\makeatletter
\@ifundefined{HCode}
{\documentclass[screen,CRCHIM,Unicode,biblatex]{cedram}
\addbibresource{crchim20260018.bib}
\newenvironment{noXML}{}{}
\def\tsup#1{$^{{#1}}$}
\def\tsub#1{$_{{#1}}$}
\def\ndash{\text{--}}
\usepackage[T1]{fontenc}
\def\thead{\noalign{\relax}\hline}
\def\tbody{\noalign{\relax}\hline}
\def\endthead{\noalign{\relax}\hline}
\def\tabnote#1{\vskip4pt\parbox{.96\linewidth}{#1}}
\def\hyphen{\text{-}}
\RequirePackage{etoolbox}
\def\jobid{crchim20260018}
%\graphicspath{{/tmp/\jobid_figs/web/}}
\graphicspath{{./figures/}}
\newcounter{runlevel}
\let\MakeYrStrItalic\relax
\csdef{Seqnsplit}{\\}
\def\refinput#1{}
\def\back#1{}
\def\botline{\\\hline}
\def\dollar{\$}
\def\ubreak{\break}
\def\og{\guillemotleft}
\def\fg{\guillemotright}
\def\mn{\phantom{$-$}}
\def\0{\phantom{0}}
\def\figType#1{}
\usepackage{multirow} 
\def\nrow#1{\@tempcnta #1\relax%
\advance\@tempcnta by 1\relax%
\xdef\lenrow{\the\@tempcnta}}
\def\morerows#1#2{\nrow{#1}\multirow{\lenrow}{*}{#2}}
\usepackage{hyperref}
\makeatletter
\g@addto@macro{\UrlBreaks}{\UrlOrds}
\gappto{\UrlBreaks}{\UrlOrds}
\def\rmmu{\upmu}
\DOI{10.5802/crchim.445}
\datereceived{2026-01-08}
\daterevised{2026-02-09}
\dateaccepted{2026-02-04}
\ItHasTeXPublished
}
{\documentclass[crchim]{article}
\usepackage[T1]{fontenc}
\def\CDRdoi{10.5802/crchim.445}
\let\refinput\input
\let\ubreak\relax
\makeatletter
\def\CDRsupplementaryTwotypes#1#2{}
\let\splittabular\relax
\newcommand\@coi{}
\newcommand\COI[1]{\gdef\@coi{#1}}
\newcommand\printCOI{\ifx\@coi\@empty\else%
\section*{Declaration of interests}
\@coi\fi
}
\let\newline\break
}
\makeatother

\usepackage{upgreek}

\COI{The authors do not work for, advise, own shares in, or receive
funds from any organization that could benefit from this article, and
have declared no affiliations other than their research organizations.}

\begin{document}

%\dateposted{2026-02-16}

\begin{noXML}

\editornote{Article submitted by invitation.}
\alteditornote{Article soumis sur invitation.}

\CDRsetmeta{articletype}{review}

\title{Properties of phosphorus dendrimers decorated with heterocycles}

\alttitle{Propri\'{e}t\'{e}s de dendrim\`{e}res phosphor\'{e}s
d\'{e}cor\'{e}s par des h\'{e}t\'{e}rocycles}

\author{\firstname{Zoia} \lastname{Voitenko}\CDRorcid{0000-0002-2408-6991}}
\address{Univ Toulouse, CNRS, LCC, Toulouse, France}
\address{Laboratoire de Chimie de Coordination, 
205 Route de Narbonne, 31077 Toulouse CEDEX 4, France}
\email[Z. Voitenko]{zoia.voitenko@lcc-toulouse.fr}

\author{\firstname{Val\'{e}rie} \lastname{Maraval}\CDRorcid{0000-0003-0900-7332}}
\addressSameAs{1}{Univ Toulouse, CNRS, LCC, Toulouse, France}
\addressSameAs{2}{Laboratoire de Chimie de Coordination, 
205 Route de Narbonne, 31077 Toulouse CEDEX 4, France}
\email[V. Maraval]{valerie.maraval@lcc-toulouse.fr}

\author{\firstname{Anne-Marie} \lastname{Caminade}\CDRorcid{0000-0001-8487-3578}\IsCorresp}
\addressSameAs{1}{Univ Toulouse, CNRS, LCC, Toulouse, France}
\addressSameAs{2}{Laboratoire de Chimie de Coordination, 
205 Route de Narbonne, 31077 Toulouse CEDEX 4, France}
\email[A.-M. Caminade]{anne-marie.caminade@lcc-toulouse.fr}

\shortrunauthors

\keywords{\kwd{Dendrimer}
\kwd{Heterocycle}
\kwd{Phosphorus}
\kwd{Catalysis}
\kwd{Materials}
\kwd{Biology}}

\altkeywords{\kwd{Dendrim\`{e}re}
\kwd{H\'{e}t\'{e}rocycle}
\kwd{Phosphore}
\kwd{Catalyse}
\kwd{Mat\'{e}riaux}
\kwd{Biologie}}

\begin{abstract}
Phosphorus dendrimers are highly branched, monodisperse macromolecules
built around a phosphorus core, and having phosphorus atoms at all
branching points. Their modular architecture allows for precise control
over size, shape, and surface functionality. When functionalized at the
surface with heterocycles---organic rings containing one or more
heteroatoms (e.g., nitrogen, oxygen, sulfur)---these dendrimers gain
catalytic activity, in the materials or biological fields. This review
is organized according to the type of heterocycles (including N, O, S,
or combined heteroatoms) linked to the surface of phosphorus dendrimers
and emphasizes their properties.
\end{abstract}

\begin{altabstract}
Les dendrim\`{e}res phosphor\'{e}s sont des macromol\'{e}cules
monodisperses, fortement ramifi\'{e}es, construites autour d'un noyau
phosphor\'{e} et comportant un atome de phosphore \`{a} tous les points
de ramification. Leur architecture modulaire permet un contr\^{o}le
pr\'{e}cis de la taille, de la forme et de la fonctionnalit\'{e} de
surface. Lorsqu'ils sont fonctionnalis\'{e}s en surface avec des
h\'{e}t\'{e}rocycles --- contenant un ou plusieurs h\'{e}t\'{e}roatomes (par
exemple, azote, oxyg\`{e}ne, soufre) --- ces dendrim\`{e}res
acqui\`{e}rent une activit\'{e} accrue en tant que catalyseurs, pour
les mat\'{e}riaux ou pour la biologie. Cette revue est organis\'{e}e en
fonction du type d'h\'{e}t\'{e}rocycles (N, O, S, ou leur combinaison)
li\'{e}s \`{a} la surface des dendrim\`{e}res phosphor\'{e}s et mettra
l'accent sur leurs propri\'{e}t\'{e}s.
\end{altabstract}

%\input{CR-pagedemetas}

\maketitle

\vspace*{3pt}

\twocolumngrid

\end{noXML}

\section{Introduction}\label{sec1}
Heterocycles of medium size (five- or six-membered rings most
generally) contain at least one heteroatom, most generally nitrogen,
oxygen, or sulfur atoms, which are gaining widespread interest due to
their numerous properties, particularly for medicinal 
uses~\cite{1,2,3,4}. Among the different types of heterocycles,
nitrogen heterocycles are the most widely represented both in Nature
and in medicinal  compounds~\cite{5,6}. Recent reviews have emphasized
the role of nitrogen heterocycles in U.S.\ FDA-approved pharmaceuticals 
(2013--2023)~\cite{7}, in particular as anticancer drugs~\cite{8,9,10}
and anti-viral  agents~\cite{11}. More specific reviews have gathered
the properties of  pyridine~\cite{12,13}, piperidine~\cite{14},
quinoline~\cite{15}, pyrrole~\cite{16}, pyrrolidine~\cite{17}, and
triazole~\cite{18,19} derivatives. A few other elements have been used
for the synthesis of heterocycles, in particular oxygen, as emphasized
in a review about recent FDA-approved drugs containing oxygen
heterocycles (including  carbohydrates)~\cite{20}. Sulfur heterocycles
are less common, but also display medicinal properties, as they make up
an important part of FDA-approved  drugs~\cite{21}. Despite the very
large number of heterocycles, their synthesis is still of current
interest, in particular for their  functionalization~\cite{22}.

\begin{figure*}
\includegraphics{fig01}
\caption{\label{fig1}Full structures and corresponding abbreviated
linear structures of different generations of PPH dendrimers built from
two different cores.}
\vspace*{-4pt}
\end{figure*}

Dendrimers are highly branched, monodisperse macromolecules,
synthesized stepwise, generally by a divergent process from a
multifunctional  core~\cite{23}. Each time the number of terminal
functions is multiplied, most frequently by  two~\cite{24}, a new
generation is created.   It should also be kept in mind that their
synthesis must be efficient and  selective to obtain high generations.
Among the different types of  dendrimers~\cite{25}, those based on
\mbox{``inorganic''}  elements~\cite{26}, in particular phosphorus, i.e.,
\mbox{having} phosphorus atoms both at the core and at the branching 
points~\cite{27,28}, display specific properties that have been
recently  emphasized~\cite{29}. Their easy characterization by 
$\mbox{}^{31}$P NMR has to be underlined in  particular~\cite{30}. In
most cases, phosphorus dendrimers were built from a trifunctional core 
(P(S)Cl$_{3}$) up to generation  12~\cite{31} or from a hexafunctional
core  (N$_3$P$_3$Cl$_6$) up to generation  8~\cite{32}. Both families
of phosphorus dendrimers will be considered in this review, but with
the emphasis of the most widely used, based on  N$_{3}$P$_3$Cl$_6$ as
core, which affords twice the number of terminal functions at a given
generation, compared to the P(S)Cl$_{3}$ core.  Figure~\ref{fig1}
displays both the full chemical structures and the abbreviated chemical
structures (linear with parentheses after each branching point) of
these dendrimers based on poly(phosphorhydrazone) (PPH) linkages. These
PPH dendrimers have, as terminal functions, either  P(S)Cl$_{2}$  or
aldehyde groups both being highly reactive and suitable for the
grafting of new chemical functions, depending on the desired 
properties. 

