\ensuremath{]_{\text{t}}}}}\rightarrow1
, which means that the affinity is to high to be observable. This 
behavior of the system can be modified by precise choice of the 
concentrations of the reagents. However the manipulation with 
reagent's concentrations is possible in some range: solubility of 
the reagents in case of weak-binding, and the detection limit in 
case of tight-binding.

  Competition titration experiments<subsec:Competition-titration-experiment>

In situations of tight binding (i.e.,\frac{1}{K_{\text{a}}}<<[L]\Rightarrow\frac{[\text{MeL}]}{[\text{Me\ensuremath{]_{\text{t}}}}}\rightarrow1
, according to equation [eq:MeL/Me_t ratio]) the metals in direct 
titration are bound to the protein in a manner that prevents the 
direct detection of free metal (in case of metals that can be 
observed directly, e.g. Fe(II) by  spectroscopy. To determine the 
 of such metal-protein complexes the  has to compete for the  
(equation ) with another compound () of which [superscript:MeC] is known.



With known [superscript:MeL] and [superscript:MeC] it is possible to define  (equation [eq:Kex_competition]).

K_{\text{ex}}=\frac{[\text{MeL}]\cdot[\text{C}]}{[\text{MeC}]\cdot[\text{L}]}=\frac{K_{\text{a}}^{\text{MeL}}}{K_{\text{a}}^{\text{MeC}}}

With defined  and known [superscript:MeC] it is possible to determine the [superscript:MeL] provided that at least one of the components can be measured at 
equilibrium or can give a measurable signal (e.g., heat change, 
fluorescence, etc.). The shortly described titration experiment 
can be designed to determine affinities of complexes with higher 
stoichiometries (i.e. ₂, ₃, etc.), as well as using the  with 
higher stoichiometries, so long as the affinities of the  are 
known, and in range of the [superscript:MeL]. Of course in practice it is better to use  with 1:1 
stoichiometry when possible. The use of competition titration 
sometimes is called a metal-ion buffer, due to the fact that a  
provides a controlled source of free metal ions analogously to 
the pH buffers.

3.2.3 Structural methods for the characterization of 
  metal-involved protein-protein interfaces

The detailed description of structural methods that are used for 
the characterization of the proteins, metal-proteins complexes, 
and  is far beyond the scope of this thesis – more comprehensive 
and elaborate resources are available. Structures of proteins and 
 can be acquired by X-ray crystallography, ,  spectroscopy, and 
other, less frequently used techniques. Due to the fact that 
during my PhD project, I have utilized some more `exotic` 
structural methods they are worth mentioning. At the outset, it 
is worth noting that the methods used by me during the PhD 
project do not give the precise coordinates of each atom in the 
protein, contrary to the above-mentioned structural methods. 

  Small-angle X-ray scattering

Similarly, like  spectroscopy or X-ray crystallography  results 
may be written into a  file. This technique provides information 
about the overall shape and size distribution of a protein.  is 
particularly useful for studying large proteins. The biggest 
advantage over the X-ray crystallography and  spectroscopy is the 
fact that  can be applied to flexible proteins. The combination 
of  with other biophysical techniques allows for the 
understanding of the shape and dynamics of the protein. Although  
itself is blind to metal ions it can be a helpful tool for the 
description of protein dynamics Padjasek 2022 (or its lack—[chap:hRad50]
, Tran, Padjasek, and Krężel 2022) induced by metal ion 
complexation.

  Extended X-ray absorption fine structure and X-ray absorption 
  near-edge structure

X-ray absorption spectroscopy is a technique used to study the 
electronic structure of elements and their local coordination 
environment in various materials, also in . X-ray absorption 
spectroscopy encompasses two techniques  and 

 involves the measurement of the intensity of X-rays scattered at 
large angles from a target atom. The scattered X-rays are 
collected and analyzed to determine the average distance and 
coordination number of the neighboring atoms surrounding the 
target atom.  provides information about the local coordination 
environment of the target atom and can be used to study the 
bonding environment of target atoms in .

, on the other hand, involves the measurement of the absorption 
of X-rays near the absorption edge of the target atom.  provides 
information about the oxidation state and coordination geometry 
of the target atom by analyzing the fine structure in the X-ray 
absorption spectra near the absorption edge.  can be used to 
determine the oxidation state of metal ions in , which can 
provide valuable information about their function and reactivity.

Both  and  have been widely used in the study of  and have 
provided important insights into the mechanisms of  function. By 
combining these techniques with other experimental and 
computational methods, scientists have been able to gain a deeper 
understanding of the roles of metal ions in enzyme catalysis, 
electron transfer reactions, and oxidative stress.

3.2.4 Biophysical tools for the characterization of 
  metal-involved protein-protein interfaces<subsec:Biophysical-tools-for>

[subsec:Equilibria] is a short introduction to the 
equilibria-related calculations, that are necessary to obtain 
accurate values, that can be later compared to other known 
affinities, which should give those values a biological context. 
However, to be able to calculate anything one requires data, this 
Subsection is devoted to the description of some methods that can 
produce the data for those calculations.

  Size Exclusion Chromatography

 is one of the first liquid chromatography techniques developed 
to investigate protein–protein interactions Boone and Adamec 2016
.  separates the molecules based on their size (thus on the 
molecular weight). With a properly calibrated  column, it is 
possible to approximate the weight of the investigated protein or 
protein complex. Provided the change in the complexes' molecular 
weight is considerable it is possible to measure the binding. In 
the case of homodimerization, the molecular weight of the complex 
is twice as big as the monomeric protein thus making  an ideal 
solution for the detection. However, in a theoretical case of 
heterodimerization of a macromolecule with a small peptide, 
assuming that the binding does not induce a huge conformational 
change in the macromolecule  seems inapplicable. 

It is possible to run directly titrated with metal samples on , 
however, in case of weak binding the lack of appropriate metal 
ions in the running buffers may cause the dissociation of the 
metal-protein complexes. To circumvent this inconvenience, the 
separation on  can be run in metal-ion buffers ([subsec:Competition-titration-experiment]
) that control the concentration of available metal. The  from 
the competition experiment performed with  can be calculated 
using the area under the peaks corresponding to the various 
protein fractions Kocyła and Krężel 2018, however, it should be 
noted that the observed peaks are not literally in an 
equilibrium.

The detection of separated protein fractions can be performed 
with almost any flow-method, e.g.  spectrophotometry, , .

  MALS

Usually  is performed in conjugation with  (-). This is due to 
the fact that, not every macromolecule is perfectly globular, and 
 separates particles based on shape and hydrodynamic radius – the 
latter is often correlated with the molecular weight of the 
particle, however, the shape is not necessarily correlated. Thus 
the application of  in-line with  allows for determination of 
mass and size in solution, overcoming many inconveniences with  
column calibration Minton 2016.

  Circular dichorism

 can be used in the determination of  in both, direct and 
competition titrations. With , it is possible to evaluate the 
secondary structure of the protein (or peptide) Greenfield 2007, 
folding and binding, provided that a change in the secondary 
structure of the protein is observed. The direct titration of 
protein with a metal ion provided tight affinities that can show 
the ratio of binding. For example,  interaction with Zn(II) can 
be monitored using  ([chap:Galvanization-of-Protein-Protein]), 
due to the tight binding it is possible to plot the ellipticity 
at the chosen wavelength in the Zn(II)/protein function, and 
observe the specific turning point at Zn(II)/protein=0.5, which 
together with the rest of the evidence suggests the formation of 
Zn()₂complex. 

  Ultraviolet-Visible spectrophotometry

 spectrophotometry is a useful tool that can be utilized as a 
detection for  separation.  spectrophotometry is an essential 
method in thhe determination of a metal–protein stabilities.  
seems an ideal method for 'spectroscopically loud' metals.[footnote:
Spectroscopically quiet metals have completely filled or empty 
d-shell. So colloquially metal ions with partially filled d-shell 
should be called `loud` Penner-Hahn 2005.
]  observation of some 'spectroscopically quiet' metal complexes 
is possible due to the  phenomenon.  is a type of electron 
transition in which the electron from the donor is transitioned 
to the acceptor's orbital. In the case of thiolates and Zn(II) 
the  transition can be observed, which shows that even in the 
case of `spectroscopically quiet` metals  can be applicable[footnote:
Sometimes it is not possible to observe  from the native metal 
coordinated by the protein of interest due to overlapping of the  
bands with the absorption bands of the protein backbone. In such 
cases one can use isostructural probes showing absorption bands 
in a different region, i.e., in case of Zn(II), Co(II) Kluska, Adamczyk, and Krężel 2018
 and Cd(II) Freisinger and Vašák 2013 can be used as 
isostructural spectroscopic probes.
] Vasak, Kaegi, and Hill 1981. 

Similarly to  in case of tight-affinities, it is possible to 
determine the binding stoichiometry from  titration.  can be used 
as well in competition titration studies, however, the ease of 
use of this technique will be dependent on the availability of 
the proper chelating probe Kocyła, Pomorski, and Krężel 2015. 

  Isothermal titration calorimetry

Virtually every chemical reaction involves heat change. This 
allows the reaction to be monitored with accurate titration 
calorimeters. Also, metal-binding can be monitored using, . As 
with the aforementioned methods, competitive titration can be 
used, provided that the competitor used is well characterized 
thermodynamically. At the same time, the fact that virtually 
every reaction involves a change in heat analysis of  data can be 
somewhat problematic, that is: proton dissociation from the 
protein during binding, protonation of the buffer, association of 
the metal ion with coordinating residues, and so on. All these 
factors should be taken into account during data analysis. It is 
more than important to convert the obtained [superscript:ITC] constant into the association constants for the corresponding 
complex. 

Recent data analysis advances in the  allow to determine the 
kinetic information from the experiments – this can be done by 
the analysis of each injection peak, which adds an additional 
cross-validation layer to the experiment, due to the fact that 
the kinetic constants are bound with the  with the equations  and 
[eq:Ka_definition] Burnouf et al. 2012. 

  Potentiometry

The use of the glass electrode in potentiometry allows for the 
quantitative determination of equation . Of course, the general 
equation  will take another form in case of different complexes 
with different stoichiometry, i.e., ₂, etc. 

Potentiometry still is the gold standard for the determination of 
complexes stabilities, however, it should be noted that all 
proton-related equilibria should be taken into account while 
calculating the  – this is often impossible for proteins due to 
large number of sidechains that can exchange protons. Moreover, 
proteins are sensitive to pH, in extremely acidic or basic 
conditions proteins may change structure, unfold and precipitate, 
thus making the potentiometry only suitable for shorter peptide 
models.

