Text crossing the margin in amsbook

Question

I am currently using the amsbook class to write my thesis. It happens quite regularly that some text lines cross the margin of the document. So far, I have no clue why and in which cases that happens (e.g., it does not only happen in math environments). Does anyone know why this happens and how to fix it?

Here is some minimal example from my thesis, where this happens. I have included the packages that I use, but the same problem occurs if I do not use any of them.

\documentclass{amsbook}

\usepackage[english]{babel}
\usepackage{amsmath, amsthm, amsfonts,amssymb}
\usepackage[T1]{fontenc}
\usepackage[utf8]{inputenc}\usepackage{lmodern}
\usepackage[bookmarks,bookmarksopen,bookmarksnumbered,colorlinks,linkcolor=blue,citecolor=red,urlcolor=grey,breaklinks]{hyperref}
\usepackage{enumitem}
\usepackage{graphicx}
\usepackage{tikz}\usetikzlibrary{positioning}
\usepackage{tikz-cd}
\usepackage{accents}
\usepackage{mathtools}
\usepackage{todonotes}
\usepackage{mathrsfs}

\theoremstyle{definition}
\newtheorem{Def}{Definition}[chapter]

\begin{document}
    \begin{Def}[Polynomial Hubbard Trees]
        Let $p$ be a post-critically finite polynomial. The \emph{Hubbard Tree} of $p$ is the unique finite embedded tree $H\subset\mathbb{C}$ for which the following conditions are satisfied:
        \begin{itemize}
            \item $\Omega^+(C(p))\subset H$ and all endpoints of $H$ are contained in $\Omega^+(C(p))$. (\textbf{Minimality})
            \item The embedded tree $H$ is forward invariant, i.e., $p(H)\subset H$. (\textbf{Invariance})
            \item For every pair of points $z,w\in H$, the in-tree connection $[z,w]_H$ is a regulated arc. (\textbf{Normalization})
        \end{itemize}
    \end{Def}
\end{document}

Answer 1

it seems that \textbf{...} in combination with left parent before it (see @egreg comment below; thank you very mach Gregorio!) block automatic hyphenation of the text inside it. it is not connected with amsbook document class, but also happen for example at use of article. so, to my opinion, this is bug ...

you have two possibilities to resolve your problem:

• manually insert hyphenation points as suggested in Red-Cloud answer
• instead (\textbf{...}) use ({\bfseries ...}) as i suggested in my comment below question

some off-topic remarks:

• the hyperref package should be loaded last (in rare exception as is cleveref)
• mathtolls load amsmath (and improve it), so you not need load again
• amssymb load amsfonts, so you not need to load again

• utf8 coding is included in recent (standard) document class version, so you can omit it in preamble

  \documentclass{article}
  
  \usepackage[english]{babel}
  \usepackage{mathtools,amssymb, amsthm}
  \usepackage[T1]{fontenc}
  %\usepackage[utf8]{inputenc}
  \usepackage{lmodern}
  \usepackage{mathrsfs}
  \usepackage{accents}
  \usepackage{mathtools, amssymb, amsthm}
  \usepackage{enumitem}
  \usepackage{graphicx}
  \usepackage{tikz}
  \usetikzlibrary{positioning}
  \usepackage{tikz-cd}
  \usepackage{todonotes}
  \usepackage[bookmarks,bookmarksopen,bookmarksnumbered,
              colorlinks,linkcolor=blue,citecolor=red,urlcolor=grey,
              breaklinks]{hyperref}
  
  \theoremstyle{definition}
  \newtheorem{Def}{Definition}[section]
  
  \begin{document}
      \begin{Def}[Polynomial Hubbard Trees]
          Let $p$ be a post-critically finite polynomial. The \emph{Hubbard Tree} of $p$ is the unique finite embedded tree $H\subset\mathbb{C}$ for which the following conditions are satisfied:
          \begin{itemize}
              \item $\Omega^+(C(p))\subset H$ and all endpoints of $H$ are contained in $\Omega^+(C(p))$. ({\bfseries Minimality})
              \item The embedded tree $H$ is forward invariant, i.e., $p(H)\subset H$. (\textbf{Invariance})
              \item For every pair of points $z,w\in H$, the in-tree connection $[z,w]_H$ is a regulated arc. (\textbf{Normalization})
          \end{itemize}
      \end{Def}
  \end{document}
  

result:

Answer 2

Use \textbf{Mini\-mality}.

But it is a bit curious, because there are alread the hyphenation points Min-i-mal-ity