This review will display the peripheral functionalization of phosphorus
dendrimers with different types of heterocycles. It will be organized
depending on the element included in the heterocycle, beginning with
nitrogen, followed by oxygen, then by sulfur, and it will end with
heterocycles containing two types of heteroatoms, such as P and N, P
and O, N and O. Each part will begin with fundamental 
\mbox{research} followed
by applications, in catalysis, for materials, and mainly for biological
uses. 

\begin{scheme*}
\includegraphics{sc01}
\vspace*{4pt}
\caption{\label{sch1}First functionalization examples of
aldehyde-decorated PPH dendrimers by nitrogen heterocycles. 
(A)~Condensation.  (B)~Other condensation followed by
protonation. (C)~Horner--Wadsworth--Emmons reaction.}
\vspace*{-14pt}
\end{scheme*}

\section{PPH dendrimers functionalized with\newline nitrogen
heterocycles}\label{secII}

\subsection{Methods of synthesis and characterization}\label{secII.1} 
The very first examples of PPH dendrimers functionalized with
heterocycles concerned nitrogen heterocycles, grafted on the aldehyde
terminal functions. Characterization was essentially carried out by
NMR. Simple condensations with various hydrazine derivatives were first
carried out on the fourth generation, bearing 48 aldehyde terminal 
functions~\cite{33}. One of them was
1-amino-4-(2-hydroxyethyl)piperazine, which was reacted with
generations 1 to 4 of PPH dendrimers built from the trifunctional core,
affording dendrimers \textbf{1-Gn}  ($\mathbf{n} = 0$ to 4) 
(Scheme~\ref{sch1}A)~\cite{34}. Another example of condensation
concerned the reaction of a third-generation PPH dendrimer with
2-hydrazinopyridine, affording dendrimer \textbf{2-G3}. To obtain a
water-soluble dendrimer, the pyridines were protonated with HCl 
(Scheme~\ref{sch1}B). The thermal 
\mbox{behavior} of both compounds and of
many other differently functionalized PPH dendrimers was studied. The
percentage of mass remaining at ${+}$1000~\textdegree C was 33.5\% and
23.5\% for the neutral (\textbf{2-G3}) and protonated
(\textbf{2-G3,HCl}) dendrimers, respectively. These values are lower
than those obtained for the dendrimers with aldehyde terminal
functions (47.0\% for \textbf{G3-CHO}$_{\mathbf{48}}$)~\cite{35}.
Generation 1 PPH dendrimers were used in Horner--Wadsworth--Emmons 
reactions~\cite{36} with different stabilized phosphonate carbanions.
The same reaction was also applied to generation~4, in particular with
the phosphonate functionalized with a piperidine  (Scheme~\ref{sch1}C).
Such a reaction afforded both  $E$- and  $Z$-isomers of the C${=}$C
double bond, in an $E/Z$ ratio of 90:10 for both the first
(\textbf{3-G1}) and fourth (\textbf{3-G4})  generations~\cite{37}. 

The P(S)Cl$_{2}$ terminal functions of PPH dendrimers are also suitable
for grafting heterocycles, using either phenol- or amine-functionalized
heterocycles. The reaction of Boc-protected 4-(hydroxyphenyl)
piperazine, in the presence of cesium carbonate as a base, afforded
dendrimer \textbf{4-G2}  (Scheme~\ref{sch2}A). Deprotection was carried
out with trifluoroacetic acid (TFA), then further functionalization was
carried out through a Michael addition of the deprotected piperazine to
vinylidene tetraisopropyl bisphosphonate to afford dendrimer
\textbf{5-G2}. The same process was applied to generations 1 and 3,
affording dendrimers \textbf{5-G1} and \textbf{5-G3}, respectively.
These dendrimers were used for complexing gadolinium from
$\mathrm{Gd}(\mathrm{SO}_{3})_{3}{\cdot}6\mathrm{H}_{2}\mathrm{O}$. 
Measurements of magnetic susceptibility for the three generations
confirmed the presence of one gadolinium per bisphosphonate group and
the occurrence of antiferromagnetic Gd--Gd  interactions~\cite{38}. 

\begin{scheme*}
\includegraphics{sc02}
\vspace*{4pt}
\caption{\label{sch2}Various examples of functionalization of PPH
dendrimers P(S)Cl$_{2}$ peripheral groups by nitrogen heterocycles.
(A)~Reaction with phenols.  (B)~Reaction with amines.
(C)~Reaction with amines on a non-symmetrical dendrimer. 
(D)~Two-step functionalization with amines.}
\vspace*{-10pt}
\end{scheme*}

The first example of heterocycles grafted to PPH dendrimers through an
amine concerned  $N$-(trimethylsilyl)imidazole  (Scheme~\ref{sch2}B).
The reaction was applied to generations 1 and 4 affording 
\mbox{dendrimers}
\textbf{6-G1} and \textbf{6-G4},  respectively~\cite{34}. A kind of
Janus~\cite{39,40} dendrimer
\textbf{G1-Cl}$_{\mathbf{2}}$\textbf{-G2-Cl}$_{\mathbf{8}}$, 
reacted on both sides with 1-(2-aminoethyl)piperidine,  in the presence
of diisopropylethylamine (DIPEA) as a base,  to afford compound
\textbf{7-G1-G2} functionalized with a  piperidine on both sides 
(Scheme~\ref{sch2}C). The same reaction was applied to the next
generation, affording dendrimer \textbf{7-G2-G3}. In both cases,
protonation with HCl led to water-soluble  compounds~\cite{41}. In a
third example, the heterocycle was not grafted directly on the
P(S)Cl$_{2}$ groups but in a second step. The first step consisted in
the substitution of the chlorines with
$S$-methyl-4-hydroxydithiobenzoate, affording dendrimer
\textbf{G1-thio}$_{\mathbf{12}}$ (Scheme~\ref{sch2}D). The second
step was a thioacylation with various amines, in particular piperazine,
to afford dendrimer  \textbf{8-G1}~\cite{42}. 

\begin{scheme*}
\includegraphics{sc03}
\vspace*{4pt}
\caption{\label{sch3}PPH dendrimers functionalized with TEMPO radicals.}
\vspace*{-9pt}
\end{scheme*}

Some dendrimers were functionalized to allow specific
characterizations. For instance, generation zero
\textbf{G0-CHO}$_{\mathbf{6}}$ was condensed with the 4-amino-TEMPO
(2,2,6,6-tetramethylpiperidine \mbox{1-oxyl}) radical, to afford compound
\textbf{9-G0}  (Scheme~\ref{sch3}). Completion of the reaction was
monitored by  $\mbox{}^{31}$P NMR, which was not too much affected by
the presence of the radicals. This compound was also characterized by
single-crystal X-ray diffraction and electron paramagnetic resonance
(EPR). Both techniques gave identical structures in the solid state and
in solution, with three arms above the cyclotriphosphazene ring, and
three  below~\cite{43}. The same synthetic method was applied to
generations 1 to 4 of PPH dendrimers. Proportionality between the EPR
signal intensity and the number of radicals of each generation was
observed. These dendrimers were studied in diluted solution to detect
intramolecular interactions between branches. A  $\vert \Delta
m_S\vert = 2$  transition at half-field was observed in all cases by
EPR, giving direct evidence of the intramolecular origin of the dipolar
interactions. The magnetic properties of dendrimers \textbf{9-G0},
\textbf{9-G1}, and \textbf{9-G4} were also investigated on a
polycrystalline sample by superconducting quantum interference device
(SQUID) magnetometry. The results indicated antiferromagnetic
interactions between the  radicals~\cite{44}.