  Fluorescence-based techniques

Fluorescence-based techniques for determination of metal-protein 
complexes affinities can be divided into two categories:

label-free techniques may rely on measurement of intrinsic 
protein's fluorescence, in which case the direct titration with 
metal may be applied. However, if the intrinsic fluorescence is 
lacking the label-based techniques may be applied, or a 
competition titration with a appropriate fluorescent probe (e.g, 
FluoZin-3 for Zn(II) , mag-fura-2 for Mg(II), etc.) Johnson 2010, 

label-based techniques may use recombinant proteins of interest 
with fluorescent proteins to measure properties like 
fluorescence, , fluorescence anisotropy and fluorescence 
anisotropy decay. Except from the heterologous expression of 
recombinated proteins, proteins of interests may be labeled by 
chemical means Qiao et al. 2006.

The biggest advantage of the use of fluorescence based techniques 
is the low protein concentration required, this if often a huge 
advantage for reasons of economy, time and workload, though care 
should be when envisioning the stabilities constants of the 
unlabeled proteins from the labeled counterparts – the 
introduction of label may change the protein's conformation with 
in turn may change the affinity towards the metal.

  Bio-layer interferometry and surface plasmon resonance

Both  and  measurement techniques provide label-free detection, 
due to the real-time monitoring and the fact that the measurement 
consists of both association and dissociation step it is possible 
to measure the kinetics of the interaction. The measurement of  
with the use of  and  implies the use of the metal ion buffers ([subsec:Competition-titration-experiment]
) and the loading of the protein on the surface. While this 
method may be useful for investigation of heteromeric complexes 
(i.e.,  ), it seems to be non-ideal in the case of the homomeric 
complexes (i.e. ), or in the case where individual components 
from the heterocomplex dimerize as a result of metal-binding 
(e.g. zinc mediated  and ) Kocyła and Krężel 2018. A certain 
limitation of  and  is also the fact, the lack of advanced models 
that can take into account more advanced binding stoichiometries. 

The determination of  in  and  is possible by the determination 
of  and , which later can be used to calculate  (see equations  
and [eq:Ka_definition]).




DNA repair protein Rad50

4.1 The Mre11-Rad50-Nbs1(Xrs2) complex—architecture and function

<ch:Rad50>

The genetic material of every cell is constantly under the stress 
– both endogenous and exogenous  damaging agents, which may 
introduce  to the chromosome. In addition, the  formation 
sometimes is even a `physiological` phenomenon—the formation of  
is an essential step in immunoglobulin gene rearrangements and 
meiotic chromosome recombination Johnson et al. 2021; Zhang et al. 2022
. The  arise from exposure to ionizing radiation, and genotoxic 
agents, and may arise spontaneously while  is replicated Tubbs and Nussenzweig 2017
. Because even the single  event may be a cause of chromosomal 
aberration, aneuploidy, tumorigenesis, and cell death, cells have 
evolved to sense, signal and repair . The  complex is crucial for 
 repair as the complex is involved in every step of repair: 
sensing,  end binding, signaling and finally repairing. Those 
processes follow two general routes:  and . In the Eukarya the 
use of one of two pathways is dependent on the cell cycle Scully et al. 2019
. 

The , also known as the  complex is a complex composed of three 
types of proteins: , , and  or . The  is essential for the 
viability of vertebrate cells (the yeast cells are viable without 
the ). The  is engaged in  metabolism, from which the most 
important is the  repair. The  orthologs have been found in every 
domain of life – which underlines the importance of the  and  Pal 2020
. The  (and its homologs) seems to be unique to Eukarya, as no 
homologs of the protein were found in the other domains. The X in 
the  stands for the  protein characteristic of , sharing a 
similar function to  despite the weak homology.

Recent five years of  research showed mechanical and structural 
insight into the modes of action of . The  complex binds to the 
homoduplex  and scans past the nucleosomes to locate  ends. The 
recognition of free  ends is inherent to the , however,  the  
sliding and nucleosome bypassing is inherent to the . This 
function separation may explain why two of the subunits are 
conserved in all domains of life as the  core complex. The  
performs the  end recognition via facilitated one-dimensional 
diffusion. The sliding (not alternate binding and dissociation) 
is facilitated by the  subunit. Though  binds and hydrolyzes , 
the diffusion coefficients are  independent, which means that  
hydrolysis does not regulate facilitated diffusion. The region 
responsible for the one-dimensional facilitated diffusion is 
located along the globular domains of the ,  is not responsible 
for the - interaction Myler et al. 2017. Due to the unique  
properties, the whole  complex is able to diffuse on euchromatin 
and nucleosome-free regions, however,  the diffusion on 
heterochromatin may be hampered, which may be the reason why the  
is limited or the repair of  at such regions is slow.

Often the free  end created by a  event will be blocked by a 
protein, e.g. . In yeast the  deficient cells are highly 
sensitive to , this phenotype is attenuated by deletion of the  
gene — the  has to remove –blocked  to enable the  resistance. 
However,  it should be noted that the  is not abundant in lower 
eukaryotes as it is in higher eukaryotes, and in most cases in  
the  end processing is not required when the  are not obstructed Llorente and Symington 2004; Mimitou and Symington 2008; Zhu et al. 2008
. The [margin:
In prokaryotes the  is homomeric.
] heterodimer is an abundant protein (especially in human cells) Beck et al. 2011
, combined with the high affinity towards the  and the 
recruitment within seconds in vivo Mari et al. 2006, the  (and 
some other –interacting factors) is essential in  processing in 
high eukaryotes Sartori et al. 2007. The process is most likely 
facilitated by the  resection and subsequent  repair, this is 
often done by the  end resection.

Because the  can form hairpins or be modified by chemicals, 
associated proteins (like ), or covalently bound proteins Morimoto et al. 2019; Johnson et al. 2021
, the  repair requires the cut. The enzymatic cleavage is done by 
, an Mn(II) dependent metalloenzyme Paull and Gellert 1998; Trujillo et al. 1998; Usui et al. 1998
, but how it is done if the recombinant  in vitro is a 3'\rightarrow
5' exonuclease? The strict polarity of the  nuclease is 
counterintuitive, as for the  the 5'- strand is required Ranjha, Howard, and Cejka 2018
. The paradox was fueled by the genetic experiments performed on  
showing that the  is responsible for the 5'\rightarrow3' 
resection in vivo. How it is possible that the  contributes to 
the 5' strand degradation, and how the 3'\rightarrow5' 
exonucleolytic activity is inhibited? The second question is 
answered by the formation of the  core complex. In the cell, the 
-loaded  inhibits sterically the 3'\rightarrow5' exonuclease 
activity. Though the ase activity of  is low, in  the core  
complex endonucleolytic–exonucleolytic balance is regulated by  
hydrolysis.  promotes endonucleolytic activity, while  hydrolysis 
switches those activities Majka et al. 2012. 

In Eukarya (and especially in mammals or humans) the  (or  in ) 
is much more important in  in the core  complex functioning. 
First of all the () is required for nuclear transport, however,   
the nuclease regulation differs between  and . In ,  is not 
necessary for the endonucleolytic activity of , however,   it 
exerts some positive stimulation on the process, thus in the 
cells with deletion of , and nuclear localization signal placed 
on  the  end resection was only minimally impaired. In  the  is 
crucial for endonucleolytic activity, which is stimulated by 
additional factors, especially the .  (and in , the ortholog of —
) stimulates the endonucleolytic nick of 5' , facilitating the 
formation of the 3' overhanging .

The  performs this by the cut at the 5' strand in the near 
vicinity of the  end, followed by exonucleolytic resection from 
3'\rightarrow5' direction Garcia et al. 2011; Deshpande et al. 2016
.[margin:
to jest znane jako jakiś model czy coś  może jakiś obrazek?
] The result of this nucleolytic cleavage is , an intermediate to 
be utilized in strand annealing in  or to be further elongated by 
a 5'\rightarrow3' degradation in the . Both  and bacterial  can 
enzymatically cleave off the second strand, which results in the 
release of the blocked  termini Deshpande et al. 2016; Connelly, de Leau, and Leach 2003
. 

The  polarity paradox is thus explained by the:

proteins that bind free  ends, block sterically the 3'\rightarrow
5' exonucleolytic activity of ,

tight interaction with , inhibiting the 3'\rightarrow5' 
exonucleolytic activity while –loaded,

, () and () stimulation of the 5'-endonucleolytic activity of .

4.2 Globular domain of MRN(X)—the catalytic core

The globular domain of the  complex harbors the active centers of 
 and . The globular domain allows association and sliding on the  
in search of the  Myler et al. 2017. The globular domain of  
contains the Walker A and B motifs, which do bind . The mutation 
that prevents the  binding in  is equivalent to [margin:
 before the gene's name indicates deletion. 
]rad50, however,   the   complex with  that does not bind  the 
complex was only not functional in the cutting of the 3' 
overhanging  Moncalian et al. 2004. The biochemical studies of  
show that  hydrolysis by  simulates the nucleolytic activity of 
the  on  Connelly et al. 1997; Deshpande, Lee, and Paull 2017; Herdendorf et al. 2011; Hopfner et al. 2000; Paull and Gellert 1999; Trujillo and Sung 2001
, while the  core complex, is active on  even in the absence of  Connelly and Leach 1996; Herdendorf et al. 2011; Hopfner et al. 2000
. The exact step-by-step catalysis (and associated conformational 
changes) of the nucleolytic activity of the  or  is unknown yet. 
Many structures of the catalytic cores of / were published to 
date Hopfner et al. 2000; Hopfner et al. 2001; Lammens et al. 2011; Lim et al. 2011; Liu et al. 2016; Möckel et al. 2012; Park et al. 2011; Seifert et al. 2016; Williams et al. 2008; Williams et al. 2011…
, however,  no structural evidence of free  ends is shown, in 
which  is bound after  hydrolysis. Moreover, the structural 
differences between the  globular domain from  and bacterial and 
archeal  complexes suggest differences in the mechanism of action 
Park et al. 2011.