Another specific analytical method is fluorescence, used for the
characterization of dendrimers functionalized with fluorophores,
especially those based on heterocycles.
$N$-(4$'$-hydroxyphenylethyl)-3,4-diphenylmaleimide, obtained by
reaction of tyramine with 3,4-diphenyl maleic anhydride, was reacted
with the P(S)Cl$_{2}$ terminal functions of PPH dendrimers from
generation 1 to generation 3, to afford dendrimers \textbf{10-Gn} 
($\mathbf{n} = 1$, 2, and 3)  (Figure~\ref{fig2}A). A linear increase
of the $\lambda_{\mathrm{max}}$ values in UV--Vis spectroscopy with the number of
chromophores confirmed the absence of large defects in the structure of
the dendrimers. The fluorescence of all these dendrimers was measured
in THF at 300 and 375 nm as wavelengths for excitation, with emission
at 499 nm. The fluorescence quantum yields decreased from generation 0
(78\%) to generation 3 (39\%). Analogous experiments were carried out
in CH$_{2}$Cl$_{2}$ at 302 or 320 and 375 nm as the wavelengths for
excitation, with emission at 506~nm. A decrease in the quantum yield
was also observed, from 72\% for \textbf{10-G1} to 23\% for
\textbf{10-G3}. Such a decrease might be due to interactions between
fluorophores in close proximity, or to interactions of the fluorophores
with the dendritic structure, leading in both cases to a nonradiative
deactivation  process~\cite{45}. Rhodamine B was functionalized with
tyramine to be grafted to the first-generation PPH dendrimer having
P(S)Cl$_{2}$ terminal functions, to afford compound \textbf{11-G1},
having both a nitrogen and an oxygen heterocycle, associated in a
spirolactam structure  (Figure~\ref{fig2}B). Such a closed form is not
fluorescent, and it was not possible to open it to recover
fluorescence, even when adding  HCl~\cite{46}. 

\begin{figure*}
\includegraphics{fig02}
\vspace*{2pt}
\caption{\label{fig2}Fluorescent PPH dendrimers based on heterocycles.
(A)~3,4-Diphenylmaleimide. (B)~Spirolactam form of
rhodamine B. (C)~Stilbazole chromophores for TPA.}
\vspace*{4pt}
\end{figure*}

Two-photon absorption (TPA) processes are involved in a wide range of
applications, and different types of dendrimers have been
functionalized with TPA  fluorophores~\cite{47}, in particular PPH
dendrimers~\cite{48}. A series of compounds functionalized with
push--pull stilbazole chromophores, including dendrimers \textbf{12-G1}
and \textbf{12-G2}, were synthesized, to study the magnitude of the
potential cooperative effects between fluorophores 
(Figure~\ref{fig2}C). A reduction in the molar extinction coefficient
per chromophoric subunit was observed in dendrimers compared to the
monomer. The same molar extinction per chromophoric subunit was also
observed when increasing the generation of the dendrimers, probably due
to an increased proximity between the chromophores. Interestingly, the
TPA maximum cross  section $\sigma_{2}^{\mathrm{max}}$ per chromophoric
subunit increased from monomer to dendrimers on the 
contrary~\cite{49}.

\subsection{Catalytic properties of PPH dendrimers functionalized with
N-heterocycles}\label{secII.2}
PPH dendrimers have many catalytic properties, as shown in a recent 
review~\cite{50}. Most of them bear nitrogen heterocycles as ligands or
organocatalysts, as it will be shown in this part of the review. A
series of dendrimers (generations 1 to 3) and the corresponding monomer
functionalized with the pyridine-imine ligand were used for the
complexation of copper, affording dendrimers \textbf{13-G1},
\textbf{13-G2}, and \textbf{13-G3}, and monomer \textbf{13-M} 
(Scheme~\ref{sch4}A). These compounds were used as catalysts in several
reactions. It should be noted that in all catalytic experiments
performed with different generations of PPH dendrimers, the same
quantity in catalytic sites is used. It means that the efficiency of
one equivalent of \textbf{13-G3} is compared with that of two
equivalents of \textbf{13-G2}, four equivalents of \textbf{13-G1}, and
48 equivalents of \textbf{13-M}. These catalysts were used in different
reactions, in particular for the arylation of pyrazole with iodobenzene
and bromobenzene  (Scheme~\ref{sch4}B). A large difference in catalytic
efficiency was observed between the monomer (very low efficiency) and
the dendrimers. A difference between the different generations of the
dendrimers was also observed in the case of bromobenzene 
(Scheme~\ref{sch4}C), displaying a nice positive dendrimer (or
dendritic)  effect~\cite{51}. Other catalyzed reactions with these
dendritic complexes comprised the coupling of 3,5-dimethylphenol with
iodobenzene  (Scheme~\ref{sch4}D) and the coupling of bromostyrene with
N- and O-nucleophiles, pyrazole and 3,5-dimethylphenol, respectively 
(Scheme~\ref{sch4}E)~\cite{52}.

\begin{scheme*}
\includegraphics{sc04}
\vspace*{5pt}
\caption{\label{sch4}Catalysis with copper complexes of dendritic
pyridine-imine ligands.}
\vspace*{-6pt}
\end{scheme*}

A terpyridine complexing scandium at the surface of a fourth-generation
PPH dendrimer (compound \textbf{14-G4} in  Scheme~\ref{sch5}) was used
as catalyst in Friedel--Crafts acylations under microwave heating. A
wide range of aromatics were used, in twelve consecutive runs, by
recovering the dendritic catalyst through precipitation with diethyl
ether and reusing it in the next run with different substrates. It can
be noted that runs 4 and 12 were carried out with the same substrates
and afforded identical  yields~\cite{53}.

\begin{scheme*}
\includegraphics{sc05}
\vspace*{5pt}
\caption{\label{sch5}Generation 4 of dendritic terpyridine--scandium
complex for catalyzing Friedel--Crafts acylations under microwaves
(MW), using a wide range of aromatics, by recovering and reusing the
dendritic catalyst eleven times.}
\vspace*{5pt}
\end{scheme*}

A tripodal C-scorpionate  ligand [tris-2,2,2-(1-pyrazolyl)ethanol], was
used both as monomer (\textbf{15-M}) and grafted to a first-generation
PPH dendrimer (\textbf{15-G1}) for complexing Pd(OAc)$_{2}$. These
complexes were used for catalyzing a Sonogashira reaction between
phenylacetylene and iodobenzene, and a Heck reaction between styrene
and iodobenzene  (Scheme~\ref{sch6}). Monomer and dendrimer gave almost
the same yield in coupling product in the case of the Sonogashira
reaction, but the dendrimer was much better (63\%) than the monomer
(24\%) in Heck  couplings~\cite{54}.

\begin{scheme*}
\includegraphics{sc06}
\vspace*{5pt}
\caption{\label{sch6}Pd-complexing monomeric and dendritic
C-scorpionate ligands used as catalysts in Sonogashira and Heck
couplings.}
\vspace*{-10pt}
\end{scheme*}

Cheaper and safer than organometallic catalysis, organocatalysis is
increasingly used. A review has emphasized the early times of dendritic
organocatalysts~\cite{55}. The J\o{}rgensen--Hayashi catalyst
$((S)$-$\upalpha$,$\upalpha$-diphenylprolinol trimethylsilyl
ether~\cite{56,57}) was grafted to the surface of generations 1 to 3 of
PPH dendrimers (compounds \textbf{16-G1}, \textbf{16-G2} and
\textbf{16-G3}), and to cobalt nanoparticles covered by a few graphene
layers (\textbf{16-NP},  Scheme~\ref{sch7}). Both types of compounds
were applied as organocatalysts (10~mol\% of catalytic sites in all
cases) in the addition of propanal to $\upbeta$-nitrostyrene at 
10~\textdegree C. Dendrimers \textbf{16-G2} and \textbf{16-G3}, and the
nanoparticles \textbf{16-NP} gave excellent yields at the first run.
However, when trying to recover the catalysts, only dendrimer
\textbf{16-G3} still gave excellent results after four runs (graph in 
Scheme~\ref{sch7}). Compounds \textbf{16-G3} and \textbf{16-NP} were
then used as organocatalysts for the coupling of a series of
nitroolefins with a series of aldehydes. Dendrimer \textbf{16-G3} gave
yields over 99\% in all cases, whereas nanoparticles \textbf{16-NP}
gave yields between 42\% and  99\%~\cite{58}.

\begin{scheme*}
\includegraphics{sc07}
\vspace*{5pt}
\caption{\label{sch7}Organocatalysis with dendrimers and nanoparticles
using 10~mol\% of prolinol moieties in all cases, and recovery and
reuse experiments.}
\vspace*{-10pt}
\end{scheme*}

Two other examples of organocatalysis involved $(+)$-cinchonine,
differently grafted to PPH 
\mbox{dendrimers.} In the first case, the ethylenic
function of $(+)$-cinchonine was involved in a thiol-ene reaction with
4-(2-mercaptoethyl) phenol, affording monomer \textbf{17-M}, which was
grafted to a branch, to afford \textbf{17-B} and to generations 1 and 4
of PPH dendrimers, to afford compounds \textbf{17-G1} and
\textbf{17-G4}, respectively. All these compounds were applied as
catalysts in asymmetric amination of $\upbeta$-keto esters, using in
particular ethyl 2-oxocyclopentanecarboxylate and benzyl
azodicarboxylate  (Scheme~\ref{sch8}). The reaction was extremely rapid
(2 min) except with the fourth-generation \textbf{17-G4}, which
necessitated 30 min. The branch \textbf{17-B} and the first-generation
\textbf{17-G1} gave the best enantiomeric excesses, whereas
\textbf{17-G4} afforded a racemic mixture. Recycling experiments were
carried out with \textbf{17-B} and \textbf{17-G1}. The branch
\textbf{17-B} could be recovered and reused five times, whereas the
dendrimer \textbf{17-G1} could be recovered at least nine times,
displaying still very good yield (90\%) and enantioselectivity (82\%)
at run 10. The scope of this reaction was evaluated using cyclic and
noncyclic esters, cyclic $\upbeta$-diketones, and opened-chain esters.
In practically all cases, the dendrimer \textbf{17-G1} was more
efficient than the branch  \textbf{17-B}~\cite{59}.