Till this year (2023) the structural detail of the interaction of 
human  with the catalytic core of  was unknown. Recent advances 
in  allowed for the visualization of the interaction of the human 
 with the  core complex. Rotheneder et al. 2023 Contrary to 
previous assumptions (and structures), single, not two  particles 
wrap around the dimeric . The binding is centered, and  wraps the 
 in a pseudo-symmetric fashion, the feature seems to be 
evolutionarily conserved across various eukaryotes. Rotheneder et al. 2023

The dynamicity of the complex differs between the domains, and 
the  complex behavior from [margin:
The  complex in  is called SbcC–SbcD complex.
] differs from the eukaryotic  from  or —quite unexpected 
observation, as the   complex adopts a closed (rod-shape) in a 
resting state Rotheneder et al. 2023, rather than the open 
conformation that is observed for the   complex Käshammer et al. 2019
. The open conformation of the complex was believed to be the 
conformation that is loaded onto , however,  at least in the case 
of the eukaryotic  the `rod` conformation is proposed as a 
loading conformation. Rotheneder et al. 2023 propose that the 
C-terminal fragment of eukaryotic  is responsible for the 
–independent  binding. The  C terminus could bind  directly, or 
orchestrate the –binding of the  catalytic core Williams et al. 2008
. However, the presented data show that the  has to be passed 
through the coiled coils of  to the catalytic core of . This 
suggests that at least two different –binding models must 
exist—an observation that fits well with the single-molecule  
curtain experiments—  binds  ends in an -dependent fashion, 
cleavage of –blocked end (and  from  itself) is an -dependent 
process as well, but the facilitated one-dimensional diffusion 
over the  is -independent Myler et al. 2017. Moreover, the C 
terminus of human  was found to be cleaved by the DNA-dependent 
metalloprotease SPRTN in cancer cells— lacking its C terminus is 
suggested to be a reason for impaired  binding and increased 
radiosensitivity Na et al. 2021. 

4.3 Coiled coil of the Rad50—flexible linker between the zinc 
  hook and catalytic core

Another unique part of the  complex is the extremely long coiled 
coil region of the  protein. The long internal coiled coil that 
constitutes the middle part of the  is formed by folding upon 
itself in the middle of the protein (the zinc hook), so the 
antiparallel coiled coil is formed in an intermolecular fashion. 
Atomic force microscopy showed that the coiled coil domains of   
span around 400 , while the coiled coil domain from human  is 
around 100  longer and reaches up to 500 . The length and the 
sequence of coiled coils are quite well conserved, moreover, the 
truncation of the coiled coils leads to severe biological 
consequences—shortening of the coiled coil of  abrogates the 
physiological meiotic  formation, telomere maintenance, and 
impairs the  repair by , interestingly  abrogation appeared after 
the deletion of the longer fragment than the deleted fragment 
that caused the problems with  Hohl et al. 2010.

The atomic force microscopy and the recent  images present the  
and  complexes in various conformations. Those conformations are 
available to form due to the flexibility of the coiled coil, this 
is probably achieved by breaks in heptade repeats of the coiled 
coil. The coiled coil conformation is dependent on the binding of 
 and  in the globular domain of  or . Käshammer et al. have 
observed that the   in the absence of  the nuclease domains were 
separated, and the stable globular domain formation was obtained 
after introduction of  to the solution. The -bound complex is 
proposed as a major autoinhibited state of the  complex ( active 
sites are blocked), due to the abundance of  in the cell and the 
low efficiency of the  in -hydrolysis. The -bound  presents a 
more compact structure, however,  the coiled coils of each  
protomer are separated, and form the flexible (thus so far 
impossible to use as input in the model reconstruction) `open` 
conformation. Further complex stiffening was observed in the 
vitrified sample containing both  and — the  complex was observed 
in the rod-shaped conformation, which allowed to reconstruct 
around one-third (from globular domains to the zinc hook) of the  
structure.



However,  as has been mentioned previously, the eukaryotic  
behaves in a different manner than the prokaryotic , the resting, 
-bound state of the  assumes mostly a rod-shape (however,  the 
open coiled coil subpopulation is observed as well). Rotheneder et al. 2023
 

4.4 The zinc hook—a tetramerization domain?<sec:The-zinc-hook>

The atomic force microscopy of the  showed that at the apex of 
the coiled coil domain, the  interacts with another  protomer, 
the follow-up X-ray crystallography studies (and more recent ) 
showed the detailed structure of the —the zinc hook of . The  
formed by two  protomers is a , where two of the protomers 
co-coordinate a single Zn(II) ion[footnote:
The first structure of the zinc hook presents zinc hook from   ([https://www.rcsb.org/structure/1L8D||1L8D]
), the structure contains Hg(II) ion, instead of Zn(II), this is 
discussed more in [chap:HgRad50].
], by the conserved CXXC motif, where C corresponds to a cysteine 
residue, and X corresponds to any of 20 canonical amino acids. 
The Zn(II) co-coordination of the cysteine's thiolates implies 
that Zn(II) directly takes part in the formation of the tertiary 
structure of the , what is more, the Zn(II) ion induces 
metal–coupled folding of the zinc hook domain, which is one of 
the reasons of the extreme stability of the Zn()₂ complex. Kochańczyk 2016; Kochańczyk et al. 2016

The zinc hook domain is not structurally separated from the 
coiled coil domain of , i.e., one cannot separate the particular 
domain just by assessment of the structure like in the case of 
pyruvate kinase ([https://www.rcsb.org/structure/1PKN||1PKN]), 
where three domain structure is easily seen at once Morris, Black, and Stollar 2022
. A similar distinction between the globular domain of the  and 
the superhelical domain is visible without any further 
investigation (see [https://www.rcsb.org/structure/7ZR1||7ZR1]). 
However,  the  marks a priori the stretch of around 100 amino 
acids flanking the canonical CXXC motif to be the zinc hook, 
however,  this approach seems to be experimentally unjustified.

In structural biology, protein domain is understood as a part of 
the protein that is folded independently of the rest of the 
protein, and is stable per se. Combining this with the a priori 
assignment of the zinc hook domain is not justified in the lack 
of the structural discriminant of the coiled coil and the zinc 
hook domain. The marking of the region that makes the zinc hook 
domain should be functional (e.g. assessment of the stability), 
not structural. This distinction is possible due to the 
experimental work performed by Kochańczyk and complemented 
further by Padjasek et al. for the zinc hook domain from  . Kochańczyk 2016; Kochańczyk et al. 2016; Padjasek et al. 2020
 The stabilities studies show that in the  from  only a 
45-amino-acid-long fragment (with the CXXC motif and flanking 
region) is sufficient to form a stable domain, further elongation 
does not result in the stabilization, thus the zinc hook in the  
should be annotated as this short region in case of . The 
stabilities studies performed during my PhD project show that in 
the case of the   the zinc hook domain is longer—an increment in 
the stability of the zinc hook domain is observed up to a length 
of ~130 amino acids (see [chap:hRad50]). Tran, Padjasek, and Krężel 2022
 

From  the first experiments investigating the zinc hook from , 
till the beginning of 2023[a] the zinc hook was thought to be a 
dimerization domain—Zn(II) is bound in a tetrahedron geometry, by 
two cysteines from each  protomer. The [https://www.rcsb.org/structure/1L8D||1L8D]
 Hopfner et al. 2002 structure presented by Karl-Peter Hopfner 
after his postdoctoral fellowship presents the  from  in a 
`V-shaped` conformation, without any possibilities of the  
tetramerization. After 15 years, Park et al. present the 
structure of a zinc hook from  —[https://www.rcsb.org/structure/5GOX||5GOX]
, the structure adopts a totally different shape, compared to the 
V-shaped conformation of the zinc hook from . The newly described 
`rod` conformation is validated by the in-solution experiments 
(inter alia  and ) and stands in agreement with the atomic force 
microscopy of . Park et al. 2017 Four years later Soh, Basquin, and Gruber
 present an almost identical sequence of   to the sequence 
crystallized in the [https://www.rcsb.org/structure/1L8D||1L8D] ([https://www.rcsb.org/structure/6ZFF||6ZFF]
), however,  the zinc hook, instead of adopting the `V-shaped` 
assembly, adopts a `rod-shape`. Soh, Basquin, and Gruber 2021 The 
[https://www.rcsb.org/structure/6ZFF||6ZFF] structure has 
suggested that the zinc hook domain from  has stabilizing 
interactions that are not present in the [https://www.rcsb.org/structure/1L8D||1L8D]
, however,  this does not stand in agreement with the 
experimental data provided by Kochańczyk and Padjasek et al.,Kochańczyk 2016; Kochańczyk et al. 2016; Padjasek et al. 2020
 this is discussed in the Relations between Structure and Zn(II) Binding Affinity Shed Light on the Mechanisms of Rad50 Hook Domain Functioning and Its …
 (see [chap:hRad50]), as the experimental description of the   
zinc hook stability stands in a great agreement with the [https://www.rcsb.org/structure/5GOX||5GOX]
 structure proposed by Park et al., making a great pivot for 
discussion of the relationship between the structure and 
sequence. After 21 years, Hopfner, now as a group leader, 
releases another  structure. Those 21 years of technological 
development allowed for the reconstruction of the complete 
eukaryotic  structure, together in complex with  and  ([https://www.rcsb.org/structure/7ZR1||7ZR1]
). The observed   complex, similar to [https://www.rcsb.org/structure/5GOX||5GOX]
 and [https://www.rcsb.org/structure/6ZFF||6ZFF] presents the  in 
the `rod` conformation. While taking  microphotographs Rotheneder et al.
 observed that the  complex formed conglomerates, connected by 
the zinc hook, however,  due to the low number of particles it 
was impossible to resolve the three-dimensional map of the 
conglomerate, to obtain a structure with better resolution and to 
investigate the  between the apexes, the crystal structure of the 
zinc hook from   was resolved ([https://www.rcsb.org/structure/7ZQY||7ZQY]
). It turned out that the  consist of the dimer of dimers (zinc 
hooks)—the same arrangement that was observed by the . Curiously 
enough the previous eukaryotic zinc hook structure, [https://www.rcsb.org/structure/5GOX||5GOX]
, shows the same tendency in the crystal lattice (when the  is 
duplicated and translated), however,  the propensity to form 
tetrameric assembly was not obvious to Park et al. when the 
structure was published, thus the  presented in the  shows the   
in a dimeric assembly. During the  experiments performed by me on 
the zinc hook from  , I was not able to observe the 
tetramerization, as the radius of gyration stands in accordance 
with the crystal structure. Tran, Padjasek, and Krężel 2022 

The ability to form various architectures by interaction in the 
zinc hook domain suggests that the  complex is able to assume 
conformations that could hold the  fragments together, which 
possible could be one of the steps involved in the  damage repair 
via . The architecture formed by the tetramerization of the  by 
apex-apex interaction could form a structure that is at least 
1000  long, which theoretically should allow the binding of the 
sister chromatids during the . However,  to date the exact 
biological role of the apex-apex tetramerization of the  is 
shrouded in mystery—the apex-apex interface is not conserved 
evolutionarily, and the tethering and its regulation probably 
differ in organisms. Precise in vivo experiments are needed to 
put the  tetramerization in the biological aspect.




Research aims

<ch:Goals>



The unique zinc binding hook from the  protein inspired me to ask 
questions: how many  and  are structurally described. The first 
goal of the research was to provide a survey of intermolecularly 
bound Zn(II) ions available in the . The idea was to 
algorithmically filter for the physiological candidates of , 
followed by the manual assessement of the validity of the 
filtered structures.

The second goal, an expansion of the first idea, was to provide 
freely, and easily-available reso urce of all  deposited in the . 
To do I have developed InterMetalDB, using atomium Ireland and Martin
 and . The service, and its open  are available online.

The third aim (which inspired formulation of two previous ideas) 
of my doctoral project was to perform biophysical 
characterization of the zinc hook domain of the human protein .  
protein, occurs in the  core complex in every domain of life—a 
key complex involved in  damage response. In eukaryotes  core 
complex is stabilized and regulated by an additional interaction 
partner— (or  in ), forming thus the  complex. The conserved zinc 
hook domain in  is an example of unique , where the dimerization 
is strictly dependent on the Zn(II) ion co-cordination by two  
protomers. The formation of  dimer is essential for the proper 
functioning of the , which is directly related to the functioning 
of the whole () complex. 

Finally, the last aim of the research was to investigate the 
potential detrimental effect of the mismetallation of the zinc 
hook from the . This has been done by investigation of the effect 
of Hg(II) and Ag(I) on the zinc hook from the  . The choice of 
the   to be the investigation target was the fact that this model 
is so far best characterized model of the zinc hook available. 






Results And Discussion

Galvanization of Protein–Protein Interactions in a Dynamic Zinc 
Interactome



<chap:Galvanization-of-Protein-Protein>

Anna Kocyła, Józef Ba Tran and Artur Krężel
Department of Chemical Biology, Faculty of Biotechnology, 
University of Wrocław, F. Joliot-Curie 14a, 50-383 Wrocław, 
Poland

Trends in Biochemical Sciences 2021 46(1), 64–79.



6.1 Brief history and bibliometric data

The first chapters presenting published work of mine ([chap:Galvanization-of-Protein-Protein]
 and [chap:InterMetalDB]) focus mostly on in silico endeavors to 
systematically describe, and characterize metal-involved . My 
contribution in neither of the articles would not be possible 
without involvement in the wet lab experiments from the very 
beginning of my master's and doctoral studies. My ongoing journey 
with computer programming began as a mere way to make my 
laboratory work easier. Producing countless amounts of data, I 
had to prepare scripts to pre-process the raw data or workflows 
to facilitate data analysis, some of those scripts were made at 
the request of my coworkers (e.g. the script that was part of the 
fluorescence lifetime analysis workflow), some already may solve 
the problems of the researchers on the other continents.[footnote:
Allowing for batch export of the data from  like [https://github.com/jzftran/jws2txt||jws2txt]
, mentioned at: 
https://www.researchgate.net/post/Batch_export_processing_of_Jasco_interval_data_files_jwb_to_csv
.
] My interest in programming led to a collaboration with Anna 
Kocyła on the Galvanization of Protein--Protein Interactions in a Dynamic Zinc Interactome
, where I was able to use my skills in preparing a structural 
data set of . After the few years I have been back to scripts 
written for our publication, I am pleased with how much I have 
grown in terms of writing clear and efficient . And though the  I 
have written at the time was not beautiful, to say, it worked and 
was one of the reasons why we were able to select around 150 
examples of the  deposited in the  and publish our results in a 
well-renowned journal Trends in Biochemical Sciences, the  of 
14.264 (as of March 6, 2023). [footnote:
https://scholar.google.com/
]

6.2 Summary

The functions that are performed by macromolecules in the cell 
are dependent on the interaction with their partners. The role of 
the macromolecule depends on what the interaction partner is. For 
example, in the case of glucose-6-phosphate isomerase, one of the 
interaction partners for the enzyme is glucose 6-phosphate, which 
is interconverted into fructose-6-phosphate. Berg, Tymoczko, and Stryer 2007
 Other proteins interact with lipids, nucleic acid, and other 
proteins. It is not an exaggeration to state that virtually every 
protein has to interact, transiently or permanently, with other 
macromolecules to realize its function in the organism. The 
holistic knowledge  about these interactions is used to plot the 
map of the interactions, which in turn can identify the functions 
of the proteins and their interdependence leading to 
understanding the cellular function. Cafarelli et al. 2017 From 
this network of interactions,  are crucial to gain comprehension 
of the functional relationships between proteins. Over the last 
decades, the development of  methods e.g. peptide phage display ,Sundell and Ivarsson 2014
 yeast surface display,Boder and Wittrup 1997 yeast-two-hybrid 
system, Fields and Song 1989 and others lead to heaping amount of 
data. Together with huge amount of data a large number of 
databases have been founded to curate and annotate the . Patil 2019
 

However, none of the vast interaction networks presents a full 
landscape, as interactions induced by small ligands as well as 
caused by metal ions may be missing from the picture. What is 
worth mentioning here, is the rough projection of how many 
proteins associate with metal ions – it is estimated that roughly 
30-40% of all proteins interact with metal ions. Andreini et al. 2008; Andreini et al. 2013; Waldron et al. 2009
 Taking into account that bioinformatic analyses estimate that in 
human  every tenth protein is a zinc-binding protein Andreini, Bertini, and Rosato 2009; Bertini, Decaria, and Rosato 2010
, and the fact that most of the  methods do not even consider the 
presence of another actor (i.e. Zn(II)) in the interaction, thus 
causing such interactions to be overlooked, makes a substantial 
amount of interactions not fully resolved. Although metal ions, 
compared to macromolecules, are small in both mass and size, they 
have a significant impact on protein structure, which implies 
consequences for the stability and function of proteins and 
protein complexes. Though the impact of various metal ions on 
living organisms has been studied for a while, and the effects of 
deficit or surplus of some microelements are known, the mechanism 
of action of those elements may be still vague. As has been 
described in [chap:Zinc proteome] [chap:Zinc proteome] Zn(II) can 
be found in various architectures. The studies of how the 
architecture translates into function led to the identification 
of catalytic and structural zinc-binding sites, however, the fact 
that Zn(II) can be located on the  is often overlooked. 

Our goal in this report was to provide a comprehensive 
description of . We wanted to underline that the presence of 
Zn(II) at protein–protein  plays a role in orchestrating the 
complex stability, adds additional level of the protein complex 
regulation, and despite the fact that it seems to be quite a rare 
phenomenon the presence of Zn(II) is crucial for fulfilling 
several essential for the organism physiological functions. As a 
showcase example – Zn(II)–protein complex, essential component of 
the circadian clock, can be evoked, where interdependent Zn(II) 
complexation and disulfide bond formation modulates the complex. Schmalen et al. 2014
 Our review article, attached bellow presents more fascinating  
examples where Zn(II) is essential for proper functioning of the 
protein complexes.

The description of outstanding  examples is only one part of our 
review, the formation of  is governed by the same thermodynamic 
and chemical principles as regular protein-protein complexes, 
however, the involvement of Zn(II) in the complexation makes the 
formation of complex dependent on the . This dependence of 
complex formation on an additional factor might be unclear, thus 
to introduce the reader into the world of chemical equilibria we 
discuss the effect of availability of the Zn(II) and its 
fluctuations caused by Zn(II) influx and efflux, and its effect 
on formation of . Because the mechanisms of action of various 
zinc transients are still unknown, our contemplation on the 
subject is purely hypothetical, nevertheless, we can deduce that 
terra incognita of Zn(II) fluctuations and the interdependence of 
 formation  is a vast, worth investigating field.

For sure the impact of the review would have been lower if no 
thorough survey on  was made. I have decided to collect 
information about discovered protein-protein complexes with 
Zn(II) on the . So how does one search for ? How common is zinc 
at protein-protein ? It is known that characteristic sequence 
motifs involving cysteines, histidines, and acidic residues[footnote:
glutamates and aspartates
][margin:
Single-letter code for these amino acids and their frequency in 
Zn(II)-binding sequences was an inspiration for the tool “CHED”. Babor et al. 2007
 
] that provide a clue to Zn(II) coordination within a 
single-chain protein. Yet, the prediction, based solely on the 
protein sequence, of metal binding on the protein-protein  to 
date is not possible. One way of identification would be a  
thorough literature search, one could hypothetically search the 
Internet using research tools for scientific literature like 
Google Scholar[footnote:
https://scholar.google.com/
], Semantic Scholar[footnote:
https://www.semanticscholar.org/
], or other similar tools utilizing keywords “zinc”, “
intermolecular binding”, “protein–protein ”. However, this way of 
analysis seems to be cumbersome, and requires tremendous amount 
of manual work, not to mention that the found literature might 
not be available. The only logical solution to the problem is to 
search online databases that provide information about proteins. 
Usually, for the researcher the first source of information about 
the protein of interest is  Bateman et al. 2022, a significant 
amount of protein sequences aggregated in  are properly annotated 
to interact with metal ions, including Zn(II) ion, alas,  does 
not provide the information whether the metal ion is bound on the 
protein-protein . Taking into account the fact of non-existence 
of metal ions at protein-protein  and the fact that it is not 
possible (so far) infer the intermolecular binding based on the 
sequence solely I have decided to use one of the most 
comprehensive resources available for any structural biologist – 
the  Berman, Westbrook, and Feng 2000; Burley et al. 2021. The 
reason why I used  as a source of information, instead of already 
established databases containing structures of metal ion-protein 
complexes (e.g. MetalPDB Andreini et al. 2013, ZincBind Ireland and Martin 2019
) is shortly discussed in the review. Using the data from  and 
simple  scripts I was able to automate much of the required 
effort.