\begin{scheme*}
\vspace*{5pt}
\includegraphics{sc08}
\caption{\label{sch8}Organocatalysis with $(+)$-cinchonine
derivatives.}
\vspace*{-12pt}
\end{scheme*}

The second method for grafting various derivatives of $(+)$-cinchonine
necessitated first to modify the surface of the first-generation PPH
dendrimer, to have iodine as terminal functions. In the last step, the
dendrimer was used to alkylate $(+)$-cinchonine, affording dendrimers
\textbf{18-G1a}--\textbf{d}, functionalized with twelve quaternary
ammonium salts, and different types of R substituents (H, allyl,
benzyl, TMS). This family of dendrimers was used as organocatalyst
(0.1~mol\%) in the reaction of glycinate with benzyl bromide at 25, 0,
and  ${-}$25~\textdegree C (Scheme~\ref{sch9}). The resulting imines were
derivatized to the corresponding trifluoroacetamides, and deprotected
with trifluoroacetic anhydride to get stable compounds, easier to
analyze. Dendrimer \textbf{18-G1b}  (R ${=}$ allyl) afforded the best
enantioselectivities, and the enantiomeric excess (ee) was better when
the reactions were performed at 0~\textdegree C. It was possible to
recover and reuse \textbf{18-G1b} at least four times with still good
yield (77\%) and enantioselectivity  (79\%)~\cite{60}.

\begin{scheme*}
\includegraphics{sc09}
\vspace*{4pt}
\caption{\label{sch9}First generation of PPH dendrimer functionalized
with twelve quaternary ammonium salts and used as organocatalyst.}
\vspace*{-10pt}
\end{scheme*}

\subsection{Materials functionalized with PPH\newline 
dendrimers bearing
N-heterocycles}\label{secII.3}   
Condensation between the aldehyde terminal functions of PPH dendrimers
and hydrazones of type Girard P (pyridinium) afforded dendrimers
functionalized with pyridiniums, either built from a trifunctional core
(family \textbf{19-Gn}) or an hexafunctional core (family
\textbf{20-Gn}). All these dendrimers were soluble in water, but when
heated at 60--65~\textdegree C for 11--13~days, the solutions became
rigid hydrogels, as illustrated by the reversed flask in 
Figure~\ref{fig3}. Each terminal pyridinium group was able to gel
around 1200--1400 molecules of water, 
{regardless} of the dendrimer {generation.} 
Gelation time can be dramatically reduced in the presence
of water-soluble substances such as acids (citric, ascorbic, lactic,
L-tartaric acid, etc.),  buffer [TRIS
(tris(hydroxymethyl)aminomethane)], dithioerythritol (DTE), sodium salt
of ethylenediamine tetraacetate (EDTA), or even metallic salts (Ni, Y,
Er acetates). Freeze-drying of these hydrogels at low temperature gave
rise to aerogels that retained the shape and size of the hydrogels,
presumably in relation to the supramolecular interactions shown in 
Figure~\ref{fig3}~\cite{61}. These dendrimers were also used for the
production of fibers when 
{depositing} them through a moving needle in a
flocculating bath containing 10 mol\% La(NO$_{3}$)$_{3}$. These fibers
displayed an elastic behavior (reversible deformation) contrarily to
fibers produced with polymers in the same condition, which displayed a
plastic behavior (irreversible deformation), as shown by mechanical 
{measurements~\cite{62}.}\looseness=-1

\begin{figure*}
\includegraphics{fig03}
\caption{\label{fig3}Hydrogel obtained with dendrimers functionalized
with the Girard P reagent.}
\end{figure*}

Later on, the hydrogels based on the \textbf{20-Gn} family were also
prepared in the presence of biocompatible additives such as glucose,
glycine, or polyethylene glycol. They were used to efficiently bind and
slowly release nucleic  acids~\cite{63}. 

\begin{figure*}
\includegraphics{fig04}
\vspace*{4pt}
\caption{\label{fig4}Dendrons functionalized with ethacrynic acid and
grafted to graphene oxide.}
\vspace*{6pt}
\end{figure*}

Graphene oxide is generally produced by oxidation of 
graphite~\cite{64}, inducing the formation of various oxygenated
functional groups, potentially suitable for grafting different
entities, including dendritic structures. Dendrons are dendritic
wedges, having one function at the core different from the functions at
the  surface~\cite{65}, which are in particular suitable for the
grafting to materials. Different types of PPH dendrons have been
grafted to graphene oxide, first modified to be suitable to react with
the function located at the core of dendrons. Modification of graphene
oxide with SOCl$_{2}$ produced acid chloride functions, suitable to
react with primary-amine functions in peptide-coupling-type reactions.
Different dendrons equipped with Boc-protected tyramine at the core and
various types of functions at the surface were deprotected, but the
deprotection was successful only in the case of ethacrynic acid
terminal functions. It was shown previously that small 
derivatives~\cite{66} and PPH dendrimers functionalized with ethacrynic
acid~\cite{67,68} displayed moderate anticancer properties. A
first-generation dendron functionalized with ethacrynic acid linked to
the dendrimer through 4-hydroxy-phenylpiperazine was grafted to
modified graphene oxide, affording \textbf{21-G1@GO} 
(Figure~\ref{fig4}). Dendron \textbf{21-G1} alone displayed a moderate
anticancer activity, which vanished when it was grafted to graphene 
\mbox{oxide~\cite{69}.} 

A series of dendrons having different types of pyridine-imine
peripheral functions, and either an  alkyne [series \textbf{22x-G1} 
$(\mathbf{x} = \mathrm{a}, \mathrm{b}, \mathrm{c})$] (Figure~\ref{fig5}A) or an azide [series
\textbf{23x-G1}  $(\mathbf{x} = \mathrm{a}, \mathrm{b}, \mathrm{c})$]  (Figure~\ref{fig5}B) at the
core, were grafted via ``click''  reactions~\cite{70} to graphene oxide
previously functionalized with azide or alkyne, respectively. Materials
\textbf{22x-G1@GO} and \textbf{23x-G1@GO} $(\mathbf{x} = \mathrm{a}, \mathrm{b}, \mathrm{c})$ were
obtained in this way, functionalized via a triazole ring for both
families  (Figure~\ref{fig5}). Only material \textbf{23a-G1@GO}
displayed moderate anticancer properties against HCT116 cells (human
colon cancer), with a percentage of viability of 60.5\% at 10$^{-5}$~M,
but this material was by far less efficient than the corresponding
dendron alone \textbf{23a-G1} (4.2\% viability at 
10$^{-5}$~M)~\cite{71}.

\begin{figure*}
\includegraphics{fig05}
\caption{\label{fig5}Graphene oxide (GO) functionalized with two series
of pyridine-imine dendrons, grafted by click chemistry.}
\end{figure*}

\subsection{Biological properties of PPH dendrimers functionalized with
N-heterocycles}\label{secII.4} 
PPH dendrimers functionalized with nitrogen heterocycles can have
either biological properties by themselves or can be used to bind and
deliver bioactive substances. The pyridine-imine functions shown in 
Figure~\ref{fig5} were also grafted to the surface of PPH dendrimers,
from generation 1 to generation 3, and they were used first for the
complexation of copper dichloride, affording dendrimers
\textbf{24-G1a,b,c-Cu}$_{\mathbf{12}}$,
\textbf{24-G2a,b,c-Cu}$_{\mathbf{24}}$, and
\textbf{24-G3a,b,c-Cu}$_{\mathbf{48}}$ (Figure~\ref{fig6}). Both
the ``free'' dendrimers and those complexing copper were tested for the
growth inhibition of HL60 cells (leukemia). The third generation was in
all cases the most efficient, thus further experiments were carried out
only with the third generation dendrimers against KB cells (epidermal
carcinoma). Dendrimers functionalized with ligands of type \textbf{a}
were more efficient than those functionalized with ligands of type
\textbf{b} or \textbf{c}, and the copper complexes were more 
\mbox{efficient}
than the free dendrimers. Thus, only dendrimers \textbf{24-G3a} and
\textbf{24-G3a-Cu}$_{\mathbf{48}}$ were tested against a panel of
cancerous and non-cancerous cells. The IC$_{50}$ values (quantity of
dendrimers necessary to kill 50\% of cells) were measured. The
Cu-complex \textbf{24-G3a-Cu}$_{\mathbf{48}}$ was more toxic
against the cancerous cells KB, HL60, HCT116 (human colon cancer), MCF-7
(hormone-responsive breast cancer), OVCAR8 (ovarian carcinoma), and U87
(human glioblastoma) than against the non-cancerous cells MCR-5
(proliferative human lung fibroblasts) and the quiescent EPC
(endothelial progenitor cells, \textit{Cyprinus carpio}), contrarily to
the free dendrimer  \textbf{24-G3a}~\cite{72}. In order to understand
the large difference observed between the different types of ligands
(\textbf{a}, \textbf{b}, or \textbf{c}), comparative electron
paramagnetic resonance (EPR) studies were 
\mbox{carried} out, in the presence
(or absence) of HCT116 cancer cells and MRC-5 normal cells. It was
shown that dendrimer \textbf{24-G3a-Cu}$_{\mathbf{48}}$ bound
copper more firmly than \textbf{24-G3b-Cu}$_{\mathbf{48}}$ and
\textbf{24-G3c-Cu}$_{\mathbf{48}}$, which may explain its better 
efficiency~\cite{73}. To gain insight in the differences observed
between \textbf{24-G3a} and \textbf{24-G3a-Cu}$_{\mathbf{48}}$,
their mode of action and cell death pathways were examined. Dendrimer
\textbf{24-G3a} moderately activated caspase-3 activity, an apoptosis
inducer leading to DNA fragmentation. Dendrimer
\textbf{24-G3a-Cu}$_{\mathbf{48}}$ induced a noticeable
translocation of Bax (pro-apoptotic protein) to the mitochondria,
resulting in the release of apoptosis inducing factor (AIF protein)
into the cytosol, which led to a severe DNA fragmentation without
alteration of the cell cycle. Such a mechanism is in line with the
higher activity of the Cu complex compared to the non-complexed 
dendrimer~\cite{74}.

\begin{figure*} 
\includegraphics{fig06}
\caption{\label{fig6}Generations 1 to 3 of PPH dendrimers
functionalized with different types of pyridine-imine fragments and
complexing Cu, Au, or Fe.}
\vspace*{8pt}
\end{figure*}

Dendrimer \textbf{24-G3a-Cu}$_{\mathbf{48}}$ was associated with
different types of anticancer drugs, cisplatin (alkylating agent that
binds to DNA), camptothecin (topoisomerase I inhibitor), paclitaxel
(antimitotic agent), doxorubicin (DNA intercalator, topoisomerase II
inhibitor), and MG132 (proteasome inhibitor)  (Figure~\ref{fig7}). The
combinations were tested at the active dose of each compound on KB and
HL60 cancer cell lines. No effect was observed with the combination of
\textbf{24-G3a-Cu}$_{\mathbf{48}}$ with camptothecin on either type
of cells. No effect was neither observed in the case of
\textbf{24-G3a-Cu}$_{\mathbf{48}}$\ ${+}$ cisplatin on KB cells, but
an additive effect was observed with this combination on HL60 cells, as
well as in the case of the combinations of paclitaxel or MG132 on both
types of cells. A synergistic effect was observed for the combination
\textbf{24-G3a-Cu}$_{\mathbf{48}}$\ ${+}$ doxorubicin, meaning that
the effect of this combination exceeds the sum of the inhibition
of each single active  compound~\cite{75}. 

\begin{figure*}
\includegraphics{fig07}
\caption{\label{fig7}Association of different anticancer drugs with
dendrimer \textbf{24-G3a-Cu}$_{\mathbf{48}}$ and the effect of these
combinations on the viability of two cancer cell lines.}
\end{figure*}

After copper, the complexation of gold by the same dendritic ligands
was attempted using AuCl$_{3}$. Interestingly, complexation
occurred differently in this case  with [AuCl$_{2}$]$^{+}$ being
complexed by the ligand, while [AuCl$_{4}$]$^{-}$ was the counter ion
(compound \textbf{24-G3a-Au}$_{\mathbf{48}}$)  (Figure~\ref{fig6}).
This gold complex \textbf{24-G3a-Au}$_{\mathbf{48}}$ was found
active at the low nanomolar range against KB  $(\mathrm{IC}_{50} =
7.5~\mathrm{nM})$ and HL60  $(\mathrm{IC}_{50} = 3.3~\mathrm{nM})$
cells, to be compared to the copper complex
\textbf{24-G3a-Cu}$_{\mathbf{48}}$ against KB  $(\mathrm{IC}_{50} =
470~\mathrm{nM})$ and HL60  $(\mathrm{IC}_{50} = 580~\mathrm{nM})$
cells. In view of this striking difference, a series of dendrimers
stochastically functionalized with copper, gold, free ligand
\textbf{a}, or polyethylene glycol (PEG) was synthesized 
(Figure~\ref{fig8}) in order to evaluate a potential synergistic
effect. It was shown that ten gold complexes per dendrimer were
sufficient to observe an activity in the low nanomolar range,
regardless of the other  substituents~\cite{76}. In addition, an iron
complex (compound \textbf{24-G3a-Fe}$_{\mathbf{48}}$) was
synthesized and tested, but it was found less active than the
corresponding Cu  complex~\cite{75}  (Figure~\ref{fig6}).

\begin{figure*}
\includegraphics{fig08}
\vspace*{5pt}
\caption{\label{fig8}Random functionalization at the surface of PPH
generation 3 dendrimers.}
\vspace*{10pt}
\end{figure*}

A series of dendrons having an alkyl chain at the core and the same
ligands at the surface as the dendrimers shown in  Figures~\ref{fig6}
and~\ref{fig8} was synthesized. These dendrons were equipped with an
alkyl chain of variable length at the core (C$_{11}$ for
\textbf{25a-C}$_{\mathbf{11}}$\textbf{-G1}, or C$_{17}$ for
\textbf{25a-C}$_{\mathbf{17}}$\textbf{-G1}), and complexed either
with copper or gold on the pyridine-imine ligands of type \textbf{a} 
(Figure~\ref{fig9}A). These dendrons were tested against aggressive
breast cancer cell lines (4T1, MCF-7). All these dendrons displayed
significant antiproliferative activities. The best results were
obtained with the shorter-chain dendron complexing gold
(\textbf{25a-C}$_{\mathbf{11}}$\textbf{-G1-Au}$_{\mathbf{10}}$). 
The mechanism of action involved the translocation of Bax into the
mitochondria as  previously~\cite{77}. Another family of dendrons was
equipped with two alkyl chains at the core linked to a triazine and
having as peripheral functions a type-\textbf{c} ligand for complexing
either copper or gold (Figure~\ref{fig9}B). These dendrons formed
micelles in water (mean diameter ${\sim}$9~nm for
\textbf{26c-G0-Cu}$_{\mathbf{5}}$) and multimicellar aggregates
(mean diameter  ${\sim}$60~nm for
\textbf{26c-G0-Au}$_{\mathbf{5}}$). Both dendritic complexes were
tested against several strains of glioblastoma, a malignant brain tumor
(BTSC233, JHH520, NCH644,  SF188 [pediatric], and U87 cell lines).
IC$_{50}$ values were in the 3--6~$\upmu$M range with
\textbf{26c-G0-Cu}$_{\mathbf{5}}$ and  11--15~$\upmu$M with
\textbf{26c-G0-Au}$_{\mathbf{5}}$, to be compared to
$\mathrm{IC}_{50}>100~\upmu\mathrm{M}$ with temozolomide, the clinical
standard used against  glioblastoma~\cite{78}.

\begin{figure*}
\includegraphics{fig09}
\vspace*{4pt}
\caption{\label{fig9}Dendrons equipped with one (A) or two
alkyl chains (B) at the core and having copper or gold
complexes at the periphery.}
\vspace*{4pt}
\end{figure*}

Two series of first- and second-generation dendrons having a
fluorescent group or an azabisphosphonate group at the core and either
pyrrolidinium (\textbf{27a-c-Gn},  $\mathbf{n} = 1, 2$) or piperidinium
(\textbf{28a-c-Gn},  $\mathbf{n} = 1, 2$) peripheral functions were
synthesized  (Figure~\ref{fig10}), and their critical micelle
concentrations (CMC) were measured. The lower CMC values (1.42 to 
3.56~$\upmu$M) were obtained with the pyrene series (dendrons
\textbf{27a-Gn} and \textbf{28a-Gn},  $\mathbf{n} = 1, 2$), whereas the
higher values (45.5 to 153.7~$\upmu$M) were obtained in the case of the
azabisphosphonate group at the core (dendrons \textbf{27c-Gn} and
\textbf{28c-Gn},  $\mathbf{n} = 1, 2$). 
The~\mbox{antiproliferative} activity
of these dendrons were tested against HL60, K562, and HCT116 tumor cell
lines. Best results were obtained with the piperidinium family, in
particular of generation 2 (dendrons \textbf{28a-c-G2}). Dendron
\textbf{28c-G2} was tested against a larger panel of cancerous and
non-cancerous cells, and it displayed lower toxicity toward normal
mouse fibroblast L929 cells (8.75~$\upmu$M) than toward nine tumor cell
lines (0.27 to  4.1~$\upmu$M)~\cite{79}. 

\begin{figure*}
\includegraphics{fig10}
\vspace*{4pt}
\caption{\label{fig10}Different types of dendrons functionalized with
N-heterocycles.}
\vspace*{10pt}
\end{figure*}

\begin{figure*}
\includegraphics{fig11}
\vspace*{4pt}
\caption{\label{fig11}(A)~Viologen dendrimer and (B)~small dendrimers,
both having antibacterial activities.}
\end{figure*}