One of the biggest pitfalls of our work, as well as similar 
studies (e.g. Laitaoja, Valjakka, and Jänis 2013; Song et al. 2014
), is the fact that the gathered knowledge simultaneously with 
the acceptance of the article stops from being updated. It does 
not mean it becomes automatically outdated, the extensive 
description and analysis of the gathered and reviewed examples 
stay scientifically true (until the next breakthrough in the 
field), however, the lists of structures, and interwoven 
statistics and numbers change over time. From the publication 
date of Galvanization of Protein--Protein Interactions in a Dynamic Zinc Interactome
  (accessed ) is richer with  Zn(II)-containing structures! To 
partially mitigate this deficiency I have developed InterMetalDB 
(see [chap:InterMetalDB]). Tran and Krężel 2021



The bioinogranic chemistry literature rarely divagate whether 
found Zn(II) is bound in a protein-protein interface, this is 
similar to classical biochemistry and structural biology 
textbooks, which if any, mention only briefly that metal ions may 
be involved in stabilization of quaternary protein structures. As 
has been mentioned in [chap:Proteins-and-cofactors] around 10% 
human proteins interact with Zn(II), and almost every protein 
have to interact with other macromolecule in order to fulfill its 
function. So why only short mention that Zn(II) can be bound by 
two or more macromolecules? Is this because the formation of  is 
a rare phenomenon or is this due to the fact that the studies of 
metals at protein-protein  is difficult to systematically study? 
Our opinion is the latter, which motivated us to describe known . 
However, many questions still are without any answer, making the 
search and study of  a terra incognita for bold researchers.













InterMetalDB: A Database and Browser of Intermolecular Metal 
Binding Sites in Macromolecules with Structural Information



<chap:InterMetalDB>

Józef Ba Tran and Artur Krężel
Department of Chemical Biology, Faculty of Biotechnology, 
University of Wrocław, F. Joliot-Curie 14a, 50-383 Wrocław, 
Poland

Journal of Proteome Research 2021 20(4), 1889–1901.





7.1 Brief history and bibliometric data

The work on the Galvanization of Protein--Protein Interactions in a Dynamic Zinc Interactome
 ([chap:Galvanization-of-Protein-Protein]), among other things, 
included the preparation of the data set containing all deposited 
 in the . To manipulate  files, and filter for  I have used the 
BioPython library Cock et al. 2009. My primary school art teacher 
used to say “If something is for everything, it's really for 
nothing.”, the similar is with BioPython, which has a wide range 
of tools used in bioinformatics, however, the offered  is rather 
clumsy in use, and BioPython itself lacks some more advanced 
functions to handle  files. Combined with the restructuration of 
the  search  and often questions from Ania Kocyła if the data set 
is ready I knew that the next project that will involve data from 
 will have to offer an easier in use libraries. Around this time 
a new  library emerged that allowed for efficient and easy  file 
parsing—atomium Ireland and Martin 2020, which allowed for easy 
manipulation with  files, building a better logic behind the 
database, and making the whole process scalable. 

The article describing InterMetalDB was accepted and published 
only after minor revision in the Journal of Proteome Research ( 
of 5.370, March 6, 2023) in 2021, and to date () the article was 
cited [footnote:
https://scholar.google.com/
]

7.2 Summary

The fact that the metal ions can be bound at the  that is created 
by two or more interacting macromolecules has been missing 
interest from researchers (see [chap:Galvanization-of-Protein-Protein]
). The first systematic review of metal ions bound at 
protein-protein  was published in 2014 Song et al. 2014, next, 
similar article (which was in part my creation) was published 
seven years later, however, the article was concerning only 
intermolecularly bound Zn(II) ions. Kocyła, Tran, and Krężel 2021 
This stands in opposition to studies devoted to complexes with 
stoichiometry of 1:1 – similar database surveys, generally 
concerning Zn(II)[footnote:
Probably due to its unique properties and abundance in . See [chap:Zinc proteome]
. 
], were done numerously in the past. Alberts, Nadassy, and Wodak 1998; Auld 2001; Patel, Kumar, and Durani 2007; Dokmanić, Šikić, and Tomić 2008; Zheng et al. 2008; Lee and Lim 2008; Hsin et al. 2008; Sousa et al. 2009; Andreini, Bertini, and Cavallaro 2011; Laitaoja, Valjakka, and Jänis 2013
 As mentioned in [chap:Galvanization-of-Protein-Protein] one of 
the biggest pitfalls of such studies is the fact that 
simultaneously with the publication of an article, the 
meticulously gathered, and classified[margin:
Laitaoja, Valjakka, and Jänis have manually analyzed around 7800 
Zn(II)-binding sites! Laitaoja, Valjakka, and Jänis 2013
] knowledge is no longer updated, combined together with the fact 
that usually the generated data set and the methodology is not 
easily and publicly available makes such studies fleetly 
outdated. This observation, which happened to occur to me during 
my work on the Galvanization of Protein--Protein Interactions in a Dynamic Zinc Interactome
 ([chap:Galvanization-of-Protein-Protein]), sprouted into an idea 
that the data set I will generate should be easily available to 
anyone, free of charge and updated even after the publication of 
the article. This idea growth further, eventually bore fruit – InterMetalDB: A Database and Browser of Intermolecular Metal Binding Sites in Macromolecules with Structural Information
[footnote:
https://intermetaldb.biotech.uni.wroc.pl/
].

My intention for the creation of InterMetalDB was to provide a 
tool, which would make researchers aware that macromolecules can 
bind metal ions at the . The goal of InterMetalDB is to search 
for macromolecular structures containing metal ions bound at the 
interface. By doing so InterMetalDB fills the free niche, that 
was left unoccupied for a long time. Because the process of data 
acquisition, analysis, and presentation is automated it does 
require little or no human intervention. The biggest issue with 
this approach is the fact that it is not possible to filter out 
adventitiously bound metal ions[margin:
One of many sources of adventitious metal ions in the structure 
may be a buffer used during the protein crystallization.
], as it would require tedious manual investigation of the 
structure, followed by (often) extensive literature research.

What is worth noting is how the InterMetalDB is built. Contrary 
to preceding databases (e.g. MESPEUS Hsin et al. 2008, ZifBase Jayakanthan et al. 2009
, MetalPDBAndreini et al. 2013; Putignano et al. 2018), 
InterMetalDB operates on . Compared [margin:
Clustering proteins based on their sequence identity is a method 
that reduces or removes redundancy in fetched protein structures.
] to Galvanization of Protein--Protein Interactions in a Dynamic Zinc Interactome
 study I have decided to use clustering based on 50% of identity[footnote:
Protein sequences that are in the cluster are have share 50% of 
their sequence identity.
], instead of 95% like it have been done previously. The gentle 
clustering threshold of 50% sequence identity ensures that the 
structures in the cluster will share the same . Instead of 
selecting the representative structures, InterMetalDB marks “
representative” metal binding sites, this is done by lookup what 
residues are in the radius of 5 , a concept similar to the 
Minimal Functioning Site adapted in the MetalPDB Andreini et al. 2013
, followed by the selection of the metal binding site 
characterized by the best resolution or if the resolution is 
equal across the cluster, selection of the first.   The exact 
methodology how the data set is generated is described in details 
in the article, the  is available under MIT License[margin:
The MIT License is one of the most permissive license used for 
software licensing.
] at https://github.com/jzftran/InterMetalDB. Making the  
published under the MIT License makes the InterMetalDB reusable 
by other researchers, it allows any part of the code to be used 
in another project. 













Relations between Structure and Zn(II) Binding Affinity Shed 
Light on the Mechanisms of Rad50 Hook Domain Functioning and Its 
Phosphorylation



<chap:hRad50>

Józef Ba Tran, Michał Padjasek and Artur Krężel
Department of Chemical Biology, Faculty of Biotechnology, 
University of Wrocław, F. Joliot-Curie 14a, 50-383 Wrocław, 
Poland

International Journal of Molecular Sciences 2022 23, 11140.



8.1 Brief history and bibliometric data

My early work as a researcher of protein-metal interactions 
involved joint work with Michał Padjasek on the Rad50 protein 
from  (data not published yet). Padjasek 2022 It was Michal 
Padjasek who largely showed me the ins and outs of the 
experiments that give insight into the world of . It was this 
knowledge that allowed me to study the behavior of the zinc hook 
from the human . Historically, my participation in the research 
is part of a larger project, already separated into several 
grants, devoted to studying the orthologs of the zinc hook from  
protein. This research was initiated by Tomasz Kochańczyk and 
continued by Michał Padjasek, myself, Olga Kerber, and Marek 
Łuczkowski. As it turned out, the zinc hook from the human  
behaves in a different way than its orthologs from  and . The 
hard work on this subject resulted in a publication in the 
International Journal of Molecular Sciences, a journal with an  
of 6.208 (as of March 6, 2023) 

8.2 Summary

 play a crucial role in many biological processes and are usually 
stabilized by an extensive network of non-covalent intermolecular 
interactions. An additional factor contributing to the formation 
of protein complexes is intermolecular metal binding, in which 
the affinity of the metal for the protein affects the stability 
of the entire complex. Most of the metal-involved  are 
uncharacterized in terms of the role of metal at the interface or 
in terms of the complex stability. One of few, quite well 
characterized  is the interaction observed in the zinc hook 
domain of the . Zn(II)- complex. Previous studies have 
investigated the relationship between the structure of the   and 
Zn(II)-complex stability. Kochańczyk 2016; Kochańczyk et al. 2016; Padjasek et al. 2020
 Those studies show that the complex stabilization occurs 
primarily in a small 45-amino-acid-long fragment consisting of 
the CXXC binding motif and the flanking region. In 2017[a] 
publication appeared that shifted the paradigm about the 
architecture of . Park et al. 2017 The newly published structure 
( 5GOX) presented the rod-shape of the  , and as it usually is 
with new data, new questions arise. How does the rod-shaped 
architecture of  determines the complex stability? What are the 
elements that contribute to the  stability? Are there any other 
factors that influence the formation and the complex stability?