Besides the anticancer properties, a few PPH dendrimers were tested
against bacterial strains. For instance, a few dendrimers shown in 
Figures~\ref{fig6} and~\ref{fig8}
(\textbf{24-G3a-Au}$_{\mathbf{48}}$,
\textbf{24-G3a-Au}$_{\mathbf{10}}$\textbf{Cu}$_{
\mathbf{20}}$\textbf{NN}$_{\mathbf{10}}$\textbf{PEG}$_{\mathbf{8}}$,
\textbf{24-G3a-Au}$_{\mathbf{20}}$\textbf{NN}$_{
\mathbf{20}}$\textbf{PEG}$_{\mathbf{8}}$, 
\textbf{24-G3a-Cu}$_{\mathbf{44}}$\textbf{PEG}$_{\mathbf{4}}$) 
were also tested against Gram-positive (\textit{Staphylococcus aureus})
and Gram-negative (\textit{Escherichia coli}, \textit{Pseudomonas
aeruginosa}) bacteria, as well as against yeasts (\textit{Candida
albicans}). Dendrimer \textbf{24-G3a-Au}$_{\mathbf{48}}$ had the
highest antimicrobial activity, whereas
\textbf{24-G3a-Au}$_{\mathbf{10}}$\textbf{Cu}$_{\mathbf{20}}
$\textbf{NN}$_{\mathbf{10}}$\textbf{PEG}$_{\mathbf{8}}$
displayed a marked synergistic effect between both metals, granting
this compound the highest anti-fungal  activity~\cite{76}. 
A~\mbox{series} of
viologen-containing dendrimers was synthesized, and the antibacterial
properties were assessed. As the viologens are inside the branches, and
not at the surface, they are out of the scope of this review, but one
of them is shown in  Figure~\ref{fig11}A (\textbf{29-G1}), as it
displayed the best antibacterial properties against the Gram-positive
strain \textit{S.~aureus}, and limited the growth of Gram-negative
strains \textit{E.~coli} and \textit{P.~vulgaris}, due to the presence
of a large number of charges inside the  structure~\cite{80}. Two
series of small dendrimers (\textbf{30a-g-G0} and
\textbf{31a-c,e-g-G0}) functionalized with diverse acetohydrazides,
most of them bearing a nitrogen heterocycle, were synthesized 
(Figure~\ref{fig11}B). All of them displayed activities against a panel
of bacterial pathogens, the most active being compound 
\textbf{30a-G0}~\cite{81}.

\begin{figure*}
\includegraphics{fig12}
\caption{\label{fig12}Three families of PPH dendrimers functionalized
with cationic N-heterocycles against tuberculosis.}
\vspace*{5pt}
\end{figure*}

Tuberculosis is an infectious disease caused by the bacillus
\textit{Mycobacterium tuberculosis} (\textit{Mtb}), which, according to the World
Health Organization (WHO), has already infected more than one third of
the world's  population~\cite{82}. Three families of phosphorus
dendrimers functionalized with nitrogen 
\mbox{heterocycles} were synthesized
(compounds \textbf{32a-e-G0}, \textbf{33a-g-G1},
\textbf{33a-b-G2},\textbf{G3},\textbf{G4}, and \textbf{34a-i-G1}) 
(Figure~\ref{fig12}). They were tested against three bacterial strains:
attenuated \textit{Mycobacterium tuberculosis} H37Ra, virulent
\textit{M.~tuberculosis} H37Rv, and \textit{Mycobacterium bovis} BCG. The
smallest compounds were found to be the most active, in particular
compounds \textbf{32a-G0} (six pyrrolidinium groups) and
\textbf{32b-G0} (six piperidinium groups). In addition, \textbf{32b-G0}
showed relevant efficiency against \textit{M.~tuberculosis} strains
resistant to widely used drugs such as rifampicin, isoniazid,
ethambutol, or streptomycin. Thus, compound \textbf{32b-G0} was tested
in vitro and in vivo. It was administered orally once a day for two
weeks to mice infected by the attenuated H37Ra \textit{Mtb} strain. A superior
efficiency of this compound compared to ethambutol and rifampicin was
observed in  vivo~\cite{83}.

\begin{figure*}
\includegraphics{fig13}
\vspace*{4pt}
\caption{\label{fig13}Different types of dendrons functionalized with
N-heterocycles and used as carriers of RNA.}
\vspace*{5pt}
\end{figure*}

Beside the properties observed as drugs {per se}, PPH dendrimers
have been also used as drug and DNA carriers. Dendrimers
\textbf{33a,h,i,j-G1,G4} 
\mbox{(Figure~\ref{fig12})} were synthesized, but
only the fourth generations were tested, except compound
\textbf{33j-G4} which was discarded as it was not soluble in water.
Cytotoxicity of dendrimers \textbf{33a,h,i-G4} was found to be low
against both cancerous and normal cells. Their ability to interact with
DNA was tested by electrophoresis. Only the dendrimer bearing the
pyrrolidinium groups (\textbf{33a-G4}) was found to be suitable for
transfection experiments, to deliver single-stranded (labelled with
FITC) and double-stranded (GFP-coding plasmids) DNA into one healthy
(HUVEC) and two cancerous (HEK 293 and HeLa) cell  lines~\cite{84}. The
third generation of the same dendrimer (\textbf{33a-G3}) was used to
carry a small interfering RNA, anti-Lyn, as a potent anticancer agent
against glioma. In addition to delivery, the dendrimer itself
influenced various cell parameters, in particular those suggested to be
regulating glioblastoma cell  invasion~\cite{85}. Generations 1 to 3 of
the same dendrimer (\textbf{33a-G1,G2,G3}) were used to condense
plasmid DNA (pDNA) encoding enhanced green fluorescent protein (EGFP),
and to deliver it to HeLa cancer cells. Compound \textbf{33a-G1}
displayed the best gene-delivery efficiency. This small dendrimer was
then used to deliver pDNA encoding both EGFP and p53 protein, resulting
in cell cycle arrest for cancer-gene-therapy applications. Such
properties were also validated in vivo with mice bearing a xenografted 
tumor~\cite{86}. Dendrimers \textbf{33a-G3} and \textbf{20-G3}
complexing Mcl-1 siRNA were used as dopants of hydrogels (agarose
gels). Combinations of both dendrimers in various proportions permitted
to adjust the speed of siRNA  release~\cite{87}. 

Different types of PPH dendrons functionalized with nitrogen
heterocycles were also used as carriers. A third-generation dendron,
having additional phenyl groups inside the structure to increase the
hydrophobicity, a C$_{17}$ alkyl chain at the core, and 40
pyrrolidinium groups at the periphery (compound \textbf{35-G3}) 
(Figure~\ref{fig13}A) formed micelles that possessed good intrinsic
anticancer activity. These micelles were suitable for encapsulation of
the hydrophobic anticancer drug doxorubicin (DOX) with high
drug-loading content (42.4\%) and encapsulation efficiency (96.7\%).
The micelles of dendrons loaded with doxorubicin acted collectively to
take down breast cancer cells, both in vitro and in vivo, in a
xenografted tumor model. Furthermore, the micelles \textbf{35-G3@DOX}
significantly decreased the intrinsic toxicity of free 
doxorubicin~\cite{88}. A first-generation dendron bearing also
pyrrolidinium peripheral groups and a long alkyl chain at the core
(compound \textbf{36-G1})  (Figure~\ref{fig13}B) also formed micelles
suitable to encapsulate doxorubicin with optimal loading content (80\%)
and encapsulation efficiency (98\%). Compound \textbf{36-G1@DOX} was
able to compress microRNA-21 inhibitor (miR-21i), and to co-deliver it
for combination therapy of triple-negative breast cancer. This polyplex
was readily phagocytosed by cancer cells, which killed them in vitro.
It was also used to efficiently treat an orthotopic triple-negative
breast tumor model in vivo, as demonstrated by a large decrease in
tumor size for the polyplex-treated mice~\cite{89}. Even a very small
dendron (\textbf{37-G0})  (Figure~\ref{fig13}C) was able to compress
microRNA-30d (miR-30d) and form polyplexes that effectively transfected
miR-30d to cancer cells. miR-30d is known to significantly inhibit the
migration and invasion of a murine breast cancer cell line. This
phenomenon was indeed observed with \textbf{37-G0@miR-30d} both in
vitro and in vivo in a subcutaneous tumor mouse  model~\cite{90}.

\begin{figure*}
\includegraphics{fig14}
\caption{\label{fig14}Dendrimers functionalized with fipronil.}
\end{figure*}

Another way to deliver active substances consists in grafting them to
the surface of dendrimers through a cleavable bond. Such a concept was
illustrated with PPH dendrimers for fipronil, an insecticide having a
poorly reactive NH$_{2}$ group, which was nevertheless reacted with the
aldehyde terminal functions of generation 1 and 4 dendrimers to
generate compounds \textbf{38-G1} and \textbf{38-G4} 
(Figure~\ref{fig14}). The imine bonds were very sensitive to water,
which would release fipronil. These dendrimers, kept as powders under
air for 35 days, released 12\% and 37\% of fipronil from \textbf{38-G1}
and \textbf{38-G4}, respectively. As it was difficult to keep these
compounds pure and prevent the hydrolysis of the imine bonds, another
attempt was carried out to have an amine instead of an imine for the
grafting to the dendrimers (compounds \textbf{39-Gn},  $\mathbf{n} = 1,
4$)  (Figure~\ref{fig14}). These compounds were stable for months as
powders under air. Despite being much less sensitive to hydrolysis than
the \textbf{38-Gn} 
\mbox{family,} the \textbf{39-Gn} \mbox{family} retained a certain
persistence of the pesticide activity, possibly through a different
mechanism of  action~\cite{91}.

\section{PPH dendrimers functionalized with oxygen
heterocycles}\label{secIII}
Few oxygen heterocycles were grafted to the surface of phosphorus
dendrimers and dendrons, all of them being of the carbohydrate type.
Very small dendrons bearing a phenyl, a ferrocene, or a phosphonate
group at the core were functionalized with $\upbeta$-D-glucoside groups
(compounds \textbf{40a,b,c-G0})  (Figure~\ref{fig15})~\cite{92,93}.
Generations 1 and 4 of dendrimers (\textbf{41-G1} and \textbf{41-G4})
were functionalized with the same $\upbeta$-D-glucoside groups. Isomers
of the hydrazone bonds were observed by $\mbox{}^{31}$P,
$\mbox{}^{1}$H,  and  $\mbox{}^{13}$C NMR~\cite{94}.

\begin{figure*}
\includegraphics{fig15}
\caption{\label{fig15}Functionalization of dendrons and dendrimers with
$\upbeta$-D-glucoside groups.}
\end{figure*}

Generations 1 to 3 of PPH dendrimers were functionalized with
\textit{para}-hydroxyphenyl-2,3,4-tri-$O$-acetyl-$\upbeta$-D-xylopyranoside, 
affording dendrimers \textbf{42a-Gn} 
$(\mathbf{n} = 1, 2, 3)$, then deprotection of the alcohols by
hydrolysis of the acetate groups produced dendrimers \textbf{42b-Gn} 
$(\mathbf{n} = 1, 2, 3)$  (Scheme~\ref{sch10})~\cite{95}.

\begin{scheme*}
\includegraphics{sc10}
\vspace*{4pt}
\caption{\label{sch10}Grafting protected D-xylose at the surface of
dendrimers and its deprotection.}
\vspace*{-10pt}
\end{scheme*}

A series of dendrimers capped with mannose units was synthesized with
the aim of mimicking the bioactive supramolecular structure of
mannose-capped lipoarabinomannan, one of the most abundant glycolipids
in  the  \textit{M.~tuberculosis}  cell wall, which decreases the
immune response in favor of promoting the infection. The series of 
dendrimers \textbf{43-Gn-X} differs by the generation  ($\mathbf{n}= 1$
to 4) and the type of mannose caps, being constituted of a single
mannose  $(\mathbf{X}= \mathrm{M})$, or two mannoses connected
$(\upalpha 1\rightarrow2)\;  (\mathbf{X}= \mathrm{D})$, or even three
mannoses  $(\mathbf{X}= \mathrm{T})$  (Figure~\ref{fig16}). The binding
avidity of all these dendrimers for DC-SIGN (dendritic cell-specific
intercellular adhesion molecule-3-grabbing non-integrin) was measured.
The best results (highest binding avidity) were obtained  with 
\textbf{43-G3-T} (48 trimannosides) and \textbf{43-G4-D} (96
dimannosides). These dendrimers inhibited proinflammatory cytokines. 
Furthermore, {per os} administration of the \textbf{43-G3-T}
mannodendrimer to mice exposed to aerosolized lipopolysaccharide (LPS),
as a model of acute lung inflammation, induced a significantly reduced
neutrophil influx. It was thus found suitable for the potential
treatment of lung inflammatory  diseases~\cite{96}. Later on,
dendrimers \textbf{43-G2-M} and \textbf{43-G2-D} were used in a
structure/function relationship study, which demonstrated that
dimannoside caps (as  in \textbf{43-G2-D}) and multivalent interactions
were required for efficient ligand  binding~\cite{97}. 

\begin{figure*}
\includegraphics{fig16}
\caption{\label{fig16}Dendrimers functionalized with 1, 2, or 3 mannose
units on each terminal function.}
\end{figure*}

\begin{figure*}
\includegraphics{fig17}
\caption{\label{fig17}PPH dendrimers functionalized with
S-heterocycles, thiophene, and TTF derivatives.}
\end{figure*}

\begin{scheme*}
\includegraphics{sc11}
\vspace*{4pt}
\caption{\label{sch11}Alternative methods of synthesis of phosphorus
dendrimers.}
\vspace*{-10pt}
\end{scheme*}

\section{PPH dendrimers functionalized with sulfur
heterocycles}\label{secIV}
Very few S-heterocycles were used for the functionalization of PPH
dendrimers, all of them were synthesized for electrochemical
experiments. In a first example, generations 0 to 4 of phosphorus
dendrimers were functionalized with two thiophene groups, using Wittig
reactions on the aldehydes (compounds \textbf{44-Gn},  $\mathbf{n} = 0$
to 4)  (Figure~\ref{fig17}A). Electropolymerization of these dendrimers
was obtained by applying recurrent potential scans, which induced the
growing of a polymer in the form of a dark blue film on the anode
surface. Polymers formed with \textbf{44-G0} and \textbf{44-G1}
essentially contained inter-dendrimer linkages, whereas an increasing
part of intra-dendrimer linkages was observed for higher 
generations~\cite{98}. PPH dendrimers were also functionalized by
\mbox{Wittig} 
\mbox{reactions} with tetrathiafulvalene (TTF) derivatives
(\textbf{45a-Gn},  $\mathbf{n} = 1$ to 5, from 6 to 96 TTF moieties) 
(Figure~\ref{fig17}B), sometimes linked to a crown ether
(\textbf{45b-G3}). Electrodes modified with these dendrimers were
obtained by electrodeposition, with no significant difference observed
between the different dendrimer generations. Electrodes modified with
dendrimer \textbf{45b-G3} were suitable for detecting and quantifying
Ba$^{2+}$ cations, which could be coordinated by the TTF--crown ether 
ligands~\cite{99}. Generation~1 of PPH dendrimers was functionalized
with TTF linked through different types of linkers
(\textbf{46a,b,c,d,-G1})  (Figure~\ref{fig17}C). Thin-layer cyclic
voltamperometry results were in full agreement with twelve TTF units
per dendrimer in all cases. Chemical oxidation of the TTF moieties with
PhI(OAc)$_{2}$/CF$_{3}$SO$_{3}$H induced the formation of mixed-valence
cation radical salts, which were conductive. They were characterized by
measurement of optical density in the solid films and by UV--Vis--NIR
spectroscopy in  solution~\cite{100}.

\section{PPH dendrimers functionalized with mixed
heterocycles}\label{secV}
A few mixed heterocycles incorporating both nitrogen and oxygen have
been already shown in  Figure~\ref{fig12}, as they were part of a
series of nitrogen heterocycles. However, other types of mixed
heterocycles incorporating other elements, in particular phosphorus,
were used for the functionalization of PPH dendrimers. In that case,
either the P-heterocycles were pre-synthesized before the grafting to
dendrimers, or they were obtained by reaction of the P(S)Cl$_{2}$
terminal functions with diamines.

Besides the classical two-step synthesis of PPH dendrimers, several
alternative methods for the synthesis of phosphorus dendrimers have
been proposed. One of them consisted in the repetition of three steps
and involved the use of a 
{chlorodiazaphospholane.} As illustrated 
in Scheme~\ref{sch11}A, this \mbox{diazaphospholane} was reacted with N--H
peripheral functions to afford dendrimer \textbf{47-G0}, which was
reacted with a phosphorus azide in a Staudinger  reaction~\cite{101},
then with methylhydrazine, and again with the chlorodiazaphospholane to
afford dendrimer  \textbf{47-G1}~\cite{102}. Another alternative method
for the synthesis of phosphorus dendrimers consisted in using P${=}$O
instead of P${=}$S derivatives. Condensation of the aldehyde peripheral
functions with a methylhydrazine derivative of oxazaphospholidine
afforded dendrimer \textbf{48-G1}  (Scheme~\ref{sch11}B)~\cite{103}. 

\begin{scheme*}
\includegraphics{sc12}
\vspace*{6pt}
\caption{\label{sch12}PPH dendrimers functionalized with cage
compounds, such as PTA, and used as catalysts.}
\vspace*{-4pt}
\end{scheme*}

Cage compounds based on phosphorus and their complexes are mainly used
in  catalysis~\cite{104}. A trihydrazino phosphoadamantane
functionalized with a pendant amine was reacted with the P(S)Cl$_{2}$
functions of a first-generation dendrimer, and then the remaining free
phosphine was complexed with ruthenium \textit{para}-cymene, to afford
complex \textbf{49-G1}  (Scheme~\ref{sch12}A), but this complex was not
used in catalysis  experiments~\cite{105}.
1,3,5-Triaza-7-phosphaadamantane  (PTA)~\cite{106} was grafted to the
surface of PPH dendrimers from generation 1 to 3 and used for
complexing ruthenium, affording dendrimers \textbf{50-Gn} ($\mathbf{n}
= 1$ to 3), and the corresponding monomer \textbf{50-M} was also
synthesized  (Scheme~\ref{sch12}B). These complexes were tested as
catalysts in the hydration of phenylacetylene, and also in the
isomerization of 1-octan-3-ol  (Scheme~\ref{sch12}B). In that case,
biphasic water/heptane reaction conditions were used, which enabled the
easy recovery of the product in the organic phase, and of the dendritic
catalysts in the aqueous phase. The efficiency of the catalysts
increased as the generation increased, and recycling experiments with
the first-generation dendrimer \textbf{50-G1} were efficiently carried
out three  times~\cite{107}. Another dendrimer having twice the number
of PTA at a given generation (\textbf{51-G1}) was synthesized. The
efficiency of dendrimer \textbf{51-G1} (24 PTA-Ru) was compared to that
of \textbf{50-G1} (12 PTA-Ru) and \textbf{50-G2} (24 PTA-Ru) in the
hydration of alkynes  (Scheme~\ref{sch12}C). The dense dendrimer
\textbf{51-G1} was the most efficient in this  reaction~\cite{108}. As
some ruthenium derivatives were proposed as anticancer agents, these
\mbox{dendrimers} were also tested for their ability to interact with DNA.
Dendrimer \textbf{50-G3} was discarded as insoluble in water, whereas
dendrimers \textbf{50-G1} and \textbf{50-G2} were poorly soluble. Only
the monomer \textbf{50-M} and the smallest dendrimer \textbf{50-G0}
were soluble in water. Tests for unwinding supercoiled DNA indicated
that dendrimer \textbf{50-G0} was the most  efficient~\cite{109}.