We have hypothesized that increasing the length of the peptides 
being studies, similarly to  Rad50 Kochańczyk et al. 2016, would 
increase the stability of the complex by increasing the number of 
inter- and intra- interactions. To test this hypothesis we used 
potentiometry and competition titration with spectropolarimetric 
detection. The results showed that in contrast to  from  the   
Zn(II)-complex stabilization occurs over much longer distance. 
The observation we have made show that the human  forms the 
strongest known Zn(II)-protein known so far. This extremely 
elevated stability is due to similar effect that was observed in 
the   – However, the extent of the effect in human  is much 
greater, the stabilization occurs over the length of 140 
amino-acids, which means that the Zn(II)-complex is stabilized by 
protein regions that are approximately 60  from the Zn(II) 
binding site. On the other hand, while trying to determine the 
factors that may modulate the Zn(II) complex stability we have 
investigated the effect of the phosphorylation at threonine 690 
of human  zinc hook domain. The phosphorylation site is located 
closely to the Zn(II) binding site, at the apex of the zinc hook. 
We hypothesized that the phosphorylation in close proximity of 
Zn(II) binding site will strongly influence the Zn(II) affinity 
towards the zinc hook. Our hypothesis turned out to be true, 
phosphorylation of short, 42-amino-acid-long fragment diminish 
the stability of the Zn()₂ complex by approximately one order of 
magnitude, However, this effect is attenuated in the 
183-amino-acid-long fragment and the conditional formation 
constant is lower by one-fourth order of magnitude. This 
apparently considerable changes in stability do not have huge 
impact on the formation of the Zn()₂ complex. The observed shift 
of the complex stability, concurrent with observable by direct 
Zn(II)-titration of 42-amino-acid-long peptide with extensive 
secondary structure change (detection by ), was not observable 
for the 183-amino-acid-long protein fragment by . Neither local 
changes at the Zn(II)-binding site was observable by means of . 
The fact that the phosphorylation effect is attenuated in the 
longer protein fragment, and the fact of the extreme complex 
stability, the phosphorylation do not exert the complex 
dissociation under physiological  concentration, However, 
phosphorylation of the threonine 690 may be involved in 
undiscovered yet signaling pathway.

This study provides new insights into the stability and 
regulation of the protein-protein interactions mediated by zinc 
binding sites (). The unique interaction of the human Rad50 
protein with zinc ions and the effect of phosphorylation on the 
stability of the complex highlight the importance of studying 
these interactions in detail. Further research on the mechanisms 
underlying the stability of protein-protein interactions mediated 
by zinc binding sites may lead to a better understanding of the 
biological processes they regulate and the development of new 
therapeutic strategies targeting these interactions.











An Extremely Stable Interprotein Tetrahedral Hg(Cys)₄ Forms in the 
Zinc Hook Domain of Rad50 Protein at Physiological pH



<chap:HgRad50>

Marek Łuczkowski[superscript:*], Michał Padjasek[superscript:*], Józef Ba Tran[superscript:*], Lars Hemmingsen[superscript:†], Olga Kerber[superscript:*], Jelena Habjanič[superscript:‡], Eva Freisinger[superscript:‡] and Artur Krężel[superscript:*]
[superscript:*]Department of Chemical Biology, Faculty of Biotechnology 
University of Wrocław, F. Joliot-Curie 14a, 50-383 Wrocław, 
Poland

[superscript:†]Department of Chemistry, University of Copenhagen, 
Universitetparken 5, 2100 København Ø, Denmark

[superscript:‡]Department of Chemistry, University of Zurich, 
Winterthurerstrasse, 190, 8057 Zürich, Switzerland

Chemistry – A European Journal 2022 28, e202202738





9.1 Brief history and bibliometric data

The last two chapters ([chap:HgRad50] and [chap:AgRad50]) 
presenting the results of my doctoral project focus on the 
mismetallation of  from . My participation in the first of this 
studies is quite astonishing, I do believe that “Insanity is 
doing the same thing over and over again, but expecting different 
results.”[footnote:
A quote wrongly attributed to Albert Einstein.
], and asking me to perform the same  experiments because the 
results are not in agreement with the initial hypothesis seemed 
to me as a waste of time, however, for some reason, maybe I had a 
hunch, I agreed to perform again the same experiments. My task 
was to conduct  experiments that would give information about the 
stability of the Hg(II)– complex. The first experiments I have 
performed gave exactly the same results as those obtained by my 
colleagues. The observed  of binding of Hg(II) to the zinc hook, 
assuming that the  of binding of Zn(II), Cd(II) and Hg(II) were 
the same, was too small for the observed lack of competition 
between the competitor (e.g., acetylcysteine, cyanides) and the 
zinc hook to be explained by the formation constant calculated 
using the measured  and the  estimated by other experiments (see [eq:deltaG]
, [eq:GandK] and the rest of the chapter). 

By default, all  experiments, I and my predecessors have 
performed, were performed in buffers containing  from  intended 
to provide ionic strength. In the case of other measurements made 
with Zn(II), Kochańczyk et al. 2016 or Cd(II), Padjasek et al. 2020
 this did not make any difference, but in this case it did. 
Browsing the literature in search of constants of formation of 
mercury complexes with various compounds, I came across 
information on enthalpies of formation of Hg(II) complexes with 
chlorides, eureka!, I have calculated the chemical speciation of 
Hg(II) and  complexes, and along with the speciation I have 
calculated the sum of enthalpies of formation of the complexes. 
According to Hess's law, the enthalpy of formation of the Hg()₂ 
complex should be diminished by the enthalpy of formation of 
various Hg(II) complexes with chlorides. One more I have repeated 
the  experiments using a different mercury salt and a different 
buffer (-free), unsurprisingly the difference in the observed 
enthalpy was consistent with my theoretical calculations. The 
measured enthalpy allowed me to estimate the formation constant 
of the Hg(II) and  complex. What was very nice, was the fact that 
the reviewers assigned by Chemistry – A European Journal ( of 
5.020, March 6, 2023) have pointed out an interesting application 
of  for measuring the complex stability. [footnote:
https://scholar.google.com/
]

9.2 Summary

The following two articles described in this Chapter and [chap:AgRad50]
 focus on the phenomenon of mismetallation of  zinc hook domain 
by other than Zn(II), potentially toxic metal ions: Hg(II) and 
Ag(I) (the later described in [chap:AgRad50]). It is a well-known 
fact that mercury contamination affects both ecosystem and human 
health. Mercury pollution stems from both sources – natural and 
anthropogenic, where after its release into the environment 
mercury undergoes various (physical and chemical, often 
redox-related changes). Driscoll et al. 2013 Mercury poisoning 
results as a consequence of exposure to mercury, the severity of 
the poisoning, as usual, is dependent on the dose and the length 
of the exposure. Another factor that impacts the poisoning 
severity is the form of the mercury – various mercury forms are 
characterized by diverse absorbability and reactivity towards 
biological targets, e.g. methylmercury cation [margin:
Methylmercury cation is a major source of mercury poisoning in 
human Bernhoft 2012.
] is easily absorbed into the brain, liver, kidneys, etc., while 
Hg(II) for various reasons, including the reactions with 
sulfhydryl groups on erythrocytes or glutathione in the plasma 
does not easily cross the blood-brain barrier Bernhoft 2012. The 
mercury accumulation mostly in mitochondria and in the nucleus 
was already proven years ago Bucio et al. 1999; Christensen, Mogensen, and Rungby 1988; Yoneyama, Sharma, and Kleinschuster 1985; Königsberg et al. 2001
. Inside the cell, organic mercury may be demethylated into 
inorganic mercury Suda and Takahashi 1992.  It was proven that 
the Hg(II) ions (a product of demethylation) may inhibit the 
activity of transcription factors that contain zinc fingers, 
however, the exact mechanism of how Hg(II) disturbs the function 
of the transcription factors is not known, as no data how mercury 
is coordinated by those transcription factors were provided Rodgers et al. 2001
. One of the mechanisms of the inhibition of the transcription 
factors could be the formation of the aberrant metal binding 
site, as Hg(II) in the aqueous environment tends to form linear 
complexes, however, the formation of complexes with higher 
coordination number is possible, those complexes usually are 
distorted, with T-shaped geometry of the complex being an extreme 
example of the three-coordinated species Chakraborty et al. 2010. 
In  living organisms, correspondingly to the , Hg(II), as a soft 
Lewis acid will preferentially react with sulfhydryl groups (soft 
Lewis bases), therefore it is not surprising that the cell 
detoxifying mechanisms copiously utilize thiolates to sense (), 
sequester (e.g. metallothioneins), and to enzymatically process 
mercury compounds to less toxic forms or to Hg(0) Ledwidge et al. 2010
. [margin:
Hg(0) is the major form of mercury emitted to the atmosphere. 
Hg(II) buried in sediments may be reduced to Hg(0), which can be 
re-emited to the astmosphere Christakis, Barkay, and Boyd 2021.
] In the cytoplasm Hg(II) will preferentially form diagonal 
complexes with low molecular weight compounds with thiol groups 
(i.e. cysteine, glutathione), however, when Hg(II) interacts with 
the protein, like in the case of Hg(II)-substituted rubredoxin, 
the protein may already impose a structurally defined metal 
binding site, where the Hg(II) will adopt a tetrahedral or 
pseudotetrahedral architecture Faller et al. 2000. The observed 
coordination in the crystal structure does not necessarily 
correspond to the Hg(II) binding site formed in the solution, as 
has been verified in the case of the —contrarily to the crystal 
structure ([https://www.rcsb.org/structure/1FE4||1FE4], dimer, 
crystallized at pH 6.5) in solution (pH 7.5) Hg(II) prefers to 
form linear complexes with monomeric protein Łuczkowski et al. 2013
. 

How does the coordination look like in case of  mismetallation by 
Hg(II)? Does the crystal structure of the zinc hook domain from  
, which was the sole model of the  zinc hook for 15 years really 
contain HgCys₄? Is the Hg(II) substituted zinc hook a proper model 
for Zn(II) loaded   zinc hook? Hopfner et al. claim that the 
presented by them zinc hook Hg(II) substituted structure ( 1L8D) 
is isostructural with Zn(II)-loaded structure, however, the 
Zn(II)-containing structure had lower resolution, which in turn 
prompted the authors to deposit in the  the Hg(II)-containing 
structure only. Hopfner et al. 2002 Olga Kerber as part of her 
doctoral project was able to obtain in similar conditions 
Zn(II)-loaded   (data unpublished as of ). The structure of 
Zn(II)-loaded  is indeed isostructural with the model proposed by 
Hopfner et al., however, it does not unveil the mystery of the 
tetrahedral Hg(II) coordination.

To answer the question about the tetrahedral Hg(II) coordination 
experiments performed in solution were required – the set of 
experiments consisting of titrations with  spectroscopic and 
spectropolarimetric detection, fluorescence anisotropy,  
spectroscopy and [superscript:199]Hg . All of those mentioned methods clearly indicate the Hg(II) 
coordination by four cysteines, two of each originating from each 
 protomer involved in the Hg(II) binding. The fluorescence 
anisotropy, largely performed by Michał Padjasek, showed that 
even 100\times excess of  over the 45-amino-acid-long peptide did 
not exhibit any traces of competition between Hg()₂and Hg( )ₓ, [margin:
The stability of Hg()ₓ complex(es) is not known, however, it can 
be assumed that the stability constant(s) are comparable to 
stability constants of other compounds with two thiol groups, 
e.g. Hg-dimercaprol complexes.
] which suggests the extreme stability of the Hg()₂ complex. 

How to measure, or at least estimate extremely high affinities? 
The answer to the question is somewhat perverse since the answer 
is . [margin:
 usually is not suited for measuring either extremely high or 
extremely low affinities.
] When tight affinities are measured with  usually the experiment 
that is performed implies the use of a well-characterized 
competitor, nevertheless in the described issue I have leveraged 
the fact of the isostructurality of the Hg(II) and Zn(II) 
structure (and probably Cd()₂ complex as well Padjasek et al. 2020
), which implies that the entropic effect that accompanies the 
metal complexation should be very similar, and allow the 
calculation of  if  and  were known, according to the Equation [eq:deltaG]
 where   denotes absolute temperature.

\Delta G\text{°}=\Delta H\lyxmathsym{\textdegree}-T\Delta S\text{°}

Using the relationship described by Equations [eq:GandK] ( 
denotes molar gas constant) and [eq:K12] I was able to calculate 
the , where  is Hg(II) and  is zinc hook from  , described as Hk 
in Equations [eq:K12] and . Equation [eq:K12] describes the 
cumulative association constant of the process described by 
equation .

\Delta G\text{°}=-R\cdot T\cdot\ln K_{12}

K_{12}=\frac{[\mathrm{HgHk_{2}}]}{\mathrm{[Hg(II)]\cdot[Hk]^{2}}}



The described methodology allowed me to estimate the highest 
known (so far) stability of Hg(II) and protein complex, not just 
any kind of the complex, but the first described, 
pseudotetrahedral Hg(II)-complex formed at the , that is formed 
preferentially at pH 7.4. 

Extremely high Hg(II) affinity towards  causes the replacement of 
physiologically occurring Zn(II) in the zinc hook domain. It 
would be an exaggeration to claim that the replacement of Zn(II) 
by Hg(II) is the sole reason for Hg(II) genotoxicity, especially 
taking into account the fact that the Hg(II)-saturated   is 
isostructural to the Zn(II) loaded protein. However, one cannot 
rule out the possibility that the affinity of the Hg(II) is not 
detrimental to the functioning of the  complex – not without 
reason the affinity towards Zn(II) across various organisms 
differ, reaching higher values in case of  from thermophilic  
when compared to  from . It is worth noting that somehow between 
the globular domains of  and the zinc hook domain communicate, 
i.e. even a single point mutation in the coding sequence of the 
zinc hook [margin:
In eukaryotes, the globular domain of  and the zinc hook can be 
as far apart as 600 ! Hopfner et al. 2002
] may abolish some of the functions fulfilled in the globular 
domains of . Lafrance-Vanasse, Williams, and Tainer 2015 Maybe 
the Hg(II) complexation by  somehow increases the rigidity of the 
protein, which in turn prevents the allosteric signal 
transduction between the zinc hook and globular domains? Maybe 
the increased complex stability somehow shifts the exchange of 
subunits (note that the association of the  with any metal ion is 
an equilibrium process as described by ), thus potentially 
influencing the  flexibility? Thangaraj et al. 2019; Merickel et al. 2002
 This and many other questions remain to be answered, however, 
the first problem to tackle should be an assessment of whether 
Hg(II) binding in the zinc hook domain influences any of the 
various functions filled by the .











Zn(II) to Ag(I) swap in Rad50 zinc hook domain leads to highly 
stable interprotein complex disruption through the formation of 
Agₓ(Cys)[subscript:y] cores



<chap:AgRad50>

Olga Kerber[superscript:*], Józef Ba Tran[superscript:*], Alicja Misiaszek[superscript:*], Aleksandra Chorążewska[superscript:*], Wojciech Bal[superscript:†] and Artur Krężel[superscript:*]
[superscript:*]Department of Chemical Biology, Faculty of Biotechnology, 
University of Wrocław, F. Joliot-Curie 14a, 50-383 Wrocław, 
Poland

[superscript:*]Institute of Biochemistry and Biophysics, Polish Academy of 
Sciences, Pawińskiego 5a, 02-106 Warsaw, Poland

Inorganic Chemistry 2023 62, 4076–4087



10.1 Brief history and bibliometric data

Bringing this article to publication was, of the other articles I 
co-authored, the hardest. Initially, we wanted to publish the 
article in Chemical Communications, as a continuation of the 
project devoted to the investigation of the role of Ag(I) on the 
intraprotein zinc-binding sites, where previously an article 
about the formation of  clusters in  was described. Finally, 
after several iterations involving the exchange of criticism from 
reviewers, the manuscript, and responses to the reviewers' 
objections, the article was published in Inorganic Chemistry ( of 
5.436, March 6, 2023).
I have mixed feelings as I think about the publication of this 
article. My reservations about the peer-review process became 
even stronger after this article. In the 21st century, with the 
ability to exchange information cheaply, and expressly, and where 
the cost of publishing information online is very low, the 
peer-review process as a rule should be completely transparent. 
In addition, the rule should be the possibility to criticize an 
article after it has been published, without having to write 
another article in which, one refers to the criticized 
publication. The publication process for this article was quite 
tedious due to the incompetence of one of the reviewers—the 
prolonged review did not bring anything new, did not improve our 
article, but only increased the time in review. Also, this 
article, is the last article included in this thesis and I am 
very pleased to have collaborated on it with my lab-mates, whom I 
saw every day (provided there were no pandemic restrictions). 
This successful cooperation between us was one of the reasons why 
the time spent on experiments, data processing, and writing the 
manuscript, was far less than the time this article spent in 
review.
[footnote:
https://scholar.google.com/
]

10.2 Summary



The medical use of silver, both as an Ag(0) and as compounds of 
Ag(I) is known for millennia, the ancient Persians, Phoenicians, 
Greeks, Romans, and others were recorded to utilize various forms 
of silver to preserve both food and water, the custom practiced 
through World War II. The first preserved records of the medical 
use of silver come from ancient Macedonians, who attempted to 
treat infections with silver plates. Hippocrates, the Father of 
Medicine, was supposed to use silver preparations to treat ulcers 
and promote wound healing. Before the introduction of 
antibiotics, silver, and various silver compounds were the most 
available antimicrobial agents. Alexander 2009 The toxicity of 
silver is low compared to most of the other heavy metals, which 
allows the use of silver in various utility goods, e. g. jewelry, 
silverware, or as a food additive. However, it has been shown 
that high doses of silver given parenterally, can cause 
convulsions or even death. Alexander 2009 Though the silver 
toxicity is low, the organisms usually lack the mechanisms for 
silver detoxification, and in the case of the human body, when 
silver is delivered (e.g., injection, ingestion, or applied 
topically) silver is accumulated irreversibly. The condition 
caused by sustained silver exposure (and light[footnote:
Like in a photography sunlight causes the decomposition of 
colorless silver compounds to Ag(0) or silver sulfides.
]) is called argyria. By 1939 Hill and Pillsbury have identified 
357 cases of recorded argyria, all of the recorded cases, except 
for the irreversible bluish skin tone, had a benign course of the 
disease. Hill and Pillsbury 1939 Despite low toxicity, the number 
of publications dedicated to silver toxicity by the year 
increases ([fig:ag_hg_count]), possibly the reason for this is 
the use of silver in everyday articles (e.g., use of silver in 
deodorants, sportswear, anti-smoking pills), of which a lot of 
these items utilize .

[float Figure:
<Graphics file: C:/Users/makro/Desktop/PhD/gfx/ag_hg_count.png> [Rysunek 10.1: 
<fig:ag_hg_count>Number of publications devoted to silver or 
mercury toxicity aggregated in PubMed (accessed ) Sayers et al. 2022
, blue and orange lines respectively. Note that since 2013 
(marked as a dashed line) more publications are devoted to the 
toxicity of silver than mercury.
]
]

The oligodynamic effect, which is the property of some metals to 
have a biocidal effect, in the case of silver is mainly caused by 
the bioactive species – Ag(I). The antimicrobial properties of  
mainly arise due to the oxidative dissolution of  in the cell, 
followed by the release of Ag(I) Lok et al. 2007; Hsiao et al. 2015
. In biological systems, the  thiols (e.g., glutathione) are 
found to be a common interaction partner for Ag(I) Betts et al. 2021; Marchioni et al. 2018
. The Ag(I) complex with cysteine is much more stable than the 
complexes of Ag(I) and methionine or histidine, thus no wonder 
that in the  (proteins), the common target for Ag(I) binding will 
be systems containing cysteines, e.g., natural 
cysteine-containing zinc-binding sites. The interaction of Ag(I) 
with intraprotein zinc-binding sites in  has been shown to swap 
Zn(II) for Ag(I) leading to - disruption, and in turn, it can 
potentially lead to loss of function. Despite the fact that the 
cysteine is the most common  in the zinc-binding sites (see [sec:Zinc-Binding-Sites]
), Laitaoja, Valjakka, and Jänis 2013 the potential interactions 
of the zinc proteome with Ag(I) are yet unexplored, including the 
impact of free Ag(I) ions on the , an interaction that is present 
in the zinc hook of the . The  forms Zn()₂ complex of extreme 
stability [chap:hRad50]), however, in this case the stability 
does not comes with selectivity, thus the Zn(II) is easily 
replaced, in accordance with , by the `softer` ions (e.g., Cd(II) 
or Hg(II)) from the tetrathiolate coordination sphere of  (see [chap:HgRad50]
). This lack of selectivity makes  a potential target for 
genotoxicity induced by heavy metal ions.