\begin{figure*}
\includegraphics{fig18}
\vspace*{2pt}
\caption{\label{fig18}Rh-complexing chiral phosphoramidite ligand
linked to the surface of dendrimers used as enantioselective
catalysts.}
\vspace*{4pt}
\end{figure*}

A chiral phosphoramidite ligand, suitable for the complexation of
rhodium, was grafted on a model \textbf{52-M}, on a branch
(\textbf{52-B}), and at the surface of dendrimers from generation 1
(\textbf{52-G1}) to generation 3 (\textbf{52-G3}) 
(Figure~\ref{fig18}). The resulting Rh-complexes were used as catalysts
in  the $[2+2+2]$-cycloaddition reaction of three alkynes, between
$N$-tosyl-1,6-diyne and 2-methoxynaphthalene alkynyl derivatives. A
strong positive dendritic effect was observed both in yield and
enantioselectivity of this reaction (determined by chiral HPLC),
leading to axially chiral biaryl  compounds~\cite{110}. 

Besides the direct grafting of phosphorus-containing heterocycles, the
P(S)Cl$_{2}$ terminal functions are particularly suitable to react with
diamines to produce five- or six-membered heterocycles involving a
N--P--N linkage. Such a reaction was first applied to generations 1 and
3, reacted with different tetraazamacrocycles
(1,4,8,11-tetraazacyclotetradecane, 1,4,8,12-tetraazacyclopentadecane,
and 1,4,8,11-tetraazacyclotetradecane-5,7-dione)  (Scheme~\ref{sch13}).
Five-membered rings (diazaphospholanes) were preferred when either
five- or six-membered rings could be produced (cases \textbf{53a-Gn}
and \textbf{53b-Gn}). A six-membered ring (diazaphosphorinane) was
obtained only in the case \textbf{53c-Gn}, where there was no
possibility of creating a five-membered ring because of the presence of
the amide  moieties~\cite{111}.

\begin{scheme*}
\includegraphics{sc13}
\vspace*{4pt}
\caption{\label{sch13}Reaction of macrocyclic tetraamine derivatives
with P(S)Cl$_{2}$ functions of dendrimers.}
\vspace*{-10pt}
\end{scheme*}

The same process was applied for the grafting of triazatriolefinic
macrocycles functionalized with a diamine linker. Generations 1, 2, and
4 were synthesized by using the diamine to produce diazaphospholane
rings (compounds \textbf{54-Gn},  $\mathbf{n} = 1, 2, 4$) 
(Figure~\ref{fig19})~\cite{112}. The dendritic macrocycles reacted with
Pt$_{2}$(dba)$_{3}$ (dba ${=}$ dibenzylideneacetone) and formed platinum
nanoparticles (Pt-NPs) in very mild conditions. Interestingly, the
dendrimers assembled these Pt-NPs in branched networks, for which the
degree of branching and the length increased as the generation of the
dendrimers increased. A very unique organization of organic dendritic
structures interweaved with inorganic dendritic structures was observed
for the first time with these  dendrimers~\cite{113}. 

\begin{figure*}
\includegraphics{fig19}
\caption{\label{fig19}Triazatriolefinic macrocycles grafted to PPH
dendrimers through diazaphospholane rings.}
\end{figure*}

A few mixed heterocycles not involving phosphorus were grafted to the
surface of PPH dendrimers, essentially for catalytic experiments. Two
families of dendrimers functionalized with bis(oxazoline) ligands,
either linked through a phenol (dendrimers \textbf{55a-Gn}, 
$\mathbf{n} = 1, 2, 3$) or through a triazole issued from a click
reaction (dendrimers \textbf{55b-Gn},  $\mathbf{n} = 1, 2, 3$, 
Scheme~\ref{sch14}) were found suitable for complexing CuCl$_{2}$. The
resulting Cu complexes were used as catalysts in asymmetric
benzoylation of linear (case \textbf{a} in  Scheme~\ref{sch14}) and cyclic diols
(case \textbf{b}). Best results were obtained in the case of the acyclic diol.
The \textbf{55b} family was more efficiently recovered than the
\textbf{55a} family, and it could be reused twice with still good yield
and  enantioselectivity~\cite{114}.

\begin{scheme*}
\includegraphics{sc14}
\vspace*{5pt}
\caption{\label{sch14}Two families of dendrimers functionalized with
bis(oxazoline) ligands and used in asymmetric catalysis.}
\vspace*{8pt}
\end{scheme*}

Thiazoles have many biological properties~\cite{115}, but they were
grafted to the surface of dendrimers for their catalytic properties.
Indeed, depending on the substituents they bear, in particular
phosphines, they can complex metals, to be used as 
catalysts~\cite{116}. A thiazolyl phosphine was grafted to the surface
of generations 1 and 3 of PPH \mbox{dendrimers} \mbox{affording} \mbox{compounds}
\textbf{56-Gn}  ($\mathbf{n} = 1$ and 3,  Scheme~\ref{sch15}), which
were suitable for complexing Pd(OAc)$_{2}$. The dendritic Pd complexes
and the corresponding monomer were used as catalysts in a panel of
Suzuki couplings between diversely functionalized phenyl bromide and
phenyl boronic acids. A negative dendritic effect was observed, as the
first generation \textbf{56-G1} and the monomer \textbf{56-M} were more
efficient than the third generation \textbf{56-G3}. The
first-generation \textbf{56-G1} could be recovered and reused at least
four times, with still the same efficiency, whereas recovery was not
possible with the monomer \textbf{56-M}. Leaching measured by
inductively coupled plasma mass spectrometry (ICP-MS) was very high for
the monomer, but undetectable with the dendrimer, displaying an
important advantage when using dendrimers instead of monomers as 
catalysts~\cite{117}.

\begin{scheme*}
\includegraphics{sc15}
\vspace*{5pt}
\caption{\label{sch15}Dendrimers functionalized with thiazolyl
phosphine and use of these Pd complexes as catalysts in Suzuki
couplings.}
\vspace*{-8pt}
\end{scheme*}

\section{Conclusion}\label{secVI}
Phosphorus dendrimers functionalized with a panel of heterocycles (N-,
O-, S- and mixed-heterocycles) represent versatile platforms for
advanced 
\mbox{catalysis,} materials, and biomedical applications. These
dendrimers with surface heterocycles can stabilize metal complexes,
forming catalytic systems for organic transformations. These systems
were used in reactions such as Stille couplings, Knoevenagel
\mbox{condensations,} and Michael additions, with the advantage of catalyst
recoverability and reusability. The high reproducibility of the results
is explained by the fact that the proposed dendrimers containing
heterocyclic rings have a specific chemical formula and a specific
stereochemical structure, which can be easily confirmed by
spectroscopic methods.

The tunable properties of dendrimers functionalized with heterocycles
and multivalency also make them ideal candidates for gene delivery and
cancer therapeutics. Such phosphorus dendrimers are being actively
explored as non-viral nanoplatforms for gene delivery, showing promise
in cancer therapeutics. Their biocompatibility and multivalency allow
for efficient encapsulation and targeted delivery of therapeutic genes
and drugs. It should be noted that by programming the synthesis,
grafting, and enhancement of potentially biologically active
heterocycles using various methods, it is possible to program the
relative stability of target dendrimers in the relevant biological
environment. Thus, stable dendrimers can be tested for their diverse
biological activities, while in dendrimers specially synthesized for
the delivery of future and already existing drugs, focus can be placed
on hydrolysis or another mechanism for releasing heterocyclic compounds
at the right time and in the required spatial location.\looseness=1

In conclusion, recent advances in the synthesis and applications of
phosphorus dendrimers functionalized with heterocycles highlight their
versatility and potential. Innovations in eco-friendly synthesis,
biomedical applications (especially in drug delivery against cancers),
and catalysis are driving the field forward. The integration of
heterocycles not only enhances the functionality of dendrimers but also
opens new avenues for targeted therapies, advanced materials, and
sustainable catalysis.

Given the high potential of useful practical properties, it is
important to emphasize that this branch of chemistry makes a
significant contribution to fundamental science, both in enriching the
fields of dendrimer chemistry and heterocyclic chemistry, if only
because the simultaneous or sequential extension of the branches of a
phosphorus dendrimer requires virtuosity and precision in practical
methods of~working with high-molecular weight compounds, especially chemo-,
regio-, and stereoselective 
\mbox{modifications.}

\printCOI

\back{}

\printbibliography
\refinput{crchim20260018-reference.tex}

\end{document}