The following article present the influence of Ag(I) on the zinc 
hook domain from the  —the use of the model was dictated by tthe 
zinc-binding sites (seehe fact that the zinc hook  from  is the 
best described zinc hook model so far. In the study we have 
utilized the 14- and 45-amino acid long fragments with the CXXC 
motif as a minimal and full domain model (see definition of the 
zinc hook domain in [sec:The-zinc-hook]), respectively. The used 
constructs allowed us to get insight into the coordination and 
stability properties of the Ag(I)- complexes. To do so we have 
applied array of biophysical methods like: ,  spectrophotometry,  
and .

Ag(I) with  forms complexes of various stoichiometries, which is 
seen by every used method during direct titration of Ag(I) into 
both apo-zinc hook and Zn(II)-saturated zinc hook. The  shows 
that Ag(I) binding to the   zinc hook induces oligomerization of 
the zinc hook. The  studies confirm that Ag(I) replaces Zn(II) 
readily (within seconds), and the -titrations are in great 
accordance with observed 1:1 and 2:1 Ag(I):zinc hook 
stoichiometries observed in  and - or -titration. Surprisingly , 
contrary to the -, -, and both spectrophotometric titrations, 
does not discriminate many multiple complexes observed during the 
rest of the titrations. This may be because the heat or formation 
constants of the individual complexes are too small to be visible 
during . Titration of Ag(I) into the pf zinc hook shows two 
equivalence points, the first one at a ratio of 0.8 Ag(I):zinc 
hook, the second one at a ratio ~2.1. The reverse titration (zinc 
hook into Ag(I)) is in agreement with the ratios resulting from 
the normal titration. Though the fitting of the multiple-binding 
model to the isotherm was possible, the obtained formation 
constants should not be treated as reliable—firstly, the 
cumulative constant  of the Zn()₂ is far out of the measurable 
range of ,Kochańczyk et al. 2016 secondly, Ag(I) displaces Zn(II) 
easily, which means that the Ag(I)-zinc hook formation constant 
is much higher than the formation constant of Zn()₂. The titration 
of the Ag(I) into the Zn()₂ results in a single-step process, in 
which the  is obtainable. Assuming the correctness of  it is 
possible to estimate (not determine) the affinity of the Ag(I)- 
complexes. The estimated affinity of the Ag(I)- complexes 
probably refers to the complex Ag₄₂, as the equivalence point for 
the titration is around two, and the rest of the experiments 
suggest that the 4:2 Ag(I): complex is the most stable, however, 
the  might not be a sufficiently resolvable method to pinpoint 
the exactly formed complex. The estimated stability of the Ag(I) 
and  species is around 5.4 orders of magnitude higher than the 
stability of the Zn()₂ complex.

The study shows that Ag(I) ions may displace Zn(II) not only from 
the intraprotein zinc-binding sites (e.g. in ), but from 
interprotein zinc-binding sites as well. The swap most likely 
will result in the  disruption, and even though the architecture 
of the complex might be saved, the changes in the metal binding 
site conformation may lead to the loss of the function and huge 
biological impact. In case of the , as previously established, 
the deleterious mutations at the CXXC motif of cysteines to 
alanine and arginine is just as severe as the  mutation He et al. 2012
. The fact that the Zn(II)-induced dimerization is essential for 
proper functioning of , thus  as well, makes even small small 
disturbance in the zinc-binding site a suspect for a 
genotoxicity. 













Summary




Summary

When starting my doctorate, even though I had a plan of research 
and experiments written out and arranged in my head, I could not 
expect the final form my doctorate would take. I expected my 
doctorate to combine different fields of science, but I could not 
expect my doctorate to be, so interdisciplinary. I could not 
expect, the pandemic that started in 2019, and that in 2020 it 
would lock me in complete isolation for two weeks. I tried to 
make the most of the time in quarantine by conducting in silico 
research, despite the ex silico research that had already begun. 

I think the great advantage of presenting one's PhD as a set of 
publications linked by a common theme is that one tends to 
present things that more or less came out well, things we are 
satisfied with—we do not have to go back to things that did not 
come out, didn't have a chance to come out, things that 
demotivated us on a daily basis and kept us up at night. Of 
course, these things also happened during my doctorate, and I 
think it's important for other, younger, researchers to realize 
that a lot of the lines of research I undertook did not result in 
pushing the state of science forward, or an article—what was 
collected and presented in this thesis is the essence of hard 
work, but also luck.

This PhD has been the investigation of properties of , 
particularly . I do believe that contributions in the form of the 
articles presented here are of importance at a basic research 
level, but they also provide a foundation on which to build 
applied research. Understanding the assemblies formed by  may 
eventually lead to the creation of a model that could predict the 
interprotein metal binding from the structure or sequence of a 
protein. Correspondingly, filling the missing pieces of the 
metal–protein interaction in the protein–protein interaction 
puzzle can offer insight into the metal-dependent regulation of 
the interaction, as well as help to identify novel interactions 
that were missing from the picture due to the metal dependence. 
From a medical perspective, understanding  can give insight into 
diseases associated with abnormal metal-induced aggregation 
(e.g., crystallins in the eye) Ramirez-Bello et al. 2022, and how 
mismetallation of those sites may cause various diseases.

I am glad that all the publications presented here are available 
in open access, this allows other researchers to freely read and 
use the information meticulously prepared by us. This is truly 
important in the case of Galvanization of Protein--Protein Interactions in a Dynamic Zinc Interactome
 where we present the curated data set of the physiological . The 
ex silico part of my doctorate focused on the  protein, as the 
data presenting the   properties are still unpublished as of 
March 6, 2023, the Relations between Structure and Zn(II) Binding Affinity Shed Light on the Mechanisms of Rad50 Hook Domain Functioning and Its …
 makes a great foundation for the discussion of the properties of 
the  protein. The study lays out the basis of the long-range 
allostery between the zinc hook and the rest of the protein. To 
my knowledge, the described stabilization of the Zn()₂ is the 
first example of complex stabilization over such a long distance. 
The description of the link between the  damage response with the 
biology of Zn(II) ions and signal transduction pathways can be a 
great step in the apprehension of the metal dynamics in cancer 
cells.

To last two articles presented in the thesis show how detrimental 
is for the zinc-binding sites the mismetallation by heavy metal 
ions. Those articles, together with Metal Exchange in the Interprotein ZnII-Binding Site of the Rad50 Hook Domain: Struct…
 Padjasek et al. 2020 are the first articles addressing the issue 
of mismetallation of interprotein metal binding sites. Our 
studies provide an explanation of a potential genotoxic mechanism 
of mismetallation of the zinc hook.

Though the article describing InterMetalDB was published in a 
journal with a lower  the article describing physiological , Kocyła, Tran, and Krężel 2021
 I think that this is a much more fundamental contribution in 
this field. Most of the previous surveys regarding the existence 
of metal ions in  structures gather the data as a spreadsheet 
that is usually available after an exchange of emails, the same 
goes for the algorithms that were used to select the data. Even 
when the data is publicly available, the data is created only 
once and is no longer updated (as in the case of Galvanization of Protein--Protein Interactions in a Dynamic Zinc Interactome
, and similar studies). InterMetalDB is different—data presented 
in InterMetalDB is available after a few mouse clicks via the web 
interface, moreover, the resource is constantly updated, without 
any human involvement. The resource will be updated as long as 
the service is maintained, unfortunately, nobody, especially in 
the academy, can guarantee that the service will be maintained 
forever, where the service does not generate any revenue. Not 
surprisingly, even the services that were published last year are 
no longer available, and the URLs point to “404 Not Found”. With 
that thought in mind, I have envisioned InterMetalDB to be an 
open  project—freely available and reproducible. Any function, 
any component of InterMetalDB can be used by anyone, researcher 
or not, one only has to click “copy”, or “fork” repository on 
GitHub, instead of wasting time to recreate the whole service 
from the source article.

I believe that conducting research in line with open science 
principles (inter alia, providing free access to publications, 
data, software, and ) is the next step in the evolution of 
scientific knowledge dissemination in the age of the Internet, as 
it was with 17th-century advent of scientific journal during the 
Scientific Revolution. Already some principles of open science 
make the publication presented in the thesis available to the 
reader without the need to pay for the information, InterMetalDB 
and its resources are pushed the furthest in accordance with the 
principles of open science, which makes the InterMetalDB easily 
reproducible even after the ceasing of hosting.



Appendix





Bibliography
















  Colophon

This document was typeset using classicthesis style developed by 
André Miede. The style was inspired by Robert Bringhurst's 
seminal book on typography “The Elements of Typographic Style”. 
It is available for LaTeX and LyX at


https://bitbucket.org/amiede/classicthesis/

Some parts of the manuscript utilize  for typesetting. To do so I 
have used Python by Geoffrey Poore . To use Python in LyX I have 
made a Python-based wrapper available at


https://github.com/jzftran/pythontex-wrapper.py

To retrieve bibliometric data about publications I have used 
scholarly .







[note:
Hermann Zapf's Palatino and Euler type faces (Type 1 PostScript 
fonts URW Palladio L and FPL) are used. The “typewriter” text is 
typeset in Bera Mono, originally developed by Bitstream, Inc. as “
Bitstream Vera”. (Type 1 PostScript fonts were made available by 
Malte Rosenau and Ulrich Dirr.)
]

[note:
The custom size of the textblock was calculated using the 
directions given by Mr. Bringhurst (pages 26–29 and 175/176). 10 
pt Palatino needs 133.21 pt for the string 
``abcdefghijklmnopqrstuvwxyz''. This yields a good line length 
between 24–26 pc (288–312 pt). Using a “double square textblock” 
with a 1:2 ratio this results in a textblock of 312:624 pt (which 
includes the headline in this design). A good alternative would 
be the “golden section textblock” with a ratio of 1:1.62, here 
312:505.44 pt. For comparison, DIV9 of the typearea package 
results in a line length of 389 pt (32.4 pc), which is by far too 
long. However, this information will only be of interest for 
hardcore pseudo-typographers like me.

To make your own calculations, use the following commands and 
look up the corresponding lengths in the book: \settowidth{\abcd}{abcdefghijklmnopqrstuvwxyz}

\the\abcd\ % prints the value of the length
 Please see the file classicthesis.sty for some precalculated 
values for Palatino and Minion.
]






 



Declaration

Put your declaration here.



, 






     
-----
     




 

Rights and Permissions

Articles that build fragments of this thesis are open access 
articles distributed under the terms of the Creative Commons 
license, which permits unrestricted use, distribution and 
reproduction, provided the proper citation of the original work. 
Thus before each article I present the citation of the original 
